Competition; *Electronics; Employment Aigher E - ERIC

ED 245 883

TITLEINSTITUTION

REPORT NOPUB DATENOTEAVAILABLE FROM

PUB TYPE

EDRS PRICEDESCRIPTORS

IDENTIFIERS

ABSTRACT

DOCU'tENT RESUME

SE 044 576

International Competitiveness in Electronics.Congress of the U.S., Washington, D.C. Office ofTechnology Assessment.OTA-ISC-200Nov 83547p.Superintendent of Documents, U.S. Government PrintingOffice, Washington, DC 20402.Reports Evaluative/Feasibility (142)

MF02/PC22 Plus Postage.Automation; *Competition; *Electronics; EmploymentPatterns; Engineering Education; Financial Support;Aigher Education; *Industry; *InternationalRelations; *International Trade; Manufacturing;*Policy; Science Education; Secondary Education;Technology*Industrial Policy; United States

This assessment continues the Office of TechnologyAssessment's (OTA) exploration of the meaning of industrial policy inthe United States context, while also examining the industrialpolicies of several U.S. economic rivals. The major focus is onelectronics, an area which virtually defines "high technology" of the1980's. The assessment sets the characteristics of the technologyitself alongside other forces that exert major influences overinternational competitiveness. Specific areas addressed include:electronics technology; srr,,Acture, trade, and competitiveness in theinternational electronics industry; quality, reliability, andautomation in manufacturing; role of financing in competitiveness andelectronics; human resources (education, training, management);employment effects; national industrial policies; and U.S. tradepolicies and-their effects. The report concludes by outlining fiveoptions for a U.S. industrial policy, drawing on electronics forexamples of past and prospective impacts, as well as on OTA'sprevious studies of the steel and automotive induStries. A detailedsummary and iltroductory comments are included. Also included inappendices are case studies in the development and marketing ofelectronics products, a discussion of offshore manufacturing, and aglossary of terms used in the assessment. (J1,0

*-*********************************************************************** Reproductions supplied by EDRS are the best that can be made *

* from the original document. , *

*************-*******p*******************,*************-:*****************

O

Inte ationaompetitiveness

In ElectronicsUS. DEPARTMENT OF EDUCATION

NATIONAL INSTITUTE OF EDUCATIONEDUCATIONAL RESOURCES INFORMA.TION

I CENTER IER1Ci

This document has been reproduced asrecewed from the person or organaationonginatmg ItMinor changes nave been made to improvereproduction quality,

Ponts of view or opinions Staled in this docd-ment do not necessarily represent official NIEposition or policy.

44

L13. ;" CONGRESS OF THE UNITED STATES*tee* Office of Technology Assessment

41Wastfington. D. C. 20610

Office of Technology =assessment

Congressional Board of th-. r.:41:h. Congress

MORRIS K. UDALL, Arzr:::-1,

TED STEVENS. Alaska,.

Senate

ORRIN G. HATCHUtah

CHARLES McC. MATHIAS. JR-Mary land

EDWARD M. KENNEDYMassachusetts

ERNEST F.. HOLLINGSSouth Carolina

CLAIBORNE PELLRhode Island

CHARLES N. KIMBALL. ChairmanMidwest Research Institute

EARL BEISTLINEUniversity of Alaska

CHARLES A. BOWSHERGeneral Accounting Office

CLAIRE T. DEDRICKCalifornia Land Commission

House

E. BROV' ", JR.' California

..7,":";"1-IN D.

Michigan

.:`..A,R.RY WINN, Ill.Kansas

CLAYA.F.NCE E.Ohio

'1..'90PER EVAIowa

JOHN H. G -1.717-.."N'..5

Advisory Courmif

JAMES C. FLETCHMtUniversity of Pittsburgh

S. DAVID FREEMANTennessee Palley Authority

GILBERT GUDECongressional Research Service

CARL N. HODGESUniversity of Arizona .

Director

RAC) '14,Um, f

Hamlweekt & Quist

DAVID PUTTERGeneral \P. Mors Corp.

LEWIS THOMASMemorLI Sloan-Kettering

I. Cancer Center

JOHN H. GIBBONS

The Technology Assessment Board approves the reh:ase of this report. The views expressed in this report are notnecessarily those of the Board, OTA Advisory Council, or of individual 'members thereof.

International--C

in Electronics

OTA Reports are the principal documentation of formal assessment projects.These projects are approved in advance by the Technology Assessment Board.At the conclusion of a project, the Board has the opportunity to review the report,but its release does not necessarily imply endorsement of the results by the Boardor its individual members.

CONGRESS OF THE UNITED STATES

Office of Technolnly Assessmentvidsn.ngton.o C 20510

Recommended Citation:International Competitiveness in Electronics (Washington, D.C.: U.S. Congress, Officeof Technology Assessment, OTA-ISC-200, November 1983).

Library of Congress Catalog Card Number 83-600610

For sale by the Superintendent of Documents,U.S. Government Printing Office, Washington, D.C. 20402

Foreword

This assessment, requested by the Senate Committee on Commerce, Science,and Transportation, the House Committee on Ways and Means, and the joint Eco-nomic Committee, completes a series of three reports on the competitiveness ofU.S. industries. The series began with Technology and Steel Industry Competitive-ness and continued with U.S. industrial Competitiveness: A Comparison of Steel,Electronics, and Aufoinobiles.

Today. the subject of international competitiveness has more visibility amongthe general public than ever before. it has emerged as one of the primary economicissues facing Congress. Debates over "reindustrialization" and "industrial policy"beginning several years ago have been renewed. This assessment continues OTA'sexploration of the meaning of industrial policy in the U.S. context, while also ex-amining the industrial policies of several of our economic rivals.

Electronics virtually defines "high technology" in the 1980's. Th;s assessmentsets the characteristics of the technology itselfa technology already of such ubiq-uity that microprocessors and computers outnumber people in the United Statesalongside other forces that exert major influences over international com-petitiveness. These factors range from human resources and costs of capital to thepriorities that corporate managers place on manufacturing Technologies and thequality of their products. The report concludes by outlining five options for a U.S.industrial policy, drawing on electronics for examples of past and prospective im-pacts, as well as on OTA's previous studies of the steel and automobile industries.

OTA is grateful for the assistance of the advisory panel for this assessment,as well as for the help provided by many individuals in other parts of the FederalGovernment. OTA assumes full responsibility for the report.

JOHN H. GIBBONSDirector

lit

Electronics Advisory Pane!

Jack C. ActonGreensburg, Pa.

Steve BeckmanAFL-CIO

A. Terry BrixTemar Ltd.

Richard P. CaseIBM Corp.

Ruth Schwartz CowanSUNY-Stony Brook

William Kay DainesAmerican Retail Federation

Leonard DietchZenith Radio Corp.

Katherine Seelman, ChairpersonNew York, N, Y.

Richard A_ KraftMatsushita Industrial Co,

E. Floyd KvammeApple. C.."61711Tater-inc.

Geraldine McArdleReston, rVa.

Charles PhippsTexas Instruments, Inc.

K. M. PooleBell Laboratories

Benjamin t`,.t RosenSevin .Rosen Management Co.

Kate WilhelmEugene, Dreg.

Isaiah Frank Robert B. WoodJohns Hopkins University International Brotherhood of Electrical

F. Willard Griffith II Workers

GC International Michael Y. Yoshinu

Robert R. Johnson Harvard Business School

Mosaic Systems Inc.

iv 7

International Competitiveness in Electronics Project Staff

Lionel S. johns, Assistant Director, OTAEnergy, Materials, and International Security Division

Peter Sharfman, Program Managerinternational Security and Commerce Program

1Tt't

Martha Caldwell HarrisRobert R. Miller*

Contributors

Robert Fisher*

Robert Rarog*

L. W. Borgman 84 Co.Princeton, N.J.

Consultant Services Institute, Inc.Livingston, N.J.

Developing World Industry &Technology, Inc,

Washington, D.C.,

Hamilton Herman, et al.Potomac, Md.H. C. Lir

Untverrsiry -6f-Ivta-ryta

OTA Publishinn Staff

OTA contract p

Phillip Otterness*

Barbara Sachs*

ContractorsMary Ann MaguireUniversity of Maryland, College Park

Mitsubishi Research Institute, Inc.Tokyo, Japan

Richard W. MoxonUniversity of Washington, SeattleSterling Hobe Corp.Washington, D.C.

John H. Wheatley, et al.University of Washington, Seattle

John C. Holmes, Publishing OfficerJohn Bergling Kathie S. Boss Debra M. Datcher Joe Henson

Glenda Lawing Linda A. Leahy Cheryl J. Manning

Contents

Chapter Page1. Part A: Summary 3

Part B: Extended Summary 232. Introduction 553. Electronics Technology 67t. trucTure and i rade in the international Electronics industry 107

5. Competitiveness in the International Electronics Industry 163

6. Manufacturing: Quality, Reliability, and Automation 2177. Financing: Its Role in Competitiveness in Electronics 2538. Human Resources: Education, Training, Management 3019. Employment Effects 343

10. National Industrial Policies 37711. U.S. Trade Policies and Their Effects 429

12. Federal Policies Affecting Electronics: Options for the United States 463Appendix A: Glossary 505Appendix B: Offshore Manufacturing 510Appendix C: Case Studies in the Development and Marketing

of Electronics Products 520Index 539

vii

CHAPTER 1

Part A: Summary

Contents

Firiciings 5

i: lli*:_li,ulliptlIlll.ti l'ilSaillli ilI illti l......:^). Lit.i.l.,:ftnlit., fliiUStr I i.1' --',II.:',. 1-.1-, L. ;p',, ,- 1 1

15

Cr)nciusif):.! IJ

2

l\lartri$();; G./11:1/11tl'iti allf1 5

'I 11fi Ii( j Ii1r)1)if 111.10.giat.lqi Cirtsuit l'ruciuct;()p. Liltv 11

t;t 1)tica!-; 72

T.)+, 1 (:(ouptit..:r ht. Value:

:1!:(1 P:\1H)ri,, 11111 1.:(plip!noW.1....11;1,-, H. I )

CHAPTER 1

Part A: Summary

Is the United States in danger of losing, tothe Japanese or others, in the race to developnew high-technolog; electronics products! ifth-generation computers, high-density inte-grated circuits, pocket televisions? Does the de-cline of the American consumer electronics in-dustry prefigure that of semiconductors orcomputers? Is U.S. standing in world marketsdeteriorating because of poor management,slipshod Government policymaking, overregu-lation of business? Will work in automobileproduction or heavy industry be permanentlyreplaced by high-technology jobs fewer in num-ber and paying wages at half the level of the$15 to $20 per hour earned by auto or steel-workers? To what extent can electronics standfor other technology-based U.S. industries?Which policies of the Federal Government areinost crucial to the international competitive-

120

100

80

60

40

20

ness of industries like electronics? Does theUnited States need a more coherent industrialpolicy?

These questions and others like them are ad-dressed in this report, which covers three por-tions of the industry: consumer products (pri-marily color television); semiconductor devicessuch as integrated circuits; and computers. Thefocus of the report is the United States, but con -side: able attention goes to the electronics in-dustries of Western Europe and Japan, as wellas several of the newly industrializing coun-tries.

Electronics in total employs more than amillion and a half Americans; 1982 sales ex-ceeded $125 billionroughly one-fifth of totalU.S. durable goods outputand have been

Sales Trends in the U.S. Electronics Market

Consumerelectronics

Components

C

Communicationsequipment

Dataprocessingequipment

Total

1- "-mt 11.. irnp

1955

IM I1960

:,...'. *414.... .°. anima.. .. m.....:drm11..."r , E.

minim . jro :. ....g 9. .17 .., . . .. -

.1 I L :'''1965

YearSOURCE: Electronics Market Data Book 1982 (Washington, D.C.: Electronic Industries Association, 1982), P. 4.

1970

12

19751-

1980

4 International Competitiveness in Electronics

growing at nearly 15 percent per year; the sec-tor is an export leader, with a surplus of about$3 billion on a total trade volume of nearly $50billion. The industry's products feed manyother portions of the U.S. economy. Not onlydoes the Nation's defense depend heavily onelectronic technologies, but both manufac-turing and service industriesranging from theproduction of numerically controlled machinetools to banking and insuranceuse electronicproducts both directly and indirectly.

The competitiveness of firms and industriesrefers to the ability of firms in one country todesign, develop, manufacture, and market theirproducts competition with firms and indus-tries in other countries. At several pointsbelow, shares of the U.S. market or of worldmarkets are used as examples of trendsin in-ternational competitiveness; in fact, however,competitiveness is a more subtle concept.While market share is one possible indicator,it is only indirectly related to competitiveness.

How an industry will fare in internationalcompetition depends on factors ranging fromtechnology itself, to industrial policies pursuedby national governments, to the human re-sourcestechnicians to upper-level manag-ersavailable in a given country. In somecases, competitiveness is primarily a functionof prices, hence manufacturing coststhem--selves determined by wage rates, labor produc-tivities, the design of both products and man-ufacturing processes. This is the case in con-sumer electronics. In higher technology por-tions of the industry, one firm may 7--e able tooffer products that are beyond the technicalcapabilities of its rivalse.g., high-density in-tegrated circuits, advanced computer software.Where this is true, costs are less important.

From the Federal perspective, shifts in theinternational competitiveness .of American in-dustries have ramifications far beyond mattersof trade balances and foreign economic policy,even military security. The competitive stand-ing of a nation's industries will determine quite

13

directly its gross domestic product, and there-fore the standard of living of its citizens.

The linkage between competitiveness andemploymentin the aggregate, in particularsectors, or in particular occupational catego-riesis much looser. Industries can rise incompetitiveness while declin ng in employ-mentthe case in the U.S. textile industry inrecent years. In other cases, competitivenessmay remain high, output may expand, but do-mestic employment may grow relatively slowlycompared to output; this has been the case inboth the U.S. semiconductor and computer in-dustries. Similarly, domestic employment isonly loosely related to trends in foreign invest-ment or to government policies directed at con-trolling flows of imported goods; trade protec-tion has helped the employment picture in theU.S. consumer electronics industry no morethan it has in the steel industry or the automo-bile industry.

While the competitiveness of a given sectorof the U.S. economy depends on both domesticand international economic forces, the domes-tic contexte.g., people and institutions here,not overseas--generally carries the most weightin determining which industries will grow incompetitiveness, which decline. As a result,public policies with domestic objectives exertthe most influence over trends in internationalcompetitiveness. These are matters of indus-trial policy. OTA uses this .term in a neutralsense to refer to tha body of regulations, laws,and other policy instruments that affect the ac-tivities of industry and the resources, includinghuman capital, that the Nation's economy de-pends on. The United States has not in the pasthad a self-conscious industrial policy, in.partbecause it had no need for one. The lesson ofthe U.S. electronics industry, along with indus -.tries like steel and automobiles that PTA hasexamined previously, is that future internation-al competitiveness may well depend on a morecoherent and consistent approach by, the Fed-eral Government to matters of industrial policy.

Ch. 1Part A: Summary 5

Principal FindingsU.S. Competitiveness in Electronics

1. 'Electronics remains a leader among Amer-ican industries. High-technology firmsinclud-ing those making microelectronic devices likeintegrated circuits and complex electronic sys-tems such as computerscontinue to be lead-ing exporters, second ,to none in technology aswell as Most measures of commercial success.Although the Nation's imports of semiconduc-tor products exceeded its exports for the firsttime in 1982 (by $160 million out of $3.8 billionin imports) more than three-quarters of theseimports were shipments by American-ownedfirms; computer exports ($9 billion in 1982) farexceed imports.

r.6

0

O

C0

4

10

8

5

This is not to say that there is little cause forconcern, or that the waves of publicity giventhe progress made by Japan's electronics man-ufacturers over the past few years have in allcases been overdramatized. If the U.S. elec-tronics industry is still strong when comparedto other domestic industries, its margins withrespect to electronics industries in other na-tions have shrunk, in some cases vanished.Moreover, the Japanese electronics industry isone of the most productive in that nation'seconomy; this high standing relative to otherdomestic sectors is a major reason for the ex-port strength of Japan's electronics manufac-turers. In almost all categories of electronicsproductsoffice copiers and typewriters, mi-

U.S. Exports and. Imports of Computers and Equipment

0 es1960 1965 1970 1975 1980 1982

Year

SOURCES: 1960.66 Gaps In rechnology: Electronic Computers (Paris: Organization for Economic Cooperation and Development.1969), p. 50,1987.81-1972, 1977, 1980, 1982 editions, U.S. Industrial Outlook, Department of Commerce.1982U.S. Department of Commerce, Bureau of industrial Economics.

6 International Competitiveness In Electronics

croelectronics, communications equipmentand consumer goodsthe U.S.-Japan trade bal-ance is strongly negative (see ch. 4).

2'. Just as the competitive positions of a na-hylastries will differ, with Some rising

and othorSdeclining, so competitive positionswithin an industry like electronics will vary.Likewise, within one portion of the industry,such as color television manufacturing, somefirms will at any given time be more competi-tive than others.

Within the U.S. electronics industry, compet-itiveness in consumer products has declinedprecipitously since the 1960's. The Nation nowimports many of its consumer electronic prod-ucts, while more than 10 foreign-owned firmsassemble and market television sets within theUnited States. In contrast, there are few signsof slackening competitiveness in the manufac-ture of computers, although the U.S., lead in

10.000

9,000

8,000

7.0000

-6 6,000l)C2 5,000

Ea) 4,000

3,000

2,000

1,000

01967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 'I'g80 1981 1982

YearSOURCES: United Statee=1987.78A Report on the U.S. Semiconductor Industry (Washington, D.C.: Department of Commerce, September 1979j, p. 39.

197740Summaty of Trade and Tariff information: Semiconductors (Washington, D.C.: U.S. International Trade Commission Publication841, Control No. 6 -5.22; July 1982), p. 28.1981, 1982-1983 U.S. Industrial Outlook (Washington, D.C.: Department of Commeice, January 1983), p. 29-7.

Japan-1967-80Japan Fact Book '80 (Tokyo: Dempa Publications, Inc., 1980), p. 188; Japan Electronics Almanac 1982 (Tokyo: Dempa Publications, Inc.,1982), pp. 149, 178.1981, 1982in-Stat Electronics Reports Feb. 21, 1983, p..5.

technology is certainly less than even half adozen years ago. American-owned firms mak-ing and selling semiconductor devices havefaced increasingly' intense competition fromJapanese manufacturers, again primarily overthe last half-dozen years; although they havelost market share both at home and abroad insome key productse.g., computer memorychipstheir overall position remains strong.

3. It is not realistic to expect that Americansemiconductor and computer firms will, in thenear future and in the absence of cataclysmicchanges in other parts of the world, return tothe preeminent positions they held at the begin-ning of the 1.970's. Nor can the United Statesexpect to achieve the technological and com-mercial leads of earlier years in other high-technology industries. The capabilities of othercountries have improved; foreign electronicsindustries have risen within their own econo-

Semiconductor Production in the United States and Japan

_la


mies; international economic conditions havechanged.

4. The United States can continue to behighly competitive in electronics and othertechnologically driven industries, with U.S.firms remaining leaders in innovation, in in-ternational trade, and in sales and profits athome and abroad. Not only is this possible, itis necessary if the United States is to maintainits standard of living, its military security, andif the U.S. economy is to provide well-payingand satisfying jobs for the Nation's labor force.Electronics is indispensable to a broad rangeof manufacturing and service functions, fromcomputer-aided design of the structures of of:fice buildings to the switching of the telephoneswithin those buildings.

5. Congress could take the initiative in devis-ing programs that would actively support theelectronics industry, and others of comparableimportance. The first requisite is broad nation-al agreement on the role of high-technologysectors like electronics as a driving force forfuture economic growth, a greater degree ofconsensus on where the U.S. economy is nowheading and where it should head. The secondis better understanding of how particularpieces of legislation affect the competitivenessof American industry, which in turn requiresdeveloping the capability of the Federal Gov-ernment for analyzing the sources of compet-itive strength.

The Role of Technology

1. One way to establish a competitive advan-tage in an industry like electronics is throughsuperior technology. Better process technology.e.g., automationcan help reduce costs. Forsimilar products, lower manufacturing costspermit lower selling prices, hence a more com-petitive product. Alternatively, higher profits'may be possible, which can help finance fur-ther improvements. Production technologiesare particularly important in consumer elec-tronics and semiconductors, less so for largecomputers.

2. Superior product technologies may com-mand premium prices in the marketplace, mak-ing manufacturing costs less significant. Prod-uct featuresranging from appearance toquantifiable characteristics such as the per-formance of a computer system in running"benchmark" programscan contribute tocompetitive advantage; in high-technologyfields as in low, product differentiation andastute marketing can be important.

Understanding customer wants and needs isvital to designing successful products; inte-grated circuits that are functionally similar,perhaps even interchangeable, may be differen-tiated through subtle variations in perform-ance; advertising strategies can be built aroundclaims of high quality or rapid delivery'; a broadarray of alternate source suppliers may re-assure prospective purchasers. Manufacturersof computers and peripherals. devote con-siderable effort to industrial design and humanfactors engineering; ease of u e is vital in sell-ing computer systems to first -time customers.

Rapid technical change reates much morescope for product technology as a competitiveweapon in microelectronics and computersthan in consumer electronics. For many years,American semiconductor and computer man-ufacturers prosper 'd by offering products thatfirms elsewhere in the world could not designor build.

Industrial Policy

1. OTA, takes industrial policy to be a neutralterm referring to the group of Federal policiesthat affect competitiveness, productivity, andeconomic efficiencysometimes directly,sometimes through influences on business de-cisions or on individuals. Industries rise andfall in international comWition for many rea-sons. Seldom can single causes be foundmoreSeldom yet simple, straightforward palicyrernedies. Plainly, industrial policy offers noquick fixes for the dilemmas of the U.S. con-sumer electronics industry, nor any sure pre-scriptions that can guarantee the future com-

,

6'


4.

\vissifieisfrodPhoto credit: Mostek Corp.

Portion of a read-only integrated circuit memory

_petitiveness of our microelectronics or comput-er sectors. Just as plainly, competitiveness inelectronicsand in other U.S high-technologyindustrieswill depend on factors including:

capable people, hence on Federal policiesdealing with education and training;capital for new business startups and forexpansion, hence on macroeconomic andtax policies;open markets for American products,hence on foreign economic policy; and.the research base that supports domesticfirms, hence on Federal technology andscience policy.

The job of industrial policy is to evaluate, link,and coordinate the many Federal efforts thatdeal with such concerns.

2. While, international competitiveness isfirmly rooted in the efforts of private com-panies, public policies set many of the rules of

the game. In the United States and in other,parts of the world, business enterprises coalpete in an environment shaped to considerableextent by government industrial policies (in-cluding elements of fiscal, monetary, tax, manpower, trade, and regulatory policies).

Foreign governments are experimenting withindustrial policies intended to aid and supporttheir own electronics industries; virtually allindustrialized and industrializing nations sin-gle out electronics for special treatment. Amer-ican firms seeking to export or to manufactureoverseas must contend with economic andsocial policies of host governments that are'more complex and sophisticated than in thepast. Rather than outright protectionism orother forms of overt discrimination against for-eign firms, host governments now adopt indirect subsidies for their own industriestax in-centives, capital allocations, funding for com-mercially oriented research and development(R&D). At the same time, governments bargainwith foreign multinationals using carrots andsticks such as investment incentives and per-formance requirements while seeking to ac-quire jobs and technology, or to improve theirbalance of payments.

3. Although a well-designed and supportiveindustrial policy is not, by itself, sufficient tobuild competitiveness in a given sector of a na-tion's economy, government policies can, un-der some circumstances, tip the balance. TheUnited States can expect no more than verylimited success in negotiations with other na-tions aimed at minimizing the impacts of thosecountries' industrial policies.

For this and other reasons, a "business-as-usual" approach is unlikely to prove sufficientto the task of maintaining U.S. competitivenessin electronics. Better prospects for strengthen-ing the U.S. position would come with theadoption of more effective industrial. policiesof our own. The American electronics industryfaces only a few major problems, mostly in thetrade arena, that are directly--susceptible toGovernment remedy. On the other hand, Fed-eral agencies could support the industrydirectly and indirectlyin many ways. Few of

17

these would have much visibility. By the sametoken, they would not necessarily cost much.Consistent and careful attention to the manysmaller matters that affect competitivenessdiffusion of technology within the UnitedStates, tax treatment of equipment contribu-tions to universities, the antitrust environmentfor joint R&D, long-term basic researcharethe necessary ingredients in a more coherentand productive industrial policy. A supercom-puter project, to take a current example, maybe glamorous as well as desirable in itself, butis no substitute.. \..

4. The choice of policy tools, and,thr designof individual measures, depend on overall ob-jectives; an industrial policy is the sum of manyparts that can be put together in different`ways.Should CongreSs wish to pursue .a more ..o-ctised industrial policy for the United States,it could choose from among five bro alter-\natives:

A protective strategy aimed at presrmvingthe domestic market base for U.S. indus-tries, along with preservation of existingjobs and job opportunities.Protection and/or support for a limitednum ibe: of industries judged Critical for theU.S. .:!,.7ratolny or. more narrowly, for na-tional semirity.Support for the technological base and inAitutional infrastructure that underlyAmerican industries, with particular atten-ticn to structural adjustwent (e.g., labor

ret;atuing and mobility).Promotion i the glolvil competitiveness of

firms inc'zYstries by encouragingt and or.,,m uwnpf:-:tition in domestic

Js int -7.:iatiDiiA morkets.2 fertwi tvhorc: r3e private sec-

tor when ..-'7cefiling industrialdevehiy-ttent an

.While these five approaches to industrial pol-i(., discussed. in chapter 12, are certainly not114.f.tually _exclusive, they represent distinctlyditerent thrusts, implying different mixes ofpolicy instruments as well as different goals.


What would be the implications of a decisionto pursue a more coherent industiial policy inthe United States? First and foremost, that toautomatically equate "industiial policy" witha greater degree of Government involvementin the economy is to view the matter from anarbitrarily narrow perspective. Industrialpolicy does not have to run counter to effortsto "get Government off the backs of business."Rather, it should be construed as an effort tomake the inevitableindeed oftentimes desir-able and necessaryFederal involvement amore consistently productive one. It implies aneffort to develop, both politically and institu-tionally, Government policies toward industrythat:

explicitly consider impacts on com-petitiveness and economic efficiency;seek to treat the problems and opportuni-ties of particular industries in the contextof the economy as a whole, rather than inisolation; and,do a better job of relating policy tools topolicy objectives.

Policy Concerns in ElectronicsAmong the elements of industrial policy, the

following are vital for the continuing com-petitiveness of the U.S. electronics industry.They might have rather different places, andbe addressed in different ways, under each ofthe alternatives listed above.

1. High-quality education and training (in-cluding retraining) for engineers, technicians,and other skilled workers.

More than anything else, the competitive po-sition of the United States in high technology,has been built on the human resources avail-able here. A renewed Federal commitment toeducation and training seems called for (seeohs. 8 and 9). Engineering enrollments runningat record levels have swamped the resourcesavailable in colleges and universities; even so,the United States graduated but 63,000 engineers in 1981 compared to 75,000 in ,Japan.

10

10 International Competitiveness in Electronics.

U.S. electronics firms have faced seriousproblems in finding adequate numbers of en-gineers, as well as technicians and service per7sonnel with needed skills and aptitudes. Inade-quaw resources in U.S. engineering schools areharming the quality of education as well as con-straining the numbers of new graduates. Train-ing and retraining for technicians and parapro-fessionals. varies widely in quality and appro-priateness to emerging needs. Many people inthe United States emerge from high schoolquite unprepared to work in technology-basedindustries.

Despite fluctuations in supply and demandover the years., engineers in principle compriseone ''of the most employable occupationalgroups in the labor force; it is hard to imaginean "oversupply" of engineers or of people withgood technical training of any of a wide vari-ety of types in an economy like that of theUnited States, provided that people are will-ing and able to shift jobs according to demandWithin the economy and organizations are will-ing to help them do so.

2. A strong technological basestemmingfrom basic research and applied R&D withlong-term objectives, including the diffusion ofresults, in-fields such as solid-state electronics,optical devices, communications technologies,computer-aided design of circuits and systems,and computer software.

The Federal .Government could not only con-tinue to fund basic research, it could establishnew mechanisms for diffusing the results ofR&D to the private sector, experiment with thesupport of commercially oriented (rather thanmilitary) research, and strengthen tax incen-tives and other encouragements for successfulinnovators.

3. Economic adjustment policies that smoothflows of capital and labor within the economy,aiding growing firms in their efforts to corn

while-

pete hile providing well-paying jobs for. thedomestic labor force.

Structural change is a fact of life in Americanindustries, driven by-the currents of an increas-ingly open international economy (see chs. 4and 5), as well as byjechnological change (ch.

3). Corporations, cities and regions. and peo--ple must adjust to changes, many of which areoutside their control. Federal attention to max-imizing the positive effectives of changee.g.,stimulating growth industrieswhile:amelio-rating the negative impacts, could be Opt1: of thecentral elements in a more coherent industrialpolicy for the United States. Policy initiativesaimed at personnel mobilitywhether \geo-graphic, inter-industry, or within organiza-tionsare one example.

4. Adequate supplies of investment capitalfor new startups as well as rapidly expandingestablished firms.

As discussed in chapter 7, venture capitalmarkets in the United States function well,although cyclic downturns are likely to recurand risk capital is often hard to find at earlystages of technology development.

Rapidly growing companies, particularly in'the semiconductor industry, do face severefinancial pressures. These stem from increas-ing capital intensity, due both to higher R&Dexpenses and to production equipment that hasgone up in cost by an order of magnitude overthe past decade, coupled with the preferenceof American managers to finance expansionfrom internally generated funds. Tax policieshave a major influence over sovirces of financ-ing and risk absorption.

While the advantages are not as great assometimes implied, large diversified electron-ics companies in Japan, and perhaps in someWestern European countries, do benefit fromreal (i.e., inflation-adjusted).costs of capital thatare somewhat lower than for merchant semi-conductor firms in the United States. By them-selves, these differencesmatters of a few per-centage pointsare not enough to weigh heavi-ly in the competitive balance. Constraints onrates of capital spendingdue in part to thepreference of American firms for internal fi-nancingare more likely to be a drag on thecompetitive abilities of U.S. manufacturers.These and other factors, primarily expectationsconcerning ;iflation, tilt the investment deci-sions of American managers toward the short-er term.

Increase in Capita! Costs for HighVolume IntegratedCircuit Production Line

S10million

1910 1975 1980 1985 1990

YearSOURCE: R. W. Broderson, "Signal Processing Using MOSVLSI Technology,

VLSI Electronics: Microstructure Science. vol. 2, N. G. Einspruch (ed.)(New York: Academic Press, 1981), p, 206

5. An international trading environment thatplaces U.S. firms on a more-or-less equalfooting with their competitors in other coun-tries, including those that have well-developedindustrial policies intended to protect or pro-mote domestic manufarturers.

As discussed in chapter 11, the frameworkfor international trade that emerged in thepostwar era is being overrun by events. The,nrust of industrial policies in many nations istoward indirect supports with effects on pricesand on competitivenes: that cannot be quan-titatively assessed (see ch. 10). Japanese in-dustrial policy, for instance, works in part bybreaking bottlenecks; the VLSI project of the1970's helped train Japanese engineers, trans-


ferred dfii;ign and processing know-how to in-dustry, rallied public support behind the struc-tural shifts th=.t were leading Japan toward an"information economy" (or at least helped dif-fuse counterpressures by those disadvantagedby such shifts). The goals of the heavily publi-cized fifth-generation computer project aresimilar. When many of the impacts of indus-trial policies are intangible, how do we coun-tervail them? Over at least the rest of the dec-ade, U.S. trade negotiators can expect to grap-ple with such issues. The prerequisite is an ana-lytical capability by the Federal Governmentadequate for understanding the ways in whichpublic policieshere and elsewhereaffect in-ternational competitiveness.

American electronics firms, particularlymanufacturers of semiconductors and com-puters, may also need the continuing supportof the U.S. Government, via both bilateral andmultilateral negotiations, in securing access onreasonable terms to foreign marketsfor ex-ports and for direct investmentif they are tomaintain their competitive position. Only bycompeting aggressively all over tho world, tak-ing advantage of scale eciro .,;,d newopportunities, can Atri4,riuun 1- .11S t.:Xlit1Ct toshare fully in the growth and expansion thatwill characterize this industry into the nextcenicry. As an example, semiconductor salesin Japal already exceed those in all of WesternEurope ',.;./ more than half; U.S. firms need ac:-cess to Japan's market comparable to that en-joyed by Japanese suppliers here.

Regardless of the overall approach and direc-tion of U.S. industrial policy, Congress couldact in support of objectives such as those out-lined above.

The Competitive Position of the U.S. Electronics IndustryConsumer Electronics

1. American firms making radios, TVs, andaudio products such as stereo receivers andtape recorders have been under severe compet-itive pressures for years; many have failed or

left the market. Few radios or black-and-whiteTVs are made in the United States. No videocassette recorders are manufactured here. Col-or television production has become largely anassembly operation, heavily dependent on im-ported componentswhether the parent firm


U.S. Sales and imports of Selected Consumer Electronic Products, 1982

U.S, sales(millions of dollars)

Imports(mAlions of dollars)

Import penetration(percent)3

Color television $4,253 $546 12.8%Blackand white TV 507 344 67.9Video cassette recorders 1,303 1,032 100.03Home and auto radiosb 1,579 1,207 76.4Stereo systemsc 1.754 1,342 76.5

$9,396 $4,471 47.6%a8ecause many Items Imported in a given year are not sold until the following year, dividing Imports during a given calendaryear by sates in that same year may give only a rough indication of Import penetration; for Instance, all video cassette recorderssold in the United States are imported even though 1982 sales figures exceed 1982 Import figures.bincluding auto tape players.

cincluding audio tape units and other component equipment.SOURCE: Electronic Market Data Book 1983 (Washington, D.C.: Electronic Industries ASsociatIon, 1983), pp. 8, 19, 31.

is American -owned (RCA, Zenith, GE) or for-eign-owned (Sony, Quasar, Magnavox). In tele-vision manufacture especially, the policies of 250

the Federal Government have contributed tothe plight of the industry. Dumping complaintsagainst importers going back to 1968 havenever bi.en fully resolved. An industry legallyentitled to trade protection has not received it.

2. Nonetheless, trade practices illegal underU.S, Inv e been only one factor in the de-clining competitiveness of the American con-sumer electronics industry. More fundamen-tally, competitive advantages have shifted toother parts of the worldfirst japan, now new-ly industrializing countries like Taiwan andSouth Korea. These countries 'have masteredthe technological requirements for mass-pro-ducing consumer products such as TV sets.They have lower labor costs than the UnitedStates, an adequate corps of skilled workersand engineers, supportive government indus-trial policies, and astute corporate manage-ments.

American firms have been reduced to a reac-tive posture; they have lost the lead in productdesign and development while moving manu-facturing operations to foreign countries inorder to keep their costs competitive. Americanproducts in consumer alectronicase.g., colortelevision receiverscontinue to he competi-tive in performance, quality, and reliability, butthey are no better than imports. The consumerelectronics market is highly price-competitive;without advantages in technology or productfeatures, . American manufacturers will behard-

0 2000II

tz, 1500)

100

50

0

Price Index for Televisions Comparedto All Consumer Durables

1973 1974 1975 1976 1977

Year

1978 1979 1980 1981

SOURCES: Consumer DurablesEconomic Report of the President 1982(Washington, D.C.: U.S. Government Printing Office, February 1982),p. 294.

TelovIslons.Electronic Market Data Book (Washington, D.C.: Elec-tronic Industries Association, 1982), p. 29.

pressed to keep up with their foreign-basedcompetitors. While U.S. firms may continue toinnovate and to be leaders in consumer prod-ucts aimed m. specialized market nichescom-puter games have been a recent examplebroadly speaking, product leadership has beenlost. At least in the short term, prospects fortaking the lead in new generations of high val-ued-added mass market products seem slim.

3. The rise of foreign firms together with pro-tracted trade disputes have contributed to a ma-jor shift .in the structure of the U.S. consumerelectronics industry. The number of firms-hasnot charged greatly since the 1960's; but whileonce there were 16 or 17 American-owned

21


manufacturers of TVs, today only 4 of 15 withplants in the United States have headquarterShere. Still, the market shares of the traditionalU.S. leadersZenith and RCAhave notchanged much; together these two companiescontinue to hold about 40 percent of the U.S.color TV market. It is the weaker AmericanmanufactUrers that have succumbed.

4. At the same time foreign enterprises wereinvesting in assembly plants in the UnitedStates, American-owned firms were transfer-ring labor-intensive manufacturing operationsto low-wage offshore locations. In general, finalassembly for the U.S. market remains here,with subassembly in Mexico and the Far East.These moves wore driven by foreign competi-tion. t;.S. color TV manufacturers felt they hadlittle option but to move production abroad ifthey were to cut costs and meet their competi-tor's prices.

Offshore production substitutes quite directlyfor jobs in the United States. Nonetheless, ifAmerican firms had not moved offshore, it isquite possible that they would have lost evenmore ground to foreign-based competition,with yet more jobs lost over the longer term.In most cases, transfers of production overseashare net impacts on U.S. employment and onthe U.S.. economy that appear relatively small;improvements in labor productivity, for ex-amplealso- driven by foreign competitionhave been at least as important as a cause ofemployment declines in television manufactur-ing. Needless to say, the impacts on individualsand communities where job losses concentrateare often severe and long-lasting; in 10 yearsthe production work force in consumer elec-tronics has been cut by more than 40 percent,from 85,000 to 50,000.

5. Beginning near the end of the 1970's, Or-derly Marketing Agreements (OMAs) limitedimports of color TVs while encouraging for-eign firms to produce here. The result was toequalize the terms of competition and to mod-erate employment declines in the UnitedStates. Otherwise, the OMAs did little to helpthe U.S. industry rebuild its competitivestrength.

In this regard, U.S. experience with OMAsrestricting color TV imports has paralleledother cases of import quotas, for instance inthe steel industry. Although the ostensible pur-pose may be to give domestic firms time to re-structure and adjust to changing competitivecircumstances, most cases protected indus-tries continue to react to pressures from abroadrather than taking strong positive steps of theirown; the notion that a respite from import com-petition will, by itself, help corporations restoretheir competitiveness gets little support fromevents in color television.

Semiconductors

1. U.S. manufacturers of semiconductorproducts such as integrated circuits remainhighly competitive in markets all over theworld. American-owned merchant firms--those that produce for the open marketar'aleaders in circuit design and process technol-ogy. While their share of world sales haschanged little over the past few years, with U.S.firms and their subsidiaries still accounting forabout 70 percent of worldwide output of inte-grated circuits,. Japarfe-Se manufacturers havebeen catching up in technology. Nonetheless,U.S. companies have the capability to maintaintheir competitiveness in most world markets.The inroads made by Japanese suppliers ofcommodity-like chips, notably random accessmemories (RA Ms); portend stronger competi-tion in other types of microelectronic devicesbut do not translate automatically into advan-tages for products such as logic chips or milcroprocessor families. There is no reason to ex-pect a loss of competitiveness in advanced mi-croelectronic products paralleling that in con-sumer electronics.

Although foreign manufacturers may some-times have advantagese.g., -supportive gov-ernment industrial policies, as in Japan orWestern Europethe U.S. merchant firmshave their own strengths. Among these are theability to rapidly develop and commercializenew technologies, to anticipate and design forshifting customer needs, and to adapt to chang-ing realities of international competition by

22


entering into joint venture and technologytransfer agreements with both domestic andforeign firms when this is advantageous.

2. The.structure of the merchant portion ofthe U.S. semiconductor industry is changing.A. number of well - established semiconductorfirms founded during the 1960's or early 1970's.have been acquired by larger, diversified enter-prises, either American- or forein-owned. Inparr, these structural shifts are associated witha trend toward aptivi production by end-prod-uct manufacturer;:.

Companies that design and build systemsranging from computers and communicationsnetworks to automobiles increasingly see needsfor internal capability in the design, develop-ment, and manufacture of state-of-the-art mi-croelectronic devices. The acquisition of mer-chant semiconductor firms by larger corpora-

Photo credit: General Motors

Microcomputer for controlling an automobile engine

Lions is a predictable trend in the, evolution ofthe industry.

3. At the same time that relatively maturecompanies like Intersilpurchased during1981 by General Electricare being acquired,new entrants continue to repopulate the mer-chant semiconductor industry. While thedownturn in venture capital markets during themiddle and late 1970's virtually halted start-ups, new firms are again being established.Since 1980, several dozen small firms produc-ing custom integrated circuits, gate arrays,specialized memory chips, ..and other nicheproducts have entered the industry. Aiming atportions of the market where the knowledgeand expertise of their founder3 can be broughtto bear, some of these start-ups will be suc-cessful and expand, some will remain small,others will be acquired by larger enterprises.

4. Captive manufacturers of semiconductor.devices make vital contributions to U.S. com-petitiveness. Such companies include IBM, thelargest producer of semiconductors in theworld, and Western Electric, which moved intothe merchant market in 1983an action madepossible by the settlement of the GoVernment'santitrust suit against AT&Tas well as a num-ber of aerospace and defense contractors. Com-panies that produce for internal use not onlyprovide a major part of the technological foun-dations for microelectronics, they spawn start-ups and give training and experience to peo-ple who later move to other companies.

5. Just as important for continuing interna-tional competitiveness are firms that design,develop, and build production equipment forapplications ranging from annealing siliconwafers to automated testing and assembly.

While the United States maintains the leadamong open-market suppliers of many typesof processing equipment, notably in lithog-raphy, other countries are catching up. Gov-ernment-sponsored R&D in Japan has focusedon production equipment.

6. R&Dparticularly that with relativelylong-term payoffswill remain a critical forcein support of U.S. semiconductor firms. In the

past, much of tho technology base has comefrom larger firms such as II3M and AT&T. Gov-ernment support for research has not been sig-nificant in recent years, although the VeryHigh-speed Integrated Circuit program of theDefense ,Department will have commercialspinoffs.

The U.S. semiconductor industry can nolonger rely on past approaches to R&D andtechnology development. The industry recog-nizes the changing situation, and is develop-ing new mechanisms for strengthening itstechnical foundations; these include closer in-teractions with universities, along with jointVentures and cooperative research efforts. Con-gress and the Federal Government could ac-tively support and encourage both basic andapplied research with longer run payoffs. Thisis one of the surest ways of supporting contin-ued U.S. competitiveness in microelectronics.

Computers

1. American manufacturers of digital com-puters have dominated world markets formany years. Much as U.S. semiconductor firmshave demonstrated the ability to rapidly capi-talize on new technological and market oppor-tunities, so have American computer firmspioneered most of the design concepts thathave driven information processing: network-ing and distributed computing, small businessmachines and minicomputers, time-sharingamong multiple work stations, cheap massstorage, desktop microcomputers.

There are few concrete signs that thisdominance by U.S.-based firms.is threatened.Nevertheless, relative positions within theworld computer' industry will continue to shift,stimulated in many cases by new applicationsof computing power. As the industry continuesto evolve, the technological leads of Americanfirms are likely to shrink, and competitive posi-tions may become more difficult to maintain.Nevertheless, the U.S. lead in worldwide mar-keting of data processing systems is so largethat prospective challengers such as Japan can-not hope for more than modest success overthe rest of the century.


2. American firms have done a much betterjob than their foreign competitors of balanc-ing what the available technology can doagainst what customers for data processing sys-tems have wanted to accomplish. This has beenan important element in Patterns of competi-tive success, which have depended as heavilyon software that could be easily used by neo-phyte purchasers and was reliablei.e., free of"bugs"as on raw hardware performance.

In fact, foreign computer firms have some-times been able to match the United States.interms of hardwam; by and large, Japan's com-puter marufactUrers can at present. But theirsystems are still behind, mostly because thesoftwareat all levels is not as good: More-over, foreign firms -- whether European or Jap-anesehave not been as adept as Americansat finding new ways to apply their hardware.For example, U.S. firms remain well ahead inoffice automation, point-of-sale terminals forretail merchandisers, and many other applica-tions of distributed intelligence.

3. The ongoing structural alterations in thedata processing industry will be deeper andfarther-reaching than those in microelectronicsor consumer electronics.

Most of the recent technological innovationsin consumer electronics have come from large,well-established firms; new products fromsmall companies have seldom reached massmarkets. In microelectronics, while start-upshave resumed in the United Statesmany striv-ing to establish themselves with the aid of in-novative productsthe path of technologicalevolution seems, for the moment, well charted;there are few signs of sudden change thatwould seriously unsettle the industry. Com-puter technology which depends on micro-electronics, but also on other feeders, primarilysoftwareis potentially more volatile. As newapplications of computing power open win-dows of opportunity for firms in many partsof the world, American manufacturers willface more intense competitive pressures. Dis-tributed intelligence will transform a broadrange of other industries as well.

29.


While the err, of the mainframe computer is ness systems, personal computers, and "smart"hardly over, the increasing importance of devices that do not even look like computet.ssinaller machinesminicomputers, small busi- will continue to provide the greatest oppor-

Market Segmentation of U.S. Computer Sales by Value

1975

Wordprocessors

10%

1985(projectec,.

1980

SOURCE: "Moving Away From Main Frames: The Large Computer Makers' Strategy for Survival," Business Week, Feb. 15, 1981, p. 78.

25

Ch. 1-Part A: Summary 17.

(unities for growth and expansion. The multi-tude of prospective applications of computingpower will offer new openings for overseasfirms as well as American companies. In someportions of the data processing equipmentiu those still in relative in-fancy. such as desktop machines and standard-ized otilc;0 automation productsforeign firmsmay eventually achieve a greater presence than

:they have managed in mainframe systems orgeneral-purpose minicomputers. To the extentthat computers become mass-market products,manufacturers in other parts of the world arelikely to emerge as more formidable competi-tors.

1%It*

N414

iro .9

)12'.etrair

4, In the computer industry, as in microelec-tronics, U.S. employment is rising much lessrapidly than output. Although new jobs are be-ing created making, operating, and maintain-ing "smart machines, other jobs are beingdestroyed; the net effects on U.S. employmentmight be positive or might be negative. Whilethere is little meaningful evidence on eitherside of the job creation/job destruction ques-tion, there is no question that skill requLementsare changing rapidly. In some cases, automa-tionaided by electronicslowers the skill re-quirements associated with the rertiaining jobs:in other cases, "upskilling" rather than "de.-skilling" results. A readily predictable conse-

Industrial robot at work

,--- Nam&

Photo credit: Unirnation

18 lrternational Competitiveness in Electronics

quence has been serious labor market disloca-tions; these seem bound to intensify. Even iflabor market shifts cannot be predicted verywell, the need for adjustment is clear. To theextent that labor market shiftsgeographical,in terms of skills, in terms of wage levelsareunexpected (and some will always be), the im-pacts will be more severe. An obvious implica-tion is that policy responses must emphasizeflexibility.

5. Japan's computer manufacturers will notbe content with narrow or specialized markets.Following strategies similar to those that havesucceeded in consumer electronics anti semi-conductors, Japanese computer firms will at-tempt to establish themselves in selected dataprocessing markets and expand from there.Backed by government efforts such as the fifth-

generation computer project, Japan's industryis bent on achieving technological and com-mercial parity (or superiority) in machinesranging from desktop processors to supercom-puters. Still, Japan's rising export strength incomputers differs in a major way from the pat-terns visible in consumer electronics or semi-conductors: the leading Japanese exporter ofcomputers, by a large margin, is IBM-japandespite the fact that it has been barred frommany of the government programs that haveaided other computer manufacturers.

While IBM has abundant resources and tech-nology to compete effectively against Japanesecomputer firms, other American manufactur-ers may face increasing difficulty in the future.AlthouE.,: the U.S. industry is not immediate-

Japanese Production, Imports, and Exports of Computersand Equipment, includingProduction and Exports of U.S.-Owned Subsidiaries

:,800

1,600

1,400

1,200

C>,) 1,000

0C

L2 800co

600

400

200

01960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980

YearSOURCES: 1987.78Japan Fact Book '80 (Tokyo: Damps Publications, Inc., 1980), pp. 173, 174.

1978-80Japan Electronics Almanac 1982 (Tokyo: Dempa PubKcalions, Inc., 1982), pp. 58, 591981, 1982 Department of Commerce, Bureau of Industrial Economics.

Ch. 1Part A: Summary a 19

ly imperiled, the Federal Government couldhelp ensure future competitiveness through abetter developed, more consistent industrialpolicy, particularly one supporting technologydevelopment and technical education.

6. As computers and their applications con-tinue to spread through the U.S, economy, theFederal Government might act to strengthenthe competitiveness of the industry both direct-ly and indirectly:

"Computer literacy"the ability to effec-Lively utilize smart machines and sys-temswill be a critical skill for the laborforce. Education and training in fieldsranging from traditional modes of quanti-tative thinking (arithmetic, algebra) to soft-ware engineering deserve renewed sup-port. Congress could provide leadership aswell as direct and tangible aid.Federal support aimed at critical bottle-necks in data processing, mostly in soft-ware, could be a vital long-term stimulusfor the American industry. Productivity in

software development has gone up onlyslowly over the years. Financing for educa-tion and training in software engineering,as well as R&D directed at computer archi-tectures, new programing languages,"andartificial intelligence appears appropriate.Smaller firms striving to establish them-selves in the data processing equipment in-dustryparticularly those developing soft-ware, peripherals, and innovative applica-tions of computing powerhave the sameneeds as do U.S. microelectronics firms:not only people with highly developedtechnical skills, but adequate supplies ofcapital for investment in R&D and produc-tion capacity and access to foreign mar-kets.

If effectively implemented, industrial policiesin support of such needs could pay vast divi-dends throughout the-U.S. economy becauseof the multitude of ways in which applicationsof computing power can enhance the compet-itiveness of firms in industries of all types.

ConclusionA nation can never be competitive in all in-

dustries at once. Not only will some rankhigher than others, but places will change overtime. Economies need to adjust; adjustmentbrings pain and distress to firms that encountertrouble, people who lose their jobs, the commu-nities affected. Even within an industry likeelectronicsin the United States, highly com-petitive as a wholesome parts, such as con-sumer electronics, face a far more problematicfuture than others. That s ch events are inevi-table does not mean that at least some of theproblems cannot be anticipated, and Some ofthe distress ameliorated by Government action.Moreover, the Federal Governthent can takepositive actions to support the develocimentand diffusion of technology, human resources,the infrastructure that companies depend onwhen pursuing' their individual competitivestrategies. Government policies can aid grow-ing sectors, help people and institutions adapt

to change. The dynamic of international com-petitiveness is continuous, and calls for a con-tinuing series of policy responses.

People can and will argue endlessly about thesuccesses and failures of industrial policies inother countries, but the primary lesson to bedrawn from foreign experience is simply this:industrial policymaking is a continuing activityof governments everywhere. In the UnitedStates, industrial policy has been left mostlyto the random play of events. Improvement isclearly possible; policymaking can be a pur-poseful activity characterized by learning frompast experience within a framework of empir7ically based analysis. Developing more effec-tive industrial policy for the United States mustbegin in this spirit, While recognizing that theprocess is inherently political. There is no onething that the Federal Government can do thatwill make a big difference for the future com-

2


petitiveness of the U.S. electronics industry,but there are many specific policy concernsthat deserve attention. Only by linking andcoordinating these more effectively can theUnited States expect to develop a coherent andforward-looking approach to industrial policy.

29

Until the Nation begins this task, Americanfirms will continue to find themselves at a dis-advantage when facing rivals based in coun-tries that have turned to industrial policies asa means of enhancing their own competitive-ness.

CHAPTER 1

Part B: Extended Summary

Contents

r,t;t:."lechnology 23(:;)n-innic; 23(;ein ic:onduct ors 24:ompnters 28

Finance 31Vt!tittn.4, Capital :31

nanc:ing Growth 32Intematior :1 I)ifferences 32.

Human Resources 34Quantity and Quality 34Continuing Education am: Training 35Preparation for Work in Technology-Based Industries 37NIIII;Welflnilt and Organization 37

Employment 39

Trade 42Antidumping Enforcement, 42The Environment for World Trade 42

U.S. Industrial Policies 46

List of TablesPage

Color TV Imports into the United States 24World Integrated Circuit Output by Headquarters Location of Producing Firms 27Predicted Growth Rates by Occupational Category in the

United States Over the 1980's 41Samico-nductor Sales in the United States, Western Europe, and Japan 44

List of FiguresPage

Japan's Production and Exports of Video Cassette Recorders (VCRs) 25.U.S. Semiconductor Sales by Typo 26U.S. Production of Computer Equipment 29Rates of Capital Spending by U.S. and Japanese Semiconductor Firms 33U.S. Employment in'Consumer Electronics 39U.S.Employment in Computer (and Peripheral EquiPment) Manufacturing 40U.S.-japan 'Trade in Integrated Circuits 44Distrilmtion of U.S. Semiconductor Sales by End Market 46

CHAPTER 1

Part B: Extended Summary

Part B of chapter 1 expands upon, withoutrepeating, the findings in the St:- .mary. In parrticular, the sections below highlight the role oftechnology as a force on competitiveness inconsumer electronics, semiconductors, andcomputers, along with factors such as capitalfor investment in research and expanded pro-duction capacity, human resources and theirdevelopment, and industrial and trade policiesboth here and abroad.

The several meanings that can be assignedto the rather amorphous concept "international

competitiveness" are discussed in detail inchapter 5. The viewpoint adopted below is firstthat of the manufacturer. Private companiesdesign, develop, produce", and market goodswhich may have more or less success in themarketplace, more or less positive impact ona nation's competitive position. Later the viewswitches to that of governments and their pol-icies, which act on competitiveness directlyand indirectly by influencing business activ-ities, supporting- education, subsidizing ex-ports, through the climate for capital formationand economic growth.

TechnologyChapter 3 covers electronics technology in

some detail (also see the Glossary, app. A, forexplanations of technical terms). Here the in-teractions between technical capabilities andmarket success are explored.

Consumer Electronics

In consumer electronic products such as coil_or TVs, both product and process technologiesare well-understood and widely fiiffused. Prod-uct differentiation strategies are more impor-tant than technical differences; componenttelevision, stereo sound, and digital chassisdesigns illustrate the frontiers of this nowiarge-ly routine field. Japan's consumer electronicsmanufacturers have benefited from the econo-mies of higher production volumes and per-haps from more extensive automation, but bothprodUct and process technologies in consumerelectronics tend to be standardized, technicalchange to be incremental. Companies any-where in the industrialized world have accessto much the same pool of knowledgethe ex-ceptions being newer product families likevideo cassette recorders (VCRs). Color TV3with similar prOduct features are made notonly in Western Europe, the United States, and

japan, but in developing countries like Taiwanand South Korea. Manufacturing technologiesare similar wherever TVs are built, with labor-intensive operations carried out in low-wagedeveloping countries by European and. Jap-anese firms, as well as American. The resultis a competitive environment in which Ameri-can firms have few unique advantages.

Differences in both product and process tech-nologies for televisions were greater during thelate 1960's and early 1970's whet Japanesefirms were beginning to invade the U.S. mar-ket. Then, firms in Japan moved more quicklythan their American counterparts toward solid-state chassis designs. By using transistors andintegrated circuits (ICs), they were able to im- .

prove the reliability of their products, and moreeasily automate portions of the productionprocess. Automation helped compensate forcomponent costs that at the time were" higherfor transistorized designS than for those rely-ing on vacuum tubes. Reliability was particu-larly important to Japanese firms because theydid not have service organizations or, dealernetworks within the United States. To increase -

their market shares, they needed to sell through.retail outlets such as discount chains. To

23


Color TV Imports Into the United States

Year

Number of color TVs impertedby origin (thousands)

Imports from all sourcesas a percentage ofU.S. consumptionJapan Taiwan Korea Total a

1967 315 318 6.7%

1969 879 22 912 15.7

1971 1,191 85 1,281 18.9

1973 1,059 325 2 1,399 15.8

1975 1,044 143 22 1,215 17.9

1977 1,975 318 92 2,476 27.0

1979 513 368 314 1,369 13.6

1981 727 514 393 1,946 15.6

aincludes imports from countries not listed individually.SOURCES: 1967. 1969Television Receivers and Certain Pads Thereof (Washington, D.C.: U.S. Tariff Commission Publica-

tion 436, November 1971), p. A62.1971.1973Television Receivers, Color and Monochrome, Assembled or Not Assembled, Finished or Not Finished,

and Subassemblies Thereof (Washington, D.C.: U.S. Interntnonal Trade Commission Publication 808, March 1977).pp. A-90, A99.

1975.79Color Telev'sion Receivers and Subassemblies Thereof (Washington, D.C.: U.S. International Trade Com-mission Publication 1068, May 1980). p. D-6.

1980Television Receiving Sets From Japan (Washington, D.C.: U.S. International Trade Commission Publication1153. June 1981), p. H.21.

1981Information from Department of Commerce

achieve credibility, they had to supply TVs thatdid not need frequent service. Japan's con-sumer electronics manufacturers succeeded inthis far from riskless strategy.

If technology is now a secondary factor forTVs, in more recently introduced product fam-iliesnot only VCRs, but video disk players,home computers, and related applications ofelectronic technologies to consumer goodsdesigns are evolving at a faster pace. Japaneseentrants spent many years and a great deal ofmoney on engineering development of VCRsMatsushita even reached production in 1973with a design that was shortly thereafter judgednot to be good enoughbefore achieving com-mercially viable products. But otherwise, com-petition in consumer electronics is largely amatter. of prices and marketing, brand loyaltiesand customer perceptions. While Japanese ex-porters have established themselves firmly inAmerican markets for TVs and audio products,individual companies have suffered frequentreverses in consumer goods ranging fromStereo receivers to CB radios and pocket cal-culators, where markets have been unpredict-able and competition always stiff.

SemiconductorsMicroelectronic devices, in contrast, are in-

termediate products sold in accordance withdetailed technical specifications to sophisti-cated customers who design them into finalproducts ranging from TVs and electronicgames to missile guidance systems and power-ful computers. To be successful, semiconduc-tor firms must not only meet the current re-quirements of such customers but do a goodjob of anticipating their future needs.

Technological Factors in Competition

As explained in chapter 3, the interdepend-ence of product and process technologies inleading-edge microelectronic devicesverylarge-scale ICsis unusual even for a high-tech-nology industry. Circuit designers must under-stand the nature and capabilities of the fabrica-tion processincluding proprietary detailsto optimize the performance of a chip. Productand process technologies advance together,with process capability a restriction on devicesthat can be fabricated with acceptable yields(the fraction of circuits that function). The in-

33

12

I

10

7

5

5

3

2

Japan's Production and Exportsof Vireo Cassette Recorders (VCRs)

1975 1976 1977 1978 1979 1980 1981 1982

Year

Ch. 1Part'8: Extended Summary 25

SOURCE "VTR Production Demand," Japan Report. Joint Publications ResearchService JPRS 011100. Jan. 28, 1983. P. 35

teractions go in both directions. Clever circuitdesign can compensate for some kinds cf proc-ess limitations. Among the examples are simplydoing more with fewer transistors or other cir-cuit elements and incorporating on-chip testingand redundancy. Some American firms addedredundant circuit elements to their 64K RAM(random access memory) designs, a step whichmay 'pay dividends in the future as they moveto still higher levels of integration.

Competition in standardized products likeRAMs depends on both price and technologychapter 5. When 64K RAMs were first intro-duced, they sold in sample quantities for about$100 each. From this level in early 1980, pricesfell to $10 to $15 by the end of that year. Afteranother year, 64K RAMs could be purchasedfor $5 or less. These rapid price declines, typi-

r),)- 1 1 1 - fl z -

cal of the semiconductor industry, are drivenby intense competition to improve processyields, reduce manufacturing costs, and cutprices to build market share. As the prices of64K RAMs dropped, prices also fell for the pre-vious-generation 16K devices, which by 1982sold for about $1 each. Similar patterns will befollowed as 256K RAMs, in pilot productionin both japan and the United States during1983, take over from 64K chips.

Despite the intense price competition inthese commodity-like circuits, product technol-ogy continues to play a role. Not only is a goodcircuit design essential for high yields and lowcosts, but a high-performance RAM can corn-mend a greater price. While the most commonvarieties of 64K RAMs have access times (theaverage time to retrieve the contents of a mem-ory cell) in the range of 200 nanoseconds (200X 10 seconds), otherwise comparable cir-cuits with lower access times sell for more;during 1982, 64K RAMs with access times of150 nanoseconds brought prices a dollar orso above those for 200 nanosecond circuits..Nonetheless, RAMsand most other memorychipsare in essence standardized items. Asfor consumer products like TVs, progress is in-cremental and predictable, at least at presentalthough the pace is much swirder.

If process technology is vital for RAMs, prod-uct technologyi.e., circuit designcarriesgreater weight in competition invol. ;` uthervarieties of ICs. Foreign firms have Len lesssuccessful in microprocessor families and thearrays of support chips designed to be usedwith the processors themselves, as well as sometypes of linear circuits, logic families, semicus-tom chips, interface circuits, and the manyother varieties of specialized microelectronicproducts. In contrast to memory chipsinessence "brute force" devicescircuits that im-plement Ionic depend more heavily on creativeengineering design, on anticipating user needs,and on recognizing new opportunities madepossible by developments in either process ordevice technology. A well-designed micro-processorone with an architecture that takes,maximal advantage of the circuit elements itemploys, with an instruction set that pro-

3,4


Linearand

otherICs

Memories

ICs

U.S. Semiconductor Sales by TypeLinear

and(tilerICs

Microrocessorsimicrocomovre.rs

ICs Microprocessors!microcomputers

1980

Standard logicfamilies

1975

Standardlogic families

SOURCE: 1975 El ec !tonics, Jan. 8, 1978, pp. 92, 93.1986 ElecfronIcs, Jan. 13, 1982, pp. 124, 125.1946 El e c t ro nics, Jan. 13, 1983, pp. 128, 129; Mar, 10, 1983, p. 8

gramers find easy to use, a convenient busstructure and input/output portscould be acommercial success even if developed by acompany with only mediocre process technol-ogy. Were this the 'case, however, alternatesource manufacturers might end up with moreof the market ;' ad/or higher profits.

International Positions inMicroelectronics Technology

While Japanese manufacturers now makeand sell many types of microprocessors andlogic circuitry, and have always had excellent

35

LinearandotherICs Standard logrc

families

ICsMemories

1986(Projected)

Microprocessors/microcomputers

technology for linear ICs, they have not beenable to match American semiconductor firmsin design-intensive products. For instance, themicroprocessors that Japanese semiconductorfirms se'l in large volume on the world marketare U.S. designs. Such patterns will probablycontinue to hold, although here as elsewherethe magnitude of the U.S. lead is likely to shrinkas the Japanese get better at circuit design, andas Japanese semiconductor manufacturers hireengineers from other countries.

In semiconductor processing, Japanese firmsare often on a par with the United States and

Ch. 1Part B: Extended Summary 27

may. be better in some cases. One reason hasbeen the VLSI research project and its severalfollow-ons, orchestrated and partially fundedby Japan's Government. By 1983, Japanesemanufacturers were, as a group, further alongin production plans for process-intensive 256KRAMs than their American competitors. Proc-ess control also exerts a major influence overquality: nevertheless, if a few years ago thequality of some types of Japanese ICsspecif-ically, RAM chipswas higher than suppliedby American firms, today any differences aremuch smaller (see ch. 6).

Semiconductor manufacturers in japan havemade great strides as well with complementaryMOS (metal oxide semiconductor) circuitry,one reason being its attractions for.certain ofthe consumer applications in which Japanesesemiconductor firms for many sqrs spe-cialized. In contrast, companies in WesternEurope are generally behind the UnitedStates and Japan in all varieties a MOS. Euro-pean nations are making determined efforts tocatch up, in several cases with strong govern-ment support. Despite underlying technologi-cal abilities that in many cases are excellent,European manufacturers have not been as suc-cessful as American suppliers at convertingtheir technology into successful commercialproducts. In circuit design, neither the Jap-anese nor the Europeans seem able to matchwits with Americans. This is an advantageasource of "technology gap" that the UnitedStates should be able to maintain. To do so,

U.S. firms must continue to vigorously pursuenew markets and American engineeringschools must retain their preeminent positionin fields related to microelectronics.

Research and DevelopmentDespite the continued prowess of American

circuit designers, the comfortable lead once en-joyed by the United States in the underlyingtechnology of semiconductor devices is nowspotty at best. American merchant semicon-ductor firms devote most of their R&D effortsto product and process developments with im-mediate application to end-products; relative-ly small companies with limited resources, theyhave had little choice but to place the greatestPriorities on R&D that will help them in nexty Ear's marketplace battles.

In the United States, more basic researchranging from studies, of the physics of electrondevices to the development of process toolssuch as ion-beam lithographyhas beenfunded and performed elsewhere. Some of thework has been supported by the Departmentof Defensee.g., research on high-speed gal-lium arsenide devices. In other cases, large or-ganizations such as IBM or AT&T's Bell Labo-ratories have carried much of the burden; BellLabs, in particular, has been responsible formany of the seminal developments in solid-state electronics. In the past, Bell diffusedthese widely to both U.S. and foreign enter-prises. Now, with AT&T entering new markets,

World integrated Circuit Output by Headquarters Location of Producing Firms

1978 1982a

Production(millions of dollars)

Share ofworld output

Production(millions of dollars)

Share ofworld output

United States $4,582 68.3% $ 9,700 69.5%Merchant 3,238 6,450Captive 1,344 3,250Captive percentage 29.3% 33.5%

Western Europe 453 6.7 620. 4.4Japan 1,195 17.8 3,440 24.7Rest of the worldb. 482 7.2 190 1.4

$6,712 $13,950aEstimated.

Includes the Soviet Union and Eastern Europe for 1978 but not 1982.

SOURCES: 1978Status '80: A Report on the Integrated Circuit industry (Scottsdale, Ariz.: Integrated Circuit Engineering Corp.. 1980). p. 4.1982Status 1982: A Report on the Integrated Circuit industry (Scottsdale, Ariz.: Integrated Circuit Engineering Corp.. 1982), p. 5.


including merchant semiconductor sales, andcompeting under new conditions, the companymay no longer feel that it has the luxury of_sup---porting basic research so heavily; at the least,it will guard its technology_rnua more close-ly (as IBM always_has). Other forces at workinclude growing software demands on micro-electronics firmsan area constrained by per-sonnel shortages, low productivity, and weaktheoretical foundations. Furthermore, thehighly competitive merchant firms have per-haps been taking advantage of new technolog-ical opportunities faster than the stockpile hasbeen replenished. The need for new sourcesand mechanisms of technology developmentand diffusion is plain.

Along with continued Federal support andincentives for R&D, particularly more basicwork, institutional innovations that would helpto build the technological.base for continuingU.S. competitiveness in microelectronicsaswell as in computer systemsappear worthyof congressional attention. U.S. competitive-ness in electronics has dep3nded heavily on thetechnical strengths of American firms. So longas the United States held a substantial overalllead in electronics technology, smaller com-panies could successfully design and developtheir products and processes without doingmuch research on their own. The foundationprovided by large companies, military spend-ing, and the universities sufficed. Today, notonly is this base eroding, but the overall tech-nical edge of the Nation has diminished. Inparticular, research capabilities in Americanuniversities have deteriorated because of obso-lete equipment and shortages of graduate stu-dents and faculty. A redefined Federal role inR&D could address the need for better mech-anisms of technology diffusion within theUnited States, as well as encouraging inflowsof technology from overseas.

A number of promising models begin-ning with dothestic ventures such as the Semi-conductor Research Cooperative and Micro-electronics & Computer Technology Corp. andincluding a number of experiments in othercountries. Some of these are aimed at enhanc-ing the diffusion of technology as well as at en-

couraging basic and applied research with po-tential commercial, rather than exclusively mil-

applications. The Fraunhofer Gesell-schaft in West Germany, as well as Japan's jointR&D programs, both discussed in chapter 10,come to mind. The U.S. electronics industry,including but not restricted to microelectron-ics, could benefit from institutional mechan-isms more closely linking R&D efforts in Gov-ernment laboratories, industry, and universi-ties. A relatively large but decentralized sys-tem of centers-of-excellence, directed towardcommercial developmentswith ample scopefor local funding and entrepreneurial partici-pationwould fit American traditions. Givensome fraction of funding, perhaps 30 or 40 per-cent, from the Federal Governmentron a con-tinuing basis, the time horizons could be longerthan those for R&D programs funded entirelyby industry.

ComputersIf manufacturing technology is critical for

cost control in consumer electronics, and bothprocess and product technologies are vital insemiconductors, the computer industry exem-plifies reliance on product technologies. Par-ticularly for larger systems, manufacturing isless significant for competitiveness becauseproduction volumes are modest compared toTVs or semiconductor devices. For small com-puters sold in large numbersand particular-ly the desktop machines offered by companieslike Appleor for peripherals such as printersand terminals, manufacturing technologies areof greater and growing importance.

Technological CompetitionWhat are the major factors in marketing com-

puters? First and foremost, performance/costratio: the computing power per dollar that amanufacturer can supply. This depends heavilyon system designboth hardware and soft-warei.e., in doing more with less rather thancutting production costs. For most computersystems, assembly is labor-intensive, costs in-creasing with overall complexity. The com-pany that can design a system offering higher


performance at a given cost has the advantage.IBM, as in so many ether instances in the com-puter industry, provides something of RTh ex-ception because its higher sales volumes meanmore pronounced scale economies. A furtherexception has developed at the lower end ofthe market, where personal machines, smallbusiness systems, and micro or minicomputerssold to original equipment manufacturers arebuilt in much greater numbers. In both cases,greater production volumes increase the sig-nificance of manufacturing technologies but inno way diminish the role of system design.

Because of these characteristics, the comput-er industry is just as design- and R&D-intensiveas microelectronics, but computer engineersare seldom as constrained by manufacturingprocesses as chip designers. They are, how-ever, constrained by the performance charac-teristics of available components, principallyICs. Microelectronic devices are the buildingblocks for processors as well as essential ele-ments in many other parts of computing sys-tems, from controllers for disk drives to semi-conductor memories themselves. Because sys-tem performance depends so heavily on ICs,many computer firms have established captivemicroelectronics R&D and production facili-ties. While component technologies ultimate-ly limit what can be done, computer designershave considerable latitude in configuring sys-tems; the many alternatives from which theycan choose are affected in different ways bythe characteristics ol both hardware and soft-ware.

Systems AspectsAlthough firms located in other countries are

nibbling at U.S. market share, our dominancein computer manufacture still continues, builtlargely on the abilities of American producersat system design and integration. Conceivingand developing new applications of computingpower depends on engineering design and onunderstanding user needsincluding fieldservice and software support. American manu-facturers opened the personal computer mar-ket, not through technical advances, butbecause they perceived a potential market

U.S. Production of Computer Equipment

1970 1975 1980

YearSOURCE: 1972, 1975, 1977, 1980. 1983 editions, US. Industrial Outlook, Depart-

ment of Commerce. 1981 and 1982 shipments estimated.

where others did not. Substantial penetrationby Japanese imports may eventually-follow, butbased more on low prices-:-stemming from thewell-established capability of Japanese elec-tronics firms to manufacture in high volumeat low costthan unique product features.Nevertheless, so long as technical evolution israpid, and software one of the keys to sales,American entrants with creative product de-signs should have little to fear from overseascompetitors. At least at first, the more suc-cessful Japanese personal computers will bebased on U.S.-designed microprocessor or mi-crocomputer chips, as well as software devel-oped in the United Statese.g., the popularCP/M or Unix® operating systems and themany applications programs that run on them.

This is only one example where Americancomputer manufacturers have been at the fore-front in spotting new applications of comput-ing technology. Among the other examples are:

Small- machines suited to the needs ofbusinesses with a few dozen to a few hun-dred employees.Fault-tolerant systems that can be usedwhere reliability is critical.Specialized data processing installationsfor banks, insurance companies, and Gov-ernment agencies.Dedicated processors to be integrated into

30


industrial controllers, scientific instru-ments, aircraft flight-control systems.Networked systems, time-sharing, satelliteterminals and other mechanisms for pro-viding users with computing power whenand where needed.Both large and small machines for special-ized scientific and technical computing,ranging from supercomputers for complexnumerical calculations in computationalfluid dynamics or the development of nu-clear weapons to array processors to beused in conjunction with dedicated mini-computers in modeling chemical reac-tions.

Sometimes market demand has driven the tech-nology, with the design efforts of computermanufacturers shaped by perceptions of theseneeds; occasionally, more raw computingpower has been available than has found im-mediate application.

System integration remains the forte ofAmerican firms, andjust as for U.S. semicon-ductor manufacturersso long as Americancompanies and American engineers continueto push aggressively into new software andhardware applications, they should be able tomaintain a technological edge sufficient to holda large fraction of the world computer market.

4ig

Experimental electronic map display for automobile dashboard

39


Nonetheless, this share may not be the 70 per-cent of 10 years ago. If so, the causes will bemultiple. as discussed in chapter 5. Rapidly ex-panding and fragmenting markets mean thatno one manufacturer can cover all the bases;windows of opportunity will open for foreignas well as U.S. suppliers.. Manufacturers inother countries may benefit from supportiveindustrial policies, as well as drawing on grow-ing pools of capable computer scientists, sys-tems engineers, and managers. The result islikely to he a narrowing of the technology gaps

Ch. 1--Part B: Extended Summary 31

that have favored American firms. Improve-ments in the standing of computer industriesin countries like Japan relative to other sectorsof these country's economies may lead togreater international competitiveness in com-puter manufacturing. Most of these forces areoutside U.S. control. Given the circumstances,it becomes particularly important that the Na-tion avoid unnecessary saciifices in competi-tiveness through missed opportunities byAmerican firms or defective policy choices bythe Federal Government.

FinanceWell-developed r:apital markets have been a

major source of strength for entrepreneurialhigh-technology firms in tha United States.Under most circumstances, both new start-upsand young, rapidly expanding companies havebeen able to find the money needed to grows.vith their markets. Still, this has not beenuniversally true; in recent years, some elec-tronics companiespreferring, in commonwith most of American industry, to fund ex-pansion with internally generated cash flowshave found themselves lacking the financingneeded to keep pace with market opportunities.Perhaps of greatest significance, volatile inter-ost rates in the united States reinforce otherfactors that bias decisionmaking by corporatemanagers toward short-term undertakings.

Venture Capital

Over the past quarter century, venture capitalin its various forms has spawned many of thenew entrants in the U.S. electronics industry:companies supplying softwar', instrumenta-

on, semiconductor devices, computers andperipherals. Some of theSeDigital EquipmentCorp., Intel have become mainstays of U.S.competitiveness. Other nationsWest Germany, the United Kingdom, even Japanhavernight to build some of the characteristics ofJ.S. venture capital markets into their indus-

trial policies. These attempts to generate thevitality and dymirnism that venture start-ups

have brought to American growth' industrieshave seldom met with success.

Bottlenecks in y'.s. capital markets are moreprobable and more significant when it comesto financing rapid 'expansion in sectors likemicroelectronics, where capital intensity isescalating along with sales, than in fundingnew ventures. Nevertheless, venture fundinghas not always been available, particularly seedcapital for developing new ideas well beforeproduction is in sight. When venture fundsdried up in the middle 1970's, new start-upsin electronics manufacture virtually halted.Around the turn of the decade, after the reduc-tion in capital gains taxes that took effect in1978one of many forces affecting venturecapital suppliesthe market revived. Most ofthose supplying venture funds look for capitalappreciation over a 3- to 5-year period, withtypical target returns being 35 to 50 percent peryear. Plainly, capital gains tax rates' are impor-tant both to individual and institutional sup-pliers of risk financing. However, for reasonsthat are poorly understood, venture capital,funding is notoriously cyclical: factors otherthan tax changes, also contributed to the revivalof the market. By mid-1980, a veritable boomin venture funding was underway, with muchof the money going to electronics. Prospectiveentrepreneurs, many in the Silicon Valley re-gion of California, saw opportunities in micro-computers and other applications of micro-processors, in software and computer periph-


erals, in semiconductor chips themselves. Cap-italiSts saw the technological windows in muchthe same light. More than 20 new microelec-tronics firms alone were established during thefirst two years of the venture capital resur-gence.

Financing Growth

As chapter 7 points out, finding capital forcontinued expansion has been a greater con-cern for many U.S. electronics companies, par-ticularly given the high growth rates in muchof the industry. While there are few if any signscf overall shortages of capital for investmentiii`the United States, financing growth is a com-mon problem for young companies anywh..rein the economy. ElectronicS firms, especiallythose producing ICs, face unusually steep hur-dles. The first is simply th9 need to keep upwith markets that in some years have grownat 25 percent or more. Firms trying to increasetheir shares of such markets have to add pro-duction capacity at rates that can severelystrain financial resources; needless to say, theinvestments must precede the added revenuesthey bring in. At times, U.S. semiconductormanufacturers may have been unable to securethe funds needed to keep up with marketgrowthor, more likely, have judged the con-ditions imposed by prospective suppliers ofcapital unacceptable.

_

Rising capital intensity for semiconchctorprocessing creates a second hurdle. Denser ICsrequire more expensive fabrication equipment.A state-of-the-art manufacturing facility, whichcost perhaps $5 million a decade ago, nowmight run $50 million. High levels of R&Dspending, mandatory for companies that hopeto compete in markets for advanced devices,.contribute a third hurdle. Thus capital demandis mounting even more rapidly than the markethas been growing, compounding the alreadydifficult financing problems of U.S. semicon-ductor firms.

In common with most of American industry,U.S. electronics firms have been reluctant torely 'heavily on external fundseither debt(loans, bonds) or equityfor financing growth.

At times over the past decade, it would havebeen difficult to issue either bonds or stock.Nonetheless, the U.S. electronics industry ex-hibits a pattern of consistently low debt/equity ratios contrasting sharply with foreign man-ufacturers. Aversion to borrowing may haveconstrained the growth of some American elec-tronics companies over the past decade.

The changes in U.S. `Ax law implemented bythe Economic Recovery Tax Act Of 1981 haveincreased cash flows for electronics firmsalong with other businesses in the economy.High-technology electronics manufacturersbenefit particularly from the R&D tax credit.Accelerated depreciation is a different matter:although more rapid capital cost writeoffs arenow available to virtually all U.S. corporations,the benefits are much greater for numerousother sectors. Because electronics firms, par-ticularly in the high-technology portions of theindustry, have always been able to,depreciateat fairly rapid rates, they have not been helpedas much as sectors like primary metals. Inearlier years, many such industries facedlengthy capital cost recovery periods. The re/-ative position of electronics has suffered underthe 1981 Tax Act to the extent that companiesin other lines of business have an easier timesecuring external funds.

International DifferencesWhy do American companies limit their use

of external funds? Most managers would an-swer by citing high costs of capital, whetherdebt or equity, as reflected in high U.S. interestrates over the past few years. American busi-nessmen have claimed that they face costs ofcapital perhaps twice those of their'competitorsin Japan. In fact, although costs of funds in theUnited States are higher than in Japan, thedifferenceswhen adjusted for inflationary ex-pectationsappear relatively small, certainlyless than 5 percentage points. While not insig-nificant, the resulting advantages for Japaneseelectronics manufacturers are hardly over-whelming, even when the tax benefits of thehigher debt/equity ratios characteristic of Jap-anese corporations are taken into account.Lower costs of capital in Japan make no more

Ch. 1Part B: Extended Summary '13

than minor contributions to differences inmanufacturing costs. A much more potentsource of advantage for large, diversified Jap-anese electronics firms, particularly in periodswhen markets are expanding rapidly, stemsfrom their ability to allocate funds internally,using moneys generated in other lines of busi-ness to finance high rates of spending on R&Dand new production capacity. U.S. semicon-ductor firms, especially those that remain in-dependent and have a limited range of prod-ucts, will always be hard-pressed to keep upwith diversified companies, Japanese or Amer-ican. A major difference between diversified.Japanese and American electronics firms is theevident willingness of Japanese semiconduc-tor manufacturers to compete in the mass mar-

0)a)

0Ca)2a)a.

C)C

a) 10a.cn

7:3-c,0

Rates of Capital Spending by 11.S.

ket for merchant products, and to aggressive-ly add new production capacity. It remains tobe seen how American firms like Mostek or In-tersil, which are now parts of large conglom-erates, will behave over the longer runandhow Western Electric will fare, now that it isentering the merchant market.

While the contrasts between financing prac-tices of American and Japanese electronics cor-porations are manyas are those with Euro-pean enterprisesthe net advantages thatJapan's companies receive from governmentguidance applied to investment funds aresmall. Japanese industrial policies continue toinfluence capital allocations and costs of funds,but the high leverage characteristic of Japanese

and Japanese SemIconduOtor Firms.

1973 1974 1975 1976 1977

Year

1978 1979

aintegrated circuits only, 1973.1979; weighted average of 12 manufacturers 1973-1979, 11 manufacturers 1980, 1981.b1973-1980. Weighted averages for 11 U.S. merchant semiconductor manufacturers; 1981 estimated:

SOURCE: Unitsd States-1973-1977, Bureau of Census; 1978-1981, Department of Commerce and Semiconductor Industry Association.Japan-1973-1979, Japan Fact Zook '80 (Tokyo: Dempa Publications, 1980), p. 203; 1980, 1981, Japan Economic Journal,

42

1980 1981

34 International.Competitiveness in Electronics

corporations, as meLaured by the ratio of debtto equity or debt to total capital, helps primarilyin terms of taxation. In Japan as in the UnitedStates, interest can be written off as an expense(while dividends paid to shareholders cannot);therefore higher proportions of debt reducecorporate tax bills. That banks in Japan lendwillingly to highly leveraged firms places thesebanks in a position more like that of equity-holders in the United States: Japanese banksabsorb higher risks than American banks, butthe impacts of this, by itself, on the competi-tiveness of Japanese companies are small. Fur-thermore, leverage ratios of Japanese firmshave been slowly declining over the yearsone example among many of the gradual move-ment of the Japanese economic system towardconvergence with other advanced nations.Likewise, the unusually high rate of personalsavings in Japan has impacts at the aggregatelevel which are only loosely coupled with costsor capital for individual firms. These costs varywidely across the Japanese erktronics indus-try, just as in the United States. Indeed, costand availability of capital differ more from firmto firm within the U.S. electronics industry

than, on the average, between the electronicsindustries of the United States and Japan.

The apparently high costs of capital in theUnited Statesas reflected in high interest,rates, stemming in the past from expectationsof continued price inflationdo have a seriousconsequence: High and uncertain interest ratesin the United States tend to skew investmentdecisions toward short-term undertakings.Although no one knows how to measure or ag-gregate the time horizons of business execu-tives in any meaningful waymuch less com-pare those of American executives with theircounterparts in West Germany or Japanallelse the same, interest levels that fluctuate un-predictably will act to shoe ten time horizons.Investments with longer payback periodsforexample in basic research or in advanced pro-dudtion equipment will appear less attractive.To the extent that capital markets in the UnitedStates continue to mirror expectations of highand uncertain interest rates, the future compet-itiveness of American industries like electron-ics may suffer.

Human Resources

Business enterprises depend on capable peo-ple for tasks ranging from assembly line workto service and repair of their products, designand development, and general management.From the standpoint of international compet-itiveness, the larger the pool of qualified peo-ple a firm or an industry can draw from, thebetter. An ample supply of engineers and tech-nicians means that companies will have theluxury of picking and choosing, while from theemployee viewpoint, salaries may be de-pressed. A small pool means potential short-ages, most likely of specialists, perhaps driv-ing organizations to move people laterally tomeet their needs. Soaring demand for comput-er professionals, for example, has drawn inmany people without formal training in thediscipline; about two-thirds of those employed

in programing and related jobs have degreesin other fields.

Quantity and Quality

For several years, during which engineeringgraduates in all disciplines were in short sup-ply, the U.S. electronics industry experienceda scarcity of entry-level electrical engineers andcomputer scientists. In the short term, demandhas droppedlargely because of recessionwhile the supply continues to rise, fed byswollen undergraduate engineering enroll-ments. The longer term pictureincludingprospects of continuing shortages of softwareengineers, integrated circuit designers, andothers with specialized skillshas not changed.Moreorier, the supply of grey-collar workers for

4d:

Ch, 1-:-Part B: Extended Summary 35

the electronics industrytechnicians, drafters,and designers, field Service repairmen, labo-ratory aidesmay also be short, although quan-titative data on supply and demand for suchjobs are scarce. There is, needless to say, noshortage of unskilled assembly workers; theheart of the problem in this, as in a number ofother American industries, is an excess of un-skilled workers coupled with sporadic short-ages of those with higher levels of educationand training.

The scarcity of recent U.S. graduates in en-gineering has been real, extending to virtuallyall specialties. Its origins lie in low enrollmentsduring the early and middle 1970's (see ch. 8).Since then, engineering enrollments have re-bounded to record levels. Educational re-sources have not kept pace, with the result thata substantial number of engiheering schoolshave had to limit the numbers of students ad-Milted. Not only is supply constricted, but thequality of engineering education is suffering.

Shortages of teaching faculty have con-strained engineering education more than anyother factor. Despite undergraduate enroll-ments that have nearly doubled over the pastdecade, trends in graduate engineering studyhave been nearly flat. In particular, studentshave been reluctant to enter doctoral programs.Fewer Ph. D.'s in engineering were graduatedin 1982 than in 1972. Nearly half those now re-ceiving Ph. D.'s from American engineeringschools are foreign nationals. Recent Ph. D.'shave been avoiding teaching careers, for whichthe doctoral degree is today virtually manda-tory. Not only are salaries low relative to in-dustry, but new teachers can anticipate heavycourse loads as a result of high undergraduateenrollments and the faculty shortages thatalready exist. Coupled with uncertain pros-pects for research support and a lack,of pros-pective graduate students of their own, univer-sity teaching is no longer an attractive prospectto many who in earlier years would have beenprime candidates. The result is 1,400 to 2,000unfilled vacancies onthe faculties of U.S. en-gineering schools.

Deteriorating laboratory facilities create asecond bottleneck. Engineering education isexpensive; curricula include numerous labora-tory courses, as well as heavy use of computingfacilities. Keeping laboratories relatively cur-rent, so that students get some experience withup-to-date equipmentinstrumentation, smallcomputers, applications of microprocessorsisa long-standing problem that has grown_m_qrsein recent years.

If the trends outlined above continue, seriousharm to the competitive prospects of , manyAmerican industries could result.

Continuing Education and Training

In contrast to constraints on supplies of newengineering graduates, the United States hashundreds of thousands of mideareer engineersalready in the labor market. If the half-life of

-a college education in engineering is, say, 10years, upgrading these peoples' skills offersvast opportunities both for individuals and forU.S. industry. In some cases periodic shortcourses or self-study may be enough to boostpeople along chosen career paths; in others,they may wish to move laterallye.g., from an-alog to digital circuit design, from hardwaredesign to software. As pointed out in chapter8, little data exists on the frequency with whichengineers take advantage of opportunities forcontinuing education and training; it appearsthat most who do are recent graduates, and thatthose with the 'greatest needi.e., people "10years or more out of schoolrarely pursue con.tinuing education beyond the occasional (andseldom very challenging) short course. Severalimplications follow: 1) the rewards of .pursu-ing continuing education and training ir en-gineering could be lowe.g.,..employers maynot support such activities extensively, prefer-ring to hire new graduates with the skills theyneed at lower salaries; 2) the programs avail-able may not be attractivei.e., working engi-neers may perceive them as academic and un-related to their jobs; 3) the quality of programsmay vary quite widely, so that those who have

.

4 4


el

or hear about bad experiences are reluctant totry again.

The paucity of information on this subjectis in itself disconcerting, but it appears thatall of these factors are at work, and others aswell. Certainly, existing incentives seem highenough to motivate only those with unusualability or perseverance. While some companieshave devised effective programs for encourag-ing employees to maintain and improve theirskills, others do little or nothing. The pictureis likewise mixed among educational institu-tions; some engineering schools have devel-oped aggressive outreach programs aimed atproviding high-quality coursework for tech-nical professionals, those who are seeking ad-vanced degrees and those who are not. Contin-uing education programs offered by profes-sional societies as well as profitmaking orga-nizations vary considerably in quality. Thequickest, surest way of providing the numbersof qualified engineers needed to maintain thecompetitiveness of American industries likeelectronics is to make high-quality continuingengineering education more widely availableand attractive to midcareer professionals.

Despite recent difficulties, engineering edu-cation in American universities remains

andabest in the world. In part because schools nduniversities in some countries do relativelypoorly at preparing their graduates for careersin industry, foreign companies resort more fre-quently than U.S. firms to internal and on-the-job training. Extensive company-run trainingprograms are prominent in the Japanese elk-tronics industry, where continuing educationis widespread among blue-collar and grey-col-lar employees as well as white-collar profes-sionals. One way for the United States to in-crease its pool of skilled grey-collar workerswould again be to develop a more effective ap-proach to continuing education and training.Such programs will be more effective _whereclosely coupled with prospects for upward mo-bility within organizations. At present, theprobability that an unskilled worker in an elec-tronics firm will be able to move tip to a higherpaid position is small.

More broadly, programs of all types aimedat vocational education in technical fields ap-pear to need reexamination and modificationif the quantity and quality of graduates is togrow.. In the United-States, as many as-8;000public and private schools offer vocational-technical education and training (comparedwith roughly 300 engineering colleges). Thequality of the courses and programs offered bythese institutions varies widely. Activities arefragmented, with little detailed informationavailable even to form a baseline for analysis.One point is clear: the fraction of the laborforce in U.S. manufacturing industries withformal training in technical fields (through ap-pi'enticeship programs or schooling) and/orcredentials (e.g., certification granted after ex-aminations) is far lower than in a number. ofother industrialized nations, including WestGermany and Japan. While correlations with.on-the-job ability may be imperfect, the preva-lence of such programs in other countries isgood evidence of a commitment by' individuals,governments, and business enterprises to build-ing a labor force that will help maintain thecompetitive ability of technologically based in-dustries into the future. So far this commitmenthas been lacking in the United States.

Congressional 1 o,atiership could have a ma-jor impact. As the pace of technical advancein industries like electronics continues or accel-erates, workers at all levels will face new de-mands on their capabilities. Given the increas-ing disjunction between the skills of the U.S.labor forcewhat 'people are capable of do-ing=and the skills that industry needs, theAmerican economy seems bound to face .in-creasing problems in meeting its manpowerneeds, as well as controlling unemployment,unless progress can be made in training andretraining. A company might, for example,lend an employee the money to cover vocation-'al schooling, retraining, or an advanced degreeprogram, with ropayment forgiven if the em-ployee remains with the firm for an agreedperiod. Tax policies and other instruments ofGovernment support could increase the incen-tives for both corporations and individuals.

fit,


Public policies might be designed to lessen therisks that companies ::;ponsoring education andtraining for their employees would lose their'investments when people switch jobs. The Fed-eral Government could provide incentivegrants to the States, to be raitched with cor-porate support.

Preparation for Work inTechnology-Based Industries

At the root of many of the present and vo-spective difficulties outlined above lies pourpreparation in science and mathematics pro-vided by the public schools. Leaving aside thelarge number of functional illiterates amongU.S. high school graduatesan illiteracy ratethat some estimates place as high as 20 per-centand the one-quarter of this age group thatdoes not even complete high schbol, manygood students get little education in science ormathematics once they reach the upper grades.The number of students electing courses thatare prerequisites for careers in technical fieldsis low and still falling; even those who choosescience often shy away from physics and chem-istry in favor of biology or geology. Technol-Ogy, as opposed to science, is invisible withinthe public schools. As the U.S. economy con-

1

aleloommie

Photo credit: Digital Equipment Corp.

Computer-aided engineering design

tinues to shift from manufacturing toward serv-ices, and toward more knowledge-intensive in-dustries, the American labor force will needto be prepared for technology-based jobs or riskdoing without. Even those performing un-skilled work will be in a position to makegreater contributions to productivity and com-petitiveness, while enhancing their own job se-curity and job mobility, if they are comfortable

. with numbers and quantitative reasoning, andhave a basic understanding of the physicalworld.

Part of the problem is again a shortage_pf--teachers; secondary schools are being stripped--of their science and mathematics instructorsby the attractions of higher paying jobs in in-:dustry. But the fundamental point is this: a stu-dent who opts out of scienceand particular-ly mathematicsat an early age has made a vir-tually irreversible decision, foreclosing a widerange of options in college and in his or hercareer. If American students continue to turnaway from mathematics and science at second-ary and high school levels, the United Stateswill find itself with an even greater fraction oftechnological illiterates in the adult population.Already, the Nation finds itself with few lead-ers in industry or Government who grasp theworkings of technology.

Management and Organization

Patterns of organization and management inbusiness enterprises mediate between the skillsand abilities that employees bring with themto the workplace and outcomes in terms ofcompetitive firms and industries. How wellcompanies utilize the talents of the people theyhire is quite as important as how good thesepeople are to begin with. Thus management,style and philosophy becomes a second criticalelement in human resources for the U.S. elec:tronics industry. While American managementincludes a "human relations" or participativemanagement tradition, employee involvementtends to be honored in theory more than inpractice. Techniques flowing frbm scientificmanagement, the other main tradition, remaindominant in U.S. industry. The recent voguefor Japanese management practices repres9nts

46


a swing of the pendulum toward the human re-lations pole, offering paths to improved com-petitiveness for some U.S. fh ms, though nosure cures.

Within a given countrywhether the UnitedStates, Japan, or one of the European nationselectronics firms show a good deal of diversityin manageMent style. Nonetheless, successfulelectronics companies in the United States andJapan exhibit more similarities than differ-ences. Despite the current fascination with the"secrets" of Japanese management, uniquelyJapanese traits are rare even in the cruder ster-eotypes. If the differences between firms with-in each country are often greater than the dif-ferences between countries, and clear-cut dis-tinctions between management styles in the:-United States and Japan less common thanoften assumed, two features of Japanese man-agement do, stand out: first, reward-structuresin Japanese companies create incentives for tal-ented people to huild careers in manufactur-ing; second, Japanese organizations tend tostress human relations more consistently andmore effectively.

Generally speaking, manufacturing and pro-duction engineering get more attention andmore status in Japanese companies than Amer-ican. This is one reason consumer electronicsand semiconductor firms in Japan could moveswiftly to create perceptions that theinproductsoffered better quality and reliability. Often, asdiscussed in chapter 6, those perceptions, werefirmly grounded in reality; although Americanfirms have largely caught up, the strong institu-tional commitments in Japanese corporationsto production engineering mean continuingpressure in this area. Furthermore, that manu-facturing managers in Japan carry more weightin corporate councils means in at least somecases faster shifts into automated production.The importance Japanese companies place onmanufacturing in their internal decisionmak-ing also translates into a greater share of re-sources for developments such as the complex

and demanding tasks of integrating roboticsand other forms of programmable automationinto the factory environment; companies thatlearn to utilize programmable automation mosteffectively will reap substantial competitivedividends in the future.

Stress on human relations is certainly notunique to Japanese organizations, but is moreconsistently visiblenotably among large com-panies characterized by low labor mobility and"lifetime" employment (ch. 8). That workers atall levels tend to spend much if not all of theircareers within a single organization createsstrong incentives for internal training, job rota-tion, and other steps aimed at improving peo-ple's skills and preventing stultification. Anumber of successful American electronicsfirms also go to considerable lengths to retaintheir employees, even in periods of businessdownturn. Rather than treating the labor forceas a variable cost, such firms, in both countries,regard their workers as a resource to be re-tained and nurtured although economic condi-tions might seem to point toward layoffs. Ac-complishing this implies more than keepingpeople at work and providing education andtraining. It also implies opportunities withinthe corporate structure to fully utilize presentand potential skills without sliding into pater-nalism or coercion. A number of the highlypublicized techniques associated with Japanesemanagement could, in fact, be as fairly termedmanipulative as participative.

A renewed commitment to the developmentand utilization of the human resources avail-able to American firms could make a majorcontribution to the future competitiveness ofthe U.S. electronics industry, indeed may becritical for the future prospects of this as wellas other high-technology sectors of the Ameri-can economy. Management practices in suc-cessful organizations, whether American orJapanese, tend to be associated with attentionto human relations and employee participation.

4 ./


EmploymentContinuing inadequacies in U.S. education

in science and mathematics will aggravatestructural unemployment caused by technolog-ical dovelopment and shifting competitivenessamong American industries. In the past, tech-nical change has generally created more jobsin the aggregate thIn have been destroyed. Un-fortunately, ;!here are no guarantees that con-tinued technological changeespecially thatresulting from applications of microelectronicsand computerswill in the future lead to ag-gregate increases in job opportunities. InEurope, the term "jobless growth" has cometo describe the widespread phenomenon ofhigh unemployment despite expanding output.This may or may not have been happening inthe United Statesthe evidence either way isscantybut structural unemployment is a reali-ty here.

Shifts in the composition of the work forcein electronics illustrate one of the conse-quences of technological and structural

120,000

a)u1 100,000

80,000

C)-.)

E

z60,000

40,000

20,000

0

1960 1965 1972 1973 1974 1975 1976

Year

SOURCES: 1960 -1965 -1977 Census of Manufactures. 1972-82Bureau of Labor Statistics.

change. In the United States, it is fair to saythat jobs in electronics are becoming more skill-intensive. Only in the manufacture of con-sumer products like TVs, a portion of the in-dustry that has been relatively stagnant, has theratio of blue-collar to white-collar employeesreniained high. In both computers and semi-conductors, the fraction of white-collar work-ers is much greater and increasing.

But a division into skilled and unskilledorwhite-collar, grey-collar, and blue-collar, not atall the same thing given the high levels of know-how associated with some but not all jobs ineach of these categoriesis too simple. Itmasks the increasing stratification and special-ization characterizing technologically based in-dustries, not only electronics but the sectorsit feeds. The journeyman machinist may go theway of the tinker as computer-controlled ma-chines replace engine lathes. The skilled me-chanic who could rebuild such a lathe canprobably no more fix the electronics of a

U.S. Employment in Consumer Electronics

e I.%. e .... .e. e - N.. .. ...

. N.

All employeesProdaction workers

NN. . ...

MO.

..1116

1977 1978 1979 1980 1981. 1982

A r--;


450,000

400,000

350,000

a) 300.0000ClE

0 250.000

200.000

150.000

100,000

50.000

U.S. Employment In Computer (and Peripheral Equipment) Manufacturing

All employees

Production workers

so* "

0

1955 1960 1965 1967 1968 1970 1972 1974 1976 1978 1980 1982

YearSOURCES: 1955-65-1977 Census of Manufactures. 186682Bureau of Labor Statistics.

numerically controlled machine than programthe computer that controls it. Specialists notonly design the parts to be made and programthe computer, but pick the feeds and speeds,specify tool materials and cutting fluids. Gag-ing and inspection may be automated, ratherthan the responsibility of practiced hands withdial gage and surface plate. As skilled jobschangeand at least some skilled work disap-pears along with unskilledpeople who haveno skills to start with will face still more trou-ble in finding satisfying, well-paying employ-ment. Those who cannot learn new skills may:find themselves outside the labor pool. Upwardmobility in the United States may. decline.

Many unskilled jobs are migrating over-seasin electronics, mostly to low-wage coun-tries in Asia. Moves offshore by American cor-porations have attracted widespread attention,and opposition on the grounds of "exportingjobs." Offshore assembly has been.much moreprevalent in semiconductors and consumerelectronics than in computers; even so, in bothsectors, other factors have often made greatercontributions to declining blue-collar job op-portunities (see ch. 9 as well as app. B). Amongthese factors, improvements in labor produc-tivity, many .stemming from investments inautomated manufacturing equipment, havegenerally had the greatest impacts. Moreover,

4 a


transfers of production offshore tend to hi,:driven by competitive pressures, domestic aswell as foreign in origin, which are largely out-side the control of individual firms. For in-stance, once a few U.S. semiconductor manu-facturers began assembling chips in low-wagecountries to cut costs, other suppliers had lit-tle choice but to follow suit. When the pres-sures are international, moves offshore may insome cases save domestic employment oppor-tunities over the longer term by helping main-tain U.S. competitiveness, though sacrificingjobs in the shorter term.

Manufacturing by American-owned as wellas foreign-owned companies has become wide-ly dispersed internationally. But this is only onecause of unemployment in the United States.Ongoing structural and demographic shifts arecausing serious and persiStent adjustment dif-ficulties. Peoplf.; with few skills or with obsoleteskills will find diminishing job opportunitiesin many of the older U.S. industries. Ten mil-lion and more Americans have been out ofwork at a time when American industry hasbeen short of as many as a million employeeswith specific skills and abilities. In the ag-agregate, and even considering multiplier' ef-fects, a million new jobs only dents the un-employment problem facing the United States.Yet from the standpoint of the individual, eachjob counts. Policyrnakers may find themselves

'unable to predict the causes and consequencesof structural unemployment with any preci-sion. This does not mean the problems cannotbe attacked. It means that adjustment measuresshould aim to enhance job mobilityintra-firmas well as inter-firmwithout depending on de-tailed predictions of supply and demand by oc-cupational category and industrial sector.

The total number of jobs created over thenext decade in electronics and other high-tech-nology industries will' not be large. After all,the entire U.S. electr:/nics industry employs onlyabout P12 million people today, and employ-ment has not expanded as rapidly as output.Still, many of the fastest growing occupationalcategories-hrthe economy will be found in thissector. The people who fill the new jobs willbenefit; at the same time, U.S. electronics com-

99-1 1 1 0 83 4

Predicted Growth Rates by Occupational Categoryin the United States Over the 1980's

Occupationa

Predicted increasein employment

(1980.90)Paralegal 109%Data processing machine mechanic 93Computer oper,Itor 72Computer systems analyst 68Business machine service technician 60Computer programer 49Employment interviewer 47Computer peripheral operator 44Psychiatric aide 40allon(nclusive; fastest growing occupations In electronics are listed together with

selected occupations outside of electronics On italics) for comparison.

SOURCE: "Testimony Before the senate Subcommittee on Employment and Pro.ductivity, March 26, 1982, by Ronald E. gutscher, Assistant Commis.stoner, Office of Economic Growth and Employment Projections,Bureau of Labor Statistics," Productivity in the American Economy,1982, huarInge, Subcommittee on Employment and Productivity, Cornmittee on Labor and Human Resources, U.S. Senate, Mar. 19 and 26;Apr, 2 and 16, 1982, p. 327.

panies need good people to remain competi-tive. Nonetheless, there has as yot been littleconcrete discussion of what is needed to pre-pare people for future job opportunities; theorganizations and institutions that deal withsuch concerns tend to be dispersed and o op-erate indepen tly of one another. Althoughthe past few years have seen considerable crit-icism of training programs said to be prepar-ing people for jobs that have already disap-peared, little usable information on such sub-jects in fact exists. Local control of secondaryand vocational education is the traditional pat-tern in the United States. Educators andschools of education seldom interact extensive-ly with industry or organized labor. Vocationaleducation and training has little visibility at theFederal level. Over the past two decades,schools have turned away from providing mar-ketable skills. Company-run training programsare limited in number and tend to be emergen-cy responses to hiring shortfalls rather thaneveryday features of corporate organization. Athorough re-thinking of the American approachto education and training, particularly for blue-and grey-collar workers, seems called for. Con-gress could decide to take the lead in reinvig-orating the traditional American commitmentto education and training.

5tl


TradeTrade policies pursued by the Federal Gov-

ernment have affected the several portions ofthe U.S. electronics industry in radically dif-ferent ways. Consumer electronics has sufferedfrom uncertain enforcementsome wouldnonenforcementof antidumping statutes, al-though other parts of U.S. trade law ! ave beencalled on to protect domestic firms L'om im-port competition. American manufacturers ofsemiconductors and computers have benefitedfrom U.S. leadership over the postwar periodin creating an open environment for interna-tional trade and investment. One of the'strengths of American semiconductor andcomputer firms has been their global approachto markets, a strategy aided by reductions inbarriers to trade and investment over the pa:three decades. Even though, semiconductor -im-ports frotn Japan have increased rapidly cloring the last few yearS, more than three -qtL ,t..

of U.S. semiconductor imports continue to c ,in-sist o. interdivisional shipmeinia of Americancompanies. In Japan, the largest exporter byfar among local computer manufacturers isIBM-Japan.

Antidumping EnforcementDumping complaints leveled at h -Torte's of

Japanese TVs as early as 1968 have never beenfully resolved. Dumpingselling imported TVsat prices below those charged in Japan wasproven under U.S. law, but legal challengesand interagency disputes have delayed finalcollection of duties for years.

During the 1960's and 1970's, Japanese TVmanufacturers, followed by those in SouthKorea and Taiwan, relied heavily on price cut-ting to force their way into U.S. markets. None-theless, dumping-Is-as neither the sole cause noreven a primary- cause of competitive shifts inconsumer electronics. The worldwide successof Japanese consumer electronics firms amplydemonstrates their ability, not only to manu-facture at loW cost, but to produce reliable-TVswith good performance and product featuresthat American consumers want. At the same

time, uncertainty created by lengthy and in-conclusive legal proceedings means that do-mestic firms could not know whether theymight eventually Iv' able to raise prices as a

I stilt of antliltillWi.6 1.tties ied imports.These added duties might have totale(" wellover $100 million; higher prices generatinghigher profits could, in principle, have aidedembattled American firms in revitalizing theirbusinesses.

Eventually, U.S. color TV manufacturers andtheir suppliers did receive trade protection, inthe form of negotiated quotas on imports fromJapan, Korea, and Taiwan. Under the name Or-derly Ma rketing ereemohL; As), thequotas followed escape clause proceedingsfiled in the wake of sharp rises in color TV im-ports during tL ff 1970's. 1 i'lfair trade practiceswere t at The iJi,f: , speeded struc-tural shifts already underway in the U.S. con-sumer electronics industry. By limiting im-ports, they created incentives for foreign firmsto invest in assembly plants within U.S. bor-ders. OMAs did little to revive the Americanconsumer electronics industry; they did accel-erate foreign investments, many of whichwould eventually have been made in any case.In effect, weaker American TV manufacturersdriven from the market by imports have beenreplaced by subsidiaries of foreign firms. Whilethese subsidiaries help to maintain domesticemployment, half or more of the value addedtypically remains overseas.

The Environment for World Trade

A number of American computer firms thatbegan as makers of office equipment, includingIBM, maintained foreign operations before thewar. During the postwar period, overseas in-vestments by American computer manufac-turers expanded; subsidiaries of U.S. firms be-came the backbone of computer industries inmost parts of the,industrialized world. Today,

along with-the-older companies whose productlines still center on general-purpose main-frames, the major American manufacturers of

minicomputers also operate all over the globe.Likewise, U.S. semiconductor firms began toinvest overseas at an early stage; these invest-ments, beginning around 1960, were made fortwo reasons: 1) to supply foreign markets vialocal production in industrialized nations; and2) to cut costs by moving labor-intensive man-ufacturing operations to low-wage countries.

Direct and Indirect Barriers

Foreign investments by U.S. computer andsemiconductor manufacturers, along with theircontinuing high levels of exports, have been fa-cilitated by a relatively open environment forinternational trade and investmentchapter11. Created largely under the auspices of theGeneral Agreement on Tariffs and Trade(GATT), in which the United States has playeda major role, this opening of opportunities formultinational firms via relaxation of direct bar-riers to trade and investmenttariffs, importqudtas, restrictions on flows of capital outwardas well as inwardhas, on the whole, been ofgreat benefit to the U.S. electronics industry.At the same time, relaxation of direct barriersto trade has been accompanied by a simulta-neous increase in less direct obstacles andcontrols.

As the industrial and trade policies of foreigngovernments have evolved, they have swungtoward more subtle combinations of indirectimport barriers, performance requirements, in-vestment incentives, and subsidies. In somecases, these measuresdescribed in chapter10are intended to influence investment andexporting. In others, t!;-- objectives are primari-ly matters of domest_ olicy: national securi-ty, employment, regional. developmc:at. Gov-ernments intent on pursuing industrial policiesthat will support local industries while attract:ing U.S. dollars and/or technology can choosefrom a well-stocked arsenal: trade barriersrange from paperwork obstacles to "buy na-tional" rules; performance requirements mayentail transferring technology, purchasing sup-plies and materials locally, or exporting aprescribed fraction of production as a condi-tion for investment; common forms of subsi-dies include R&D funding, capital preferences,

Ch. 1Part B: Extended Stimmary 43

and guaranteed procurements. European na-tions, in particular, have sometimes used in-vestment incentives to attract American elec-tronics firms in the name of jobs and tech-nology.

Over the past half-dozen years, spokesmenfor the U.S. semiconductor industry have fre-quently complained that the trade practices ofsome foreign enterprises have been unfair,while also voicing concern over governmentindustrial policies in countries such as Franceand especially Japan. (U.S. computer firmshave seldom been as vocal over trade practicesor internal subsidies benefiting their foreignrivals.) Among the restrictive practices that stillexist in many parts of the world, the relativelyhigh tariffs levied by the European Communi-ty (EC) on semiconductors=17 percentareperhaps the most visible. One consequence hasbeen to encourage investments within the ECby American firms, but the European marketis in any case large enough that these invest-ment patterns could have been anticipated. Inneither semiconductors nor computers haveEuropean suppliers been very successful in ap-proaching the EC market as a whole. With onlya few exceptions, local firms have exhibitedrelatively fragmented patterns of productionand sales. In contrast, American-owned enter-prises have often done a better job of treatingEurope as a unified regional market. Butwhereas the trade practices of Western Euro-pean nations may have ended by harming theability of local firms to compete with theAmericans more 'than they have helped, thesituation has been vastly different in Japan.

Japan

For many years the Japanese Government ef-fectively protected the country's electronics in-dustry, including manufacturers of consumerproducts such as TVs, through controls overforeign investment as well as restrictions onimports. With only a few exceptionse.g.,IBM-JapanAmerican-owned computer andsemiconductor firms have had no more thanmodest success in selling their productsthrough either exports or local production. Acomplex of factors ranging from chauvinism


400

350

300

0 .50

-ra; 150

100

U.S.-Japan Trade in Integrated Circuits

0

U.S. importsfrom

Japan

-r

U.S. exportsto

Japan

1975 1076 1077 1978 1979 1980 1981 1982

Year

SOURCE: 1983 U S. Industrial Outlook (Washington, D.C.: Dapariment of CornmerCe. January 1983, p. 29 -5. 1982 figures estimated.

to explicit government policies has impededboth exports and investments in Japan byAmerican electronics suppliers. The negativeimpacts on U.S. competitiveness have been fargreater than those visible anywhere else in theworld.

American companies have been able to sellproducts that the Japanese could not make forthemselvesadvanced integi.ated circuits,state-of-the-art semiconductor fabricationequipment, some types of computers. But prod-ucts available from Japanese suppliers tend tobe purchased locally, in part because of deep-ly ingrained "buy Japanese" attitudes. Struc-tural differences also play a role, particularlyin microelectronics: the half-dozen large com-panies that produce most of Japan's semicon-ductors also consume perhaps two-thirds ofthese same semiconductors; such a market isdifficult to attack from the outside. Whileforeign investment is in theory much :less re-

:_stricted today than in the past, only a fewAmerican semiconductor and computer firmshave as yetestablished wholly owned opera-tions of any size within Japan.

Given the rapidly improving technologicalabilities and competitive postures of Japaneseelectronics manufacturers, investment in Japanby American firms appears vital for maintain-ing U.S. competitiveness; while many in Jap-anese Government and industry will no doubtcontinue to oppose such investments, Japan'sGovernment has officially endorsed liberaliza-tion many times, and should be held to thesestatements in practice as well as in principle.The Japanese market for electronics productsis now second only to that of the United States;for many types of products, sales within Japanexceed those for all of Western Europe. Notonly will local manufacturing help expandmarkets for American firms, enabling them tocompete more effectively with Japanese com-panies in third countries as well as insideJapan, it will accelerate flows of technologyfrom Japan to the United States. As in indus-tries such as steel or automobiles, Americanelectronics companies can now learn fromtheir Japanese counterpartsand not only inconsumer products. U.S.-owned R&D labora-tories in Japan could help compensate for per-sonnel shortages here, as well as improving ac-cess to the results of subsidized research pro-grams such as the fifth-generation computer ef-foil. Full participation in the dynamic Japaneseelectronics market is critical to the continuingcompetitiveness of American computer and

Semiconductor Sales in the United States,Western Europe, and Japan

Sales (billions of dollars)1974 1982

United StatesDiscrete semiconductors . $0.88 $1.3Integrated circuits 1.2 6.3

$2.1 $7.6Western Europe

Discrete semiconductors . : $0.77 $0.77Integrated circuits 0.52 1.5

$1.3 $2.3Japan

Discrete semiconductors _ $0.55 $1.2Integrated circuits P.59 2.4

$1.1 $3.6SOURCES: 1974Electronics. Jan 8, 1976. pp. 92. 93. 105.

1982Electronlcs. Jan. 13, 1983, pp. 128, 142, 150; Mar. 10, 1983, p. 8.


semiconductor manufacturers; Congress couldlielp ensure that the Federal Government ac-tively supports such endeavors by Americanfirms, which are fully consistent with thiscountry's historic commitment to open tradeand investment. Competition in,Japan on termsperceived to be fair will yield dividends withinthe United States by creating conditions underwhich American companies can better main-tain their competitiveness.

Recent Developments

Broadly speaking, the Tokyo Round multilat-eral trade negotiations, completed in 1979 andimplemented shortly thereafter, in the UnitedStates by the Trade Agreements Act of 1979,are having generally positive though small im-pacts on the American electronics industry.Continuing tariff reductions will help U.S. ex-ports; accelerated duty reductions on semicon-ductor products and computers by Japan areespecially significant, though perhaps as sym-bol more than substance.

Nonetheless, tariffs are no longer the prin-cipal barrier to international trade in elec-tronics. They are being replaced by indirectand nontariff barriers, including a wide rangeof implicit and explicit subsidies. In particular,American electronics firms continue to com7plain over government-funded R&D programsin Europe, Japan, and a number of developingcountries. Although the Tokyo Round yieldeda new subsidies code intended to deal with thisand related issues, the prospects for substan-tial progress seem slim.

Taken one at a time, individual programs inforeign countriesincluding such prominentexamples as Japan's VLSI R&D effort, and,prospectively, the fifth-generation and super-computer projects now underwayhave oftenhad no more than modest impacts on interna-tional competitiveness. At the same time, theirgoals often include intangibles such as technol-ogy diffusion or improvements in the skills of

the labor force; these make outcomes difficultto evaluate. Subsidies directed at commercialtechnologies and typically rationalized as do:mestic support measures rather than exportpromotion policies have few counterparts inthe United States. It is the justification in terms,of domestic objectives rather than strengthenedexport competitiveness that makes such poli-cies a problematic subject for internationalnegotiations. Public funds for R&D, the use ofgovernment procurement to favor domestic in-dustries, and the many related instruments ofindustrial policy detailed in chapter 10 havebecome part of the conventional approach byforeign governments. Countries in many parts..of the world pursue such measures in hopesof building their competitiveness in high-tech-nology sectors like electronics. .

Given the growing prevalence of plannedprograms of industrial development, virtuallyall of which give electronics a prominent place,.it seems unlikely that continued U.S. attackson such policies as "export subsidies" will havemuch effect in arresting the trend. This is par-ticularly true given the indirect and intangible.impacts of programs directed at infrastructuralsupport, precompetitive technology develop-ment, or human resources. Many of Japan's in-dustrial policy initiatives have' been directedat overcoming structural obstacles such aslimited labor mobility and a less than stimu-lating working environment for technical pro-fessionals. Totaling the monetary value of suchsubsidies, even where possible, is an exercisethat holds little meaning. And, in the end, mostforeign governments will regard such effortsas too important to give up; certainly they donot welcome them as legitimate topics of in-ternational negotiations, bilateral, or multi-lateral. While it is clearly in the interests of theUnited States to press for clarification andagreement on the "rules of the game," it maynot be very productive to devote a great dealof effort to combating on a case-by-case basiswhat have become standard tools of industrialpolicy in other. countries.

54


U.S. Industrial PoliciesAs many observers have noted, industrial

policy in the United States has been a largelyad hoc construct of unrelated measures aimedat diverse object' vr,s; not infrequently, policymeasures have vcaLl..:.ed at cross purposes or ledto unanticipate i outrnmes. Seldom have they_represented core,.-.4,inv., <attempts to stimulate thecompetitivrmess of American industries. Tradepolicy, t;:eated separately above, is only a par-tial exception.

In recent years, U.S. industrial policies haveoni had major impacts on the electronics

industry; however, early developments in bothcomputers and semiconductors benefited fromCovernment procurement and from R&Dfunded by Federal agencies concerned withdefense and space. Since the 1960's, overlapsbetween military/space applications and civil-ian needs have diminished. Today, militaryelectronic systems are seldom as advanced ascivilianl-it;has-been-many years since Federalspending-has had much influence over elec-tronics technology or cornpetitivenefs.

Distribution of U.S. SemiconductorSe, by End Market

coro_cfner Computer andindustrial

SOURCES: 1980, 1968: "Innovation, Competition, and Governmental Policy In the- -- Semiconductor Industry,... Charles River Associates, Inc., final

'report tor Experiment at-Technology Incentives-Program, Depart-ment of Commerce, March, 1980, p. 2-13.

1980: Status '80: A Report on the Integrated Circuit Industry (Scotts-dale, Ariz.: Integrated Circuit Engineering Corp., 1980), p. 34.

Faced with increasing competition in manyindustrial sectors, slower economic growth,and a multitude of adjustment problems, thequestion for the United States has become: Canthe country continue with a de facto industrialpolicy or is a new approach needed? The sur-prising variety of programs intended to -nur-ture technologically based industries in Japan,Western Europe, and several newly industri-alizi-3 countries reveal an attentiveness toeconomic development simply lacking here.One response has been to argue that the UnitedStates needs to find ways of negating or coun-tering foreign industrial policies. Alternative-ly, rather than a reactive posture, the UnitedStates could itself move toward policies intended to stimulate and support industrialdevelopment.

Foreign industrial policies have had their fail-uresand successes too. The important pointis that countries which have adopted relative-ly systematic industrial policies continue to ex-periment with policy tools, to develop new pro-gramsin short, to accumulate expn:ience andimprove effectiveness. The U.S. system hasstrengths and weaknesses different from anyand all of the nationsJapan, France, SouthKoreathat have pursued industrial policiesaimed at economic growth and development.What sort of industrial policy could help theUnited States to maximize its own strengths,minimize its weaknesses? To trap frame thisquestion, OTA suggests five possible orienta-tions that Congress may wish to consider fora mole coherent U.S. industrial policy:

1. A policy approach aimed at ensuring astrong domestic market base for U.S. in-dustries, along with preservation of exist-ing jobs and job opportunities.

2. Policies designed to protect and/or supporta limited number of industries judgedcritical to national security, defined nar-rowly-or broadly.

3. Measui4s that will suppart_the_technologr___:.ical-base infrastructure

for American industries, particularly thoseundergoing structural change.

Ch. 1Part B: Extended Summary r 47

4. Policies intended to promote the globalcompetitiveness of American industries.

5. An orientation that would defer whereverpossible to the private sector when choicesconcerning the development of industryare to be made.

These policy directions, examined in detail inchapter 12, are by no means mutually exclu-sive; they might draw, for example, on similar,policy tools in areas such as international tradeor technology development. Nonetheless, they'represent distinctly different thrusts: the goalsdiffer even where the instruments are ali'7e.

The five alternatives, outl: -led below in thecontext of the electronics industry, carry impli-cations for the entire economy, as well as thepolitical environment where any policy wouldhave to be implemented. These broader dimen-sions are emphasized below because focusingtoo strongly on specific policy tools e.g., thoseaddressing problems visible at the moment inelect7onicswould simply repeat the ad hocapproach to U.S. industrial policies nowcurrent.

Each of the five approaches has positive andnegative aspects. They can be usefully con-trasted in terms of differential effects on sec-tors of the economy as well as susceptibilityto the political forces that corporations andtheir employees bring to bear on the policymak-ing process. The intrinsically political charac-ter of this process, now or in the foreseeablefuture, has often been couched in terms of Gov-,ernment's ability to pick and choose among"winners" and "losers." Early debates over in-dustrial policy in the United States tended tofocus on such questionsrather pointless giventhat many. Federal policies have always hadthis effect. The Economic Recovery Tax Actof 1981 treats some industries much more fa-vorably than others; trade protection has in re-cent years been extended to manufacturers ofcolor TVs, automobiles, and clothespins; politi-cal pressures routinely affect decisions onpublic works and defense projects.

When industrial policy decisions are madeon an ad hoc basiswithout linking one sec-tor of the economy to others, without setting

the problems of a domestic industry into thecontext of the world economypolitical con-siderations can more easily predominate. Tobegin coordinating such decisions more closelycarries two quite different implications: 1)greater reliance by the Federal Government onempirically grounded analysis of industrialcompetitiveness, productivity, and economicefficiency; and 2) risks thatbeyond'influenc-ing policy decisions on a case-by-case basis, ashappens alreadypolitical pressures will skewthe policy approach as a whole. The first is oneof the potential advantages of a mare coherentindustrial policy, the second, one of the pit-fallsa pitfall because companies and indus-tries in trouble, and their employees, have amore obvious stake in policy decisions, hencebring more pressure to bear, than sectors of theeconomy on the upswing.

The first two of the policy orientations listedabove carry the greater risks of political deflec-tion. Ensuring the domestic market base forU.S. industries could easily amount to nothingmore than a protectionist response to tradepressures and the rise of competitive enter-prises in other countries. Basically an inwardlooking, defensive strategy, it equates importpenetration with damage to U.S. interests. Anindustrial policy centered on safeguardingAmerican markets and American jobs wouldbe largely congruent with the political forcesthat will always advocate protectionist meas-uresfirms and industries in competitive de-cline, their employees, the communities and re-gions in which they are located.

Decline may be temporary and reversible, orit may be the consequence of deeply rootedshifts in the international economy that, overthe longer term, are iikely..to force contractionregardless of public policy responses. A marketprotection strategy implies, first of all, deter-mining whether protection is needed becauseof short-term problemswhich might rangefrom macroeconomic dilemmas to misjudg-ments by corporate managements. For suchreasons, temporary protection is sanctioned byinternational trade la ''u under circumstancesas described in chapter 11. Indeed, temporarytrade restrictions might find a place in any in-


dustrial policy alternative. Longer term declinebrings a different set of issues to the policymak-ing process; the options may range from man-aging decline (via adjustment measures in-tended to ameliorate the most immediate prob-lems)-to wholesale protection and subsidization(as several European nations have occasional-ly attempted).

A critical industries strategywhether"critical" refers narrowly to military strengthor carries sc,ne broader economic connota-tionwould also lead to a great deal of politicaljockeying among firms and industries bent ondemonstrating their criticality. Snch an out-come is virtually inevitable because few objec-tive criteria exist that would allow essential orcritical industries to be identified beyond thebroadest and most general level. Under virtual-ly any criteria imaginable, electronics wouldbe judged vital to "economic security" as wellas military security. Even so, when the industryis disaggregated, judgments at finer levels im-mediately become difficult.

Electronics would probably not suffer undereither a protectionist or a critical industries ap-proach; although backlashes by other countriesare always a possibility, nations that importhigh-technology electronics products usuallyneed them badly enough that they would pickother alternate targets for retaliation. Even so,a number of other U.s. industries would be like-ly to benefit more. For at least some compenies,the lob'iying involved would be business -as-usual.. Large and powerful corporations experi-enced in dealing with the Federal Government,defense contractors, and firms in heavilyunionized industries would tend to have anedge over smaller, technology-based concerns.The more aggressive and outward looking high-technology portions of electronics couldnot ex-pect as much positive support as they mightget' under other policy decisions.

Under any of the five alternatives, politicalforces would bear heavily on policy outcomes.Firms and industries will always have strongincentives to press for direct and indirect sub-sidies flowing from Federal decisions. This isquite understandable, and built into the Amer-

ican political system, but has consequencesthat are largely undesirable if a basic objectiveof industrial policy is to improve U.S. competi-tiveness. Industrial policies are most likely tobe productive and effective when they comple-ment ongoing changes in the world econ-omye.g., by aiding structural adjustment.When industrial policies oppose long-termshifts in comparative advantage, they are gen-erally doomed to high costs, inefficiencies, andmarginality if not ultimate failure.

This could well be true, for instance, in thecase of Federal actions that would steer capitalto selected industries. Such policies have fre-quently been advocated by those favoring "re-industrialization," as well as a critical in-dustries orientation. However, targeting of in-vestment in a consc:ous waya key elementin many foreign industrial policiesseems anunlikely prospect for the United States, if onlybecause capital markets here work much bet-ter than in most other economies. Moreover,the Federal Government's experience with in-vestment, leaving aside sectors such as hous-ing, has been restricted mostly to aggregatemeasures and to a few well-known bailouts ofloubled corporations. Finally, the records of

foreign countries that have tried to channel in-vestment into industries intended as mainstaysof economic growth and competitiveness aredecidedly mixed.

Everyone knows what the future growth in-dustries will be. The current list includescomputer-aided manufacturing and robotics,biotechnology, new nonmetallic materials,microelectronics, computers and communica-tions. U.S. capital markets have been "picking"these winners quite effectively. An industrialpolicy intended to support future U.S. growthindustriesunder a critical industries rubricor some other policy approachwould have todo more. Specifically, it would have to searchout cases where markets were not performingconsistently well. These do exist. The timehorizons of markets may be shorter than desir-able from a public standpoint (there are manyexamples in R&D, most notably in basic re-search but also in the development of generic


technologies that could benefit a wide rangeof firms -While being difficult to protect ormonopolize). Bottlenecks are always possible(the un,linhiguous successes of foreign indus-trial policies often involve breaking bottle-necks). Response times can be excessively loag(as in the case of the educational system, heavi-ly dominated by government bodies whichcreate inertia and slow responses, while alsosuffering from cloudy perceptions of futureneeds and opportunities). Such examples pointto approaches that would not be explicitly sec-toral.

The third of the policy orientations consid-ered by OTApolicies that would provide gen-eralized support for technology and infra-structural development, cutting across sectorsof the economywould reduce the leveragethat special interests could exert by avoiding,where possible, policies with strong sector-specific thrusts. Instead, the tools of first choicewould have more aggregate objectivesnotonly R&D and its diffusion, but education andtraining, open competition, structural adjust-ment. At the same time, sectoral policies wouldnot be totally ruled out.

A variety of instruments are available:manpower training and retraining;new institutional mechanisms for tech-nology development (emphasizing, for ex-ample, cooperative efforts among Govern-ment, business, and universities);incentives as well as direct funding forresearch and development;the infrastructure for diffusing availabletechnologies as well as new R&D resultsthrough the U.S. economy (including tech-nologies from overseas); andpolicies aimed at stimulating capital for-mation and investments in new and pro-ductive technologies.

By supporting the technological and human re-sources underlying competitive industries, in-terest groups anglingfor special favors wouldhave fewer obvious and attractive targets, atleast in terms of immediate financial rewards.Primarily future-oriented, this policy orienta-tion is based on the assumption that the Federal

Government can help build competitiveness bypromoting evolutionary shifts in the i7.onomy,as well as by easing the negative impacts of ad-justment on particular groups and regions.

In terms of R&D, the chief difference be-tween private sector and Federal Governmentdecisions lies not in the ability to evaluate op-portunities but in the longer time horizons thatGovernment can bring to such questions. Mo-tivated by social rather than private returns toinvestment, unconcerned with capturing im-mediate rewards, public policy initiatives canbe formulated with a longer term view than pri-vate corporations take. This is as true for ma-ture industries like steel or automobiles as for"sunrise" or growth sectors. One of the tasksof an industrial policy oriented toward adjust-ment and infrastructural support would be tofind such opportunities and develop appropri-ate responses. To develop an industrial policycapable of attacking problems of this sort, theFederal Government would need to understandindustries and their workings on a concrete,practical levelthe level of the shopfloor andthe R&D laboratory as well as the boardroom.The Government does not now have this ca-pability, a capability it would need in order to

Photo credit: Intel Corp.

Silicon wafer after chip fabrication

50 International Competitiveness in Electrobics

implement with reasonable effectiveness anyconsistent and explicit industrial policy.

As pointed out above, competitive industriesdepend on the human resources available,while training and retraining are essential toeconomic adjustment. But what, specifically,should people be trained to do? What kinds ofskills will be needed 20 years from now? Whatare the best ways of reaching people alreadyin the labor force? Are institutional changesneeded? Should the United States continue toleave training and retraining largely to localinitiative, or is a continuing but redefined Fed-eral role needed? These are among the ques-tions with which this third approach to indus-trial policy would have to come to grips. Theyillustrate the need for a well-developed analyti-cal capability within the Federal Government.

Like the first of these five policy alternatives,preservation of domestic markets, the fourthpromoting the global competitiveness of U.S.industriescenters on trade issues. However,promoting competitiveness implies an outwardlooking, export-oriented stance, an emphasison openness in international trade coupledwith stimuli for emerging, competitive sectorsof the economy. Taking as its starting point thedynamics of international competitivenessthe rise and decline of industries over timethe global trade alternative would seek out,even accelerate, processes of change, attempt-ing to keep American industries technologi-cally and commercially ahead of their foreignrivals. To the extent that such policies hastenedthe decline of other portions of the economy,adjustment measures aimed at speeding re-source floWs out of these sectors, as well ascushioning the impacts of decline, might alsobe called for.

The global approach to industrial policybuilds naturally on the traditional U.S. attitudethat international trade benefits all parties andshould be encouraged. It is the option furthest

__removed from the common notion of industrialpolicy as necessarily working against Ofiennessin trade. The Federal Government might notonly continue to press for access for U.S. ex-ports and investments in other countries, but

reciprocally keep the domestic market open,while vigorously pursuing antitrust enforce-ment in the name of competition. Rather thanresorting to bilateral trade negotiations, theUnited States could continue to work towardmultilateral agreements aimed at reducing bar-riers to tradein the current climate, primarilynontariff and indirect barriers. Tax incentivescould be used to reward competitive, export-oriented firms. While more direct forms of ex-port promotion might also find a place, directmeasures always carry the danger of becom-ing subsidieswhich, in the name of competi-tion, this policy orientation would seek toavoid. Instead of protectionist measures foraiding troubled industries, the Governmentmight attempt to manage decline and en-courage restructuring.

If interestrgroups in the United States see theNation opening its ;own markets to foreigngoods and foreign investment an intrinsicpart of the global approachwithout corre-sponding openings in other parts of the world,this option could invite a strong backlash. Evenif the United States persuaded its trading part-ners to join in a thoroughly open and com-petitive world market system, the acceleratedprocesses of domestic change might generatec,f rong sentiments in favor of protection as wellas adjustment assistance. Open world trade hasmany attractions as one element in a morecohesive U.S. industrial policy,- but by itselfmight not offer advantages great enough orvisible enough to attract the political supportneeded for implementation.

The last alternative is built around giving in-dustry a free hand, where possible, in deci-sions that affect productivity and internationalcompetitiveness. This alternative fits the recentmood in the United States: that Governmenti:ivolvement in economic affairs is counterpro-ductive, that business activities should be de-regulated,. that markets work best and indus-tries compete best when the Federal presenceis minimized. Like the global trade alternative,it could mean more rapid- rises-and declineswithin the U.S. economy. Unlike that alterna-tive, it implies less attention by Governmentto structural adjustment and less support for

5.,

the efforts of American firms to export and/orinvest overseas. Nor would protection againstimport competition be looked on with favor.

Such a policy approach would have to con-frOnt and resolve the following dilemma. Amer-ican businessmen direct many of their com-plaints at foreign industrial policies that in-tervene in markets by, for example, encourag-ing mergers or allocating capital. Spokesmenfor U.S. industry often hold, on the one hand,that industrial policies in other countries arenot only unfair, but serve to tilt the competitivebalance by strengthening or even creating com-parative advantage. On the other hand, thesesame spokesmen frequently argue that Govern-ment policies could not do so here. Some ofthese statements may simply express a desirefor unfettered competition among all corners;in other cases, they appear to imply thatgovernment actions are counterproductive inthe United States but not overseas. In anyevent. the fundamental question is: Given thatforeign governments are not likely to abandontheir industrial policies so long as they considerthem useful, can the United States counterdiem simply by avoiding policy interventions?

More positively, then. this fifth policy ap-proach might include tax incentives for capi-tal formation and investment, deregulation,and free competition. Control of inflation andmacroeconomic stability wo! .-,ertainly re-main a Federal. responsibility i toser examina-tion of re^.ent changes in tax' policy' points toone of the. central issues raised by this alter-native: Cali Government really be a neutral ar-biter of economic competition? Past experiencegiVes little evidence in favor of the proposition.

The 1981 Tax Act seems on balance to havebeen a move away from neutrality in treatmentof the various sectors of the economy. Notingthat accelerated depreciation has varying con-sequences for manufacturers of consumer elec-tronics and semiconductorsand that these

4wo.parts-of-the electronics industry are, treated..quite differently than producers of heavy elec-trical machinery, much less nonelectricalmachineryindicates some of' the potentialproblems. Differential effects on various parts

Ch. 1Part 6: Extended Summary 51

of the economy are an unavoidable conse-quence of any industrial policy, and it may bebetter to confront such issues directly than tryto avoid them, as this last alternative would ingeneral do. While true neutrality can never beachieved, an industrial policy ostensibly in-tended to "get Government off the backs ofbusiness" would more likely end up rewardingthose who could bring the most political pres-sure to bear. These interests would probablybe able to perturb the policymaking processtax policy being only one exampleto theirown benefit, aided by the illusion that theFederal presence was diminishing. Industrieswith less political strength or sophisticationwould, in a relative sense, fare less well.

Indeed, it seems wishful thinking to argueagainst Government involvement in economicaffairs, although not against counterproductiveor excessive involvement. The fact is, of course,that governments here and elsewhere do in-tervene; it is part of their job. Moreover, aseconomies grow more complex and more heav-ily dependent on advanced technologies, theforces that governments seek to modify or con-trol may become more powerful, the need forgovernment action greater. When, where, why,howthe circumstances in which govern-ments intervene, the effects of the involve-mentare the crucial questions.

What does this mean for industrial policy inthe United States? First, more effective policiestoward industry in the United States will re-quire relatively broad agreeMent on objectives.Second, the Federal Government would needto develop an analytical capability adequate tothe task of reaching these objectives. Both areefforts to which Congress could turn its atten-tion. The first is largely a political task, thebasis of the argument that our standard of liv-ing depends on the international competitiveness of industries like electionics. The seconddemands that Government go beyond the large-ly static and abstract economic perspective thatin many agencies is now called on 'to justifypolicies adopted for-other red-sells:

The political, environment in the UnitedStates makes movement toward a more con-sciously developed industrial policyfollowing


any of the five alternatives outlined abovenot only slow and painful, but an endeavor thatrisks being turned to ends far removed fromeconomic efficiency. (This is not to imply ateconomic efficiency is the only goal of indus-trial pOlicy, but that one of the purposes of amore coherent approacli would be to bring thisand related objectives closer to the forefront.)But even where decisions are made largely onpolitical gioundsas will frequently be thecasea more explicit industrial policy couldhelp frame the questions, bound the responses,increase the probability that individual policyinstruments function as expected and in-tended. Given an international economy pop-

ulated by countries experimenting with indus-trial policies, and learning to use them moreeffectively, a pragmatic orientation. , by theUnited States, grounded in empirical analysis,could be viewedby Congress and the FederalGovernment as a whole, and by both partiesas a vital support for our own economy. Suchan attitude toward industrial policy would helpto ensure that the U.S. electronics industry andother high-technology sectors would get theirfair share of the resources neededlo competeeffectively in world markets. It isalso the besthope, in the longer run, for older industriesranging from primary anetals to machine toolsand textiles.

CHAPTER 2

Introduction

62

Contents

1.:kctrunics as a high-Techntdogy Industry 5611)fviduintient 56

Clania!titiut! 57Structural 1.:itangt 58

Thc Importance of Competitiveness 60

industrial P 61

. 62

CHAPTER 2

Introduction

electronics industry provides exam-ples that can support almost any perspectiveon competitive trends in the American econ-omy over the past decade. That portion of theindustry manufacturing computers has been achampion of U.S. economic strength bothdomestically and internationally. Here andabroad, American computer firmsparticular-ly IBMhave been symbols of technologicalprowess, market power, and multinationalmarketing and production. In Europe, U.S.computer manufacturers have been models tobe emulatedfor in,ii:;enotis companies likeICL in Great Britain, or for joint ventures suchas CH-Honeywell Bull in Franceand targetsto be displaced with the aid of national in-dustrial policies. In Japan, American computerfirms have been explicitly depicted as the en-emyIBM as a stateless, global giant, withJapanese firms urged to mount fierce effortsagainst it. Meanwhile, in the United States, theDepartment of Justice had in 1969 begun an an-titrust suit aimed at dismembering IBM, a suitthat was finally dismissed 13 years later.

American consumer electronics firms havebeen pictured much differentlyparticularlythe old-line manufacturers of televisions andother home entertainment equipment, such asZenith and RCA.-Many firms in this part of theindustry have seen themselves as victims of un-fair trade practices by overseas rivals, primarilyJapanese. Foreign firms selling TVs in thiscountry have been accused of dumping (andfound guilty of this), attempted monopoliza-tion, and of receiving subsidies from their owngovernments. To other observers, the U.S. con-sumer electronics industry has been a victimof management failures, has lacked the will tocompete internationally, has.ceded some seg-ments of its markets too easily to imports, andhas lagged in adopting manufacturing methodsthat could have cut costs and increased thequality and reliability ui its products.

Semiconductor manufacturers in the UnitedStates have, over the past several years, pointed

ble harbinger of their own fate if the U.S. Gov-ernment does nothing to support them in theircompetitive battles with foreign (i.e., Japanese)rivals. At the same time, American semicon-ductor firms share with our computer manu-facturers a deserved reputation as worldwideleaders in technology, innovation, and entre-preneurial zeala reputation which the 1980-83 round of new startups in Silicon Valley canonly enhance.

These three portions of the electronics indus-trycomputers, consumer electronics, andsemiconductorsare the focus of this report.But other parts of the industry could illustratemany of the same themes. Electronic compo-nent productionswitches, resistors and ca-pacitors, printed circuit boardshas beenmoved to offshore locations as part of the re-sponse to competitive threats from imports.Professional and industrial equipmentinstru-mentation, industrial process control, medicalelectronicsis a continuing U.S. strength, butagain the technological leads that Americanfirms once held have narrowed. In telecommu-nications, American firms have lost out in ;;ev-eral promising developing country markets.While boundaries between information proc-essing and information transmittal have beenblurring for years, and communications is cer-tainly one of the central electronics-related por-tions of U.S. industry, this report touches oncommunications only in passingnot becausethis portion of the industry is unimportant, butonly to keep the study to manageable propor-tions.

The breadth and diversity of the electronicsindustry contrasts with industries such as steel,which are often pictured as monolithic. Evenhere, however, specialty steel Sand noninte-grated "minimills" have prove% notable excep-tions to the commonly accepted no'ion of de-clining U.S. competitiveness., Steel is an old,

'Technology and Steel Industry Competitiveness (Washington,D.C.: U.S. Congress, Office of Technology Assessment, OTA-M-122, Juno 1980); U.S. Industrial Competitiveness: A Corn.parison of Steel, Electronics, and Automobiles (Washington,D.C.: U.S. Congress, Office of Technology Assessment. OTA-


established industry compared to electronics;vet the electronics industry has roots goingback to the early part of the century, in con-trast to biotechnology and genetic engineeringfor xvhich international competition hashardly begunthough here, too, there are rootsin fields like plant breeding and pharmaceuti-cals. Emerging industries like biotechnologyare important for future economic growth;electronics is critical right now. Moreover,lessons learned from electronics might applyto older, "mature" industries such as steel, aswell as to nascent sectors like biotechnology.

What .can be learned from electronics, pa'-ticularly the last 10 or 15 years? That is onethe questions this report attempts to answer.Is the apparent decline of the American con--sumer electronics industry irreversible? Arethe threats to U.S. computer and semiconduc-

tor firms real, or are they better considerednatural consequences of the growth and matur-ing of these portions of the industry? How havepolicies adopted by the Federal Governmentaffected the industry? How do public policieshere differ from those of foreign governments,both in their forms and in their effects? Towhat extent have foreign industrial policiessucceeded in strengthening the electronics in-dustries of other countries, in affecting the in-vestment and export strategies of Americanfirms, in replacing tariff and nontariff barriersto international trade with less visible but noless effective constraints? `:an governmentscreate comparative advantage? If the UnitedStates were to pursue a more consciously de-veloped industrial policy, what should be theobjectives in the context of a high-technologyindustry like electronics? How might the policytools be formulated and implemented?

Electronics as a High-Technology IndustryThe electrical equipment and electronics in-

dustries have been known for technical leader-ship and innovation since their beginnings atthe close of the 19th century. While progressin electrical equipmentthat which producesor utilizes electric poweris now mostly incre-mental, electronicsreferring to devices andsystems that operate on the information con-tent rather than the power transmitted by anelectrical signalremains a technology in rapidflux. Developments in electrical machinerysuch as practical applications of superconduc-tivity can still promise significant gains in theefficiency of energy conversion and powertransmission; advances in electronics will haveeffects that reach further, and affect the Amer-ican economyindeed, society as a wholemore deeply. An obvious case will be the con-tinuing applications of distributed computing.The impacts will be broad as well af deepmanufacturing industries as a whole will betransformed by applications of electronics toautomated production equipment. Productiv-

ity will rise, the skill mix needed by the workforce continue to shift. In service industries,office and workplace automation will also dis-place people while creating new jobs needingnew s'cills.

Patterns of Development

The portions of the electronics industrywhere American firms remain preeminent arejust those where the pafle of technologicalchange continues to be most rapide.g., com-puters and semiconductors. The United Stateshas been a leader in both the technology andthe science that underly these sectors: elec-tronic properties of solids and the materials sci-ences more generally; electrical engineering;computer science and software engineeringand also in the development of new and suc-cessful commercial products. Nonetheless, al-though Americans have been among the lead-ers in the technology and science of electricalmachinery and electronics, many of the impor-

Ch. 2Introduction 57

tant prewar developmentse.g., understandingof band gaps in solids and the dynamics of con-ductionoriginated in Europe.

The Second World War pushed electronicsto the forefront of engineering science, creatinga momentum that still exists. Developments inradar and computing, both analog arid-digital:proved especially significant.2 Agaii many ofthe advances came from Europe, particular-ly the United Kingdom, where considerablestrides were made in radar technology.3 How-ever. American industry was in a far superiorposition to capitalize on these new technol-ogies in the aftermath of the war. By the late1950's, the United States had what appearedto be an unchallengeable lead in fields such asdigital computers and semiconductors.

Hindsight shows the more temporary natureof this lead, the result of an infrastructure fortechnology and science that emerged from thewar not only intact, but strengthened, coupledwith an industrial base that was likewise farstronger than in countries that had been eitherallies or enemies a few years earlier. The push_created by new technologies, coupled with thepull of war-starved markets in the UnitedStatesmarkets that were eager recipients ofthe products of these technologies, ratherthan devastatedcreated an environment forgrowth and innovation unmatched in the restof the world. Meanwhile, trading partners andpotential competitors such as Japan, Great Brit-ain, and West Germany had to rebuild. Nations

=H. H. Goldstine, 'lir. Computer From Pascal to von Neumann(Princeton. N.J.: Princeton University Press, 1972), especiallyPart Two.

Kraus. "The British Electron-Tube and Semiconductor In-dustry, 1935.62," Technology and Culture, vol. 9, 1958, p. 544.

4111111111W

like Taiwan, South Korea, Brazil, and Mexiconow factors in at least the lower technologysegments of the electronics industrywere,before 1960, simply irrelevant.

Rising Competition

Much of the impetus this strong postwar startgave to the U.S. electronics industry has nowdissipated. Gloomy predictions for the futurecompetitiveness of even the strongest sectors,such as semiconductors, have been heard. Thebusiness press reminds us incessantly that Jap-anese firms captured 40 percent of the U.S.market for 16 kilobit random access mem-ory circuits (integrated circuits called 16K

__IrtAMs), more than 50 percent for 64 kilobit cir-cuits. Market analysts predict that Japanesemanufacturers could have 30 percent of theworld computer market by the end of the1980's.4

In the past, competitors in countries likeJapan relied to considerable extent on electron-ics technology first developed by U.S. firms;now they have independent capabilities andneed not follow paths broken here. As Japaneseelectronics companies have become less de-pendent on American technology, their exportsof microelectronic devices to the United Stateshave grown faster than their imports of U.S.semiconductors. And, where once they ex-ported mostly discrete semiconductors and thesimpler integrated circuits, now firms basedin Japan are exportingor assembling in theUnited Stateslarge-scale integrated circuits

------.-4"No. l's Awesome Strategy," Business Week, Juno 8, 1981,

p. 84.

'Photo credit: Smithsonian Institution

Harvard Mark I electromechanical computer, 169-44

58 2 International Competitiveness in Electronics

(ICs) at the leading edge of the technology.*Japanese computer firms are not yet exportinglarge numbers of systems to this country, butclearly iwend to try. The government-sup-ported fifth-generation computer project is onlyone recent signal of the seriousness of Japan'sefforts.

Needle 7,F to say, this resurgence by America'scompetitofs has not been an overnight phe-nomenon, nor should it be unexpected. TheJapanese presence in consumer electronics be-gan to be felt in the 1950's with the transistorradio; by the late 1970's, firms based in Japanheld, strong positions worldwide in audioequipment, digital watches, calculators, andTV receivers. Their burgeoning capability inhigh-technology electronics builds naturally onearlier developments.

Interactions within the industry often stimu-late technological and commercial develop-ments; ICs have made possible new familiesof consumer products, such as hand calcula-tors. as well as cheaper and more powerfulcomputers. Semiconductor devices are becom-ing indispensable for the products of more andmore industries outside electronics; emissionscontrol systems for automobile engines dependheavily on microprocessors and related de-vices; more than half the cost of an airplanecan be electronics. As one result, electronicstechnologyand particularly microelectron-icshas come to be widely regarded as criticalto a modern, competitive economy, hence ac-cess to this technology a vital strategic.weaponof national industrial policies. Government at-tention to computer industries goes back to theearly years of this technology; a number ofcountries began in the 1970's to subsidize semi-conductor research and development; other:2have, felt applications of ICsrather thancapability for designing and manufacturing thecircuits themselvesto be more important, andhave channeled government funds to this end.

ICs Mcorporate of the, order of hundreds of cir-cuit elements, large-scale ICs of the order of thousands to tensof thousands, very laixe-scale ICse.g., 64K RA Ms, 16 bit micro-processorsof the order of a hundred thousand. See ch. 3.

Technological and Structural Change

The rather complex structure of the electron-ics industry in the United States is describedin more detail in chapter 4. The diversity of theindustry has already been pointed out; thereare more than 6,000 electronics firms in theUnited States. Only a small fraction could legit-imately be called "high-technology" com-panies. But this smaller fractioncompaniesbuilding computers, designing and manufac-turing large-scale ICs, supplying capital equip-ment such as microprocessor development sys-tems or plasma etchers, developing softwarepackages for computersis a driving force forthe rest of the industry, as well as for much ofthe rest of the economy.

By the standards of computers or microelec-tronics, consumer electronics cannot be con-sidered high technology. Yet the manufactureof cathode ray (picture) tubes is a sophisticatedprocess, and TV receivers are now designedaround ICs, some of rather advanced design.Digital TV and digital audi,) are on their wayto commercialization, while consumer prod-ucts are providing some of the first applicationsof speech synthesis; the same will be true ofvoice recognition. Solid-state displays as re-placements for picture tubes are a demandingtechnical challenge. Indeed, the low costs re-quired for practical consumer applicationscreate technological constraints that are, intheir own way, more severe than those im-posed on designers in portions of the industrymore commonly associated with high technol-ogy. At the same time, consumer electronicsproducts such as table radios or conventionalTV receivers are simple enough that they canbe manufa7.tured and marketed competitivelyby firms in industrializing countries such asSouth Korea and Taiwan; the same is true ofmany types of discrete semiconductor devicesand small-scale ICs.

As such examples indicate, even consumerelectronics is changing more rapidly than in-dustries like steel or automobiles. Despite thepace of technological change, electronics is notonly much larger and better established but

Ch. 2Introduction .59

more stable and predictable than, for instance,biotechnology. But again, the industry is farfrom monolithic. Consumer electronics has ori-gins in the 1920's, when radio broadcasting be-came widespread. The computer and semicon-ductor sectors are basically post-World IIphenomena, though many of the leadni8 com-panies--e.g., IBM, Western Electric, Me:4:01:11ahave prewar origins. Thus, the three porl;r1f.of the industry on which this report concentratlis include examples of both well-estab-lished, "mature" sectors, and more volatile,rapidly growing, technology-driven sectors.There are lessons to be learned from each.

One of the lessons that even a superficial lookat the computer industry teaches is the impor-tance of marketing, sales, customer supportand service, and related nontechnical factorseven in a technology-driven industry. IBM hasbeen a dominant force worldwide in comput-ers since the beginning of the 1930's. But IBM'sstrength has beennot only hardwarebutmarketing, software, and customer support. Inmany cases, IBM's competitors have offeredconsiderably more computing power for a giv-en price, but IBM has only slowly lost marketshare because of its many strengths beyondhardware technology. In some contrast, otherU.S. electronics firms have sometimes seemedto rely primarily on advanced technology towin markets. As other countries catch up intechnical capability, a technology-based mar-keting strategy may no longer be enough. Inmicroelectronics, for example, the ability topack many circuit elements onto a single inte-grated circuit chip is still important, but com-petition is more and more a matter of the sys-tems which the ICs comprise or can be inte-grated into. Moreover, as microelectronicstechnology continues to evolve, one path tocompetitive success will be the creation of newend-products incorporating ICs. The skills re-quired for this differ from those needed to es-tablish and maintain leadership in the underly-ing technology, as shown -by_the_examples ofpocket calculators or digital watchesand alsoby the failure of the West German electronicsindustry, which has access to excellent funda-mental technology, to develop into a strong in-

Although many of the major technological in-novations in electronics have originated in theUnited Statese.g., color TV, computer time-sharing, most of the important developmentsin semiconductor devicesAmerican firmshave not always been the leaders when itcomes to product innovations that depend, notnecessarily on new technology, but on productplanning, engineering design, productionskills, and marketing. Although transistorradios were developed in the United States, itwas Japanese products that reshaped the en-tire audic, market.5 Analogous strategiescon-centrating on product design and engineering,originally perhaps imitative, rather than hightechnologyhave led to success by the Japa-nese in fields such as cameras and automobiles.Japanese firms, aided by their skills at low-cost

nia:aufacturing, have recently done much bet-ter at this than companies in the European na-tions with which the United States also com-petes.

It White, "Management Criteria for Effective Innovation,"Technology Review, February 1978, p. 15.

Photo credit: Bell Laboratories


Thus, overall, the cushion that greater tech-nical capability once provided U.S. productsis eroding. And of course, in some technologiesthe United States has never had an advantage.In optical communications, for example, Jap-anese companies have always been near the

.forefront. ade rship_in_ele c tronic s_e quipm e nt_used for certain types of scientific research haslong resided overseasone example being elec-tron microscopes. That this need not alwaysbe a handicap is shown by current develop-ments in electron-beam lithographic equip-

ment. Electron-beam lithography is now essen-tial for making the masks that are, in turn, usedto fabricate large-scale ICs (in a few cases elec-tron-beam lithography is applied directly infabricating the chips). AlthOugh the equipmenthas its roots in technology developed for scan-ning- olectron-microscopes.---virtually all ofwhich are designed and built in Europe orJapanthe United States has not thus far beenhandicapped. Several U.S. firms are, in fact,leaders in electron-beam lithography.

The Importance of CompetitivenessOTA's earlier comparison of steel, electron-

ics, and automobiles provides background andillustrations for many of the questions concern-ing competitiveness, economic efficiency, and"industrial policy that remain of concern to Con-gress, to employees of the U.S. electronics in-dustry, and to the public at large.° The prac-tical meaning of "competitiveness" in the con-text of electronics is discussed in chapter 5. Inessence, the term refers to the ability of elec.':tronics firms located in one country to design,develop, manufacture, and market their prod-uctsdomestically and by exportingin com-petition with foreign enterprises. (For somepurposes, subsidiaries of foreign firms that pro-duce and sell electronics products in theUnited States are considered part of the U.S.industry, but in general the consequences offoreign direct investment must be treated ona case-by-case basis.)

The:competitiveness of an induitry like elec-troniCs is important not only intrinsically, butalso because of interactions with other partsof the economy. Still, there is no -meaningfulway of measuring the competitiveness of an en-tire economy. Competitiveness must be exam-ined on an industry-specific basis, although itcan also be difficult to generalize about an in-dustry as large and diverse as electronics,

Industrial Competitiveness: A Comparigonof Steal. Elec-tronics. and Automobiles.' op. cit.

which for many purposes must be further dis-aggregated.

Considering the electronics industry itself,competitiveness is one of the factors that deter-mines, among other things: employment pat-terns within the industry (size of the workforce, wage levels, skill mix); balance of tradefor electronics products; and the value that pur-chasers of electronics products receive for theirmoney. Electronics products are used by manyindustries-7whether components such as semi-conductors costing a few cents, 'or capitalequipment that sells for hundreds of thousands,even millions, of dollarsand can affect theircompetitiveness. Computers are the mostprominent example, but are far from alone. Nordo they always fill the role of capital equip-ment; many smaller computers are integratedinto more complex electronic systems. Indus-trial process control, scientific equipment, office machines, and communications apparatasare further examples where electronics or electronics-related products can affect the compet-itivenessmore generally, the economic per-

.. formanceof other parts of the economy.

On the bioadest levels, then, the competitive -.ness of the electronics industry affects aggre-gate employment 'levels, trade balances (moreimportantly, the ability to pay for imports), andliving standards. How this industryfares in in-

'Ch. 2Introduction 61

ternational competition influences the types ofjobs available, the country's military strength,and overall rates of economic growth. In turn,the health of the aggregate economy, the qual-ity and quantity of employees available to firmsin the industry, the market provided by the mil--itary, are among the factors that determine thecompetitiveness of American electronics firms.

Ultimately, however, the competitiveness ofany industry in the United States orwhere=depends on thefforts of individualfirms. Policies adopted by the Federal Govern-ment influence these efforts in many ways,often indirectly. Foreign industrial policies arepart of the same.context. Among the more im-portant domestic measures are those dealingwith taxes, Government spending, and mone-tary policy, as well as research and develop-ment (both basic and applied), internationaltrade, and many types of regtilatory policies.Sc:.neticries Federal policies affect only one ora .kiW idustriese.g., regulation of TV broad-casting,Others are broader. Tax treatment.ofincome from overseas investments affectsfirms with multinational operations regardlessof industry. Some policies affect the entireeconomymacroeconomic policies or thosedealing with education.',"=

Generally within the province of individualfirms are factors associated with manufactur-ingincluding costs, the quality and reliabili-ty of finished products, and decisions to man-ufacttire domestically or overseas (offshore as-sembly, wherein some but not all manufactur-

-IndustrialPublic policies that affect competitiveness

can be considered elements of "industrial pol-icy."' The term is intended to embrace FederalGovernment policies of whatever origin thataffect the activities of private industry, particu-larly its competitiveness, productivity, and eco-nomic efficiency.

'Ibid., ch. 8.

ing operations are carried out in other coun-tries to take advantage of low labor costs, iscommon in electronics). The ability to raise ex-ternal capitalwhether equity or debtand togenerate capital for reinvestment through sales,is crucial to firms in any industry, but particu-larly when markets grow as fast as those forsemiconductors and computers. As with off-shore manufacturing, which is favoredby U.S.tarifi laws, sources and costs of capital for elec-tronics-firms-are affected by public policiestax policies and many others, including thoseaimed at controlling inflation.

With respect to consequences of shifts incompetitiveness, employment receives themost attention in this reportboth in terms ofjob opportunities and in terms of the skillsneeded. This and many other topics are dis-cussed, where possible, in the confect of inter-national comparisons drawn between theUnited. States and its trading partners andrivalsusually one and the same. Japan, atpresent, is the home of the strongest competi-tors, in electronics as in many other industries.Japanese firms are likely to continue to be thechief rivals for U.S. electronics manufaCturersover the remainder of the century. But severalEuropean nations have strong technologicalbases in electronics, as well as supportive'gov-ernmental policies. And rapidly industrializingcountries will rise in competitive strength inthe future; TVs from Taiwan and South Koreaare growing factors in the U.S. market.

Policy

The United States does not at present havea coherent or consciously developed industrialpolicy, in contrast to nations such as Japan orFrance. This is not to imply that industrial pol-icies like those of the Japanese are necessarilyeffective in promoting international competi-tiveness, but simply that the United StateS hasnot attempted to develop a coherent industrialpolicy. Instead, policies affecting industries


and their competitivenesshave been formu-lated and implemented on an ad hoc basis. Asa result, industrial policy in this ,country hasbeen fragmented, sometimes contradictory,often inconsistent and lacking in continuity.

These characteristics of U.S. industrial pol-icyreflecting our pluralistic political tradi-tionshave sometimes served the Americaneconomy well, lending flexibility and the po-tential for innovative response to changing cir-cumstance. But the OTA report cited aboveconcluded that this approach to industrial pol-icywhile it might have been well-suited to anearlier period when U.S. industries were rela-tively, isolated from foreign competition, andpossessed advantages in technologyin morerecent years has too often contributed to de-clines rather than improvements in competi-tiveness.

Foreign industrial policies often include di-rect subsidies to industriesperhaps to main-tain employment, or for reasons of nationalsecurity. Export incentives and protection fordomestic industries are common. Foreign in-vestors may face a complex set of carrots andsticks. Cooperation among nominally compet-ing firms may be encouraged. Governments insome countries have engineered "nationalchampions" in attempts to increase competi-tiveness. Restrictive business regulations may

be relaxed, government procurements chan-neled to favored companies, which in somecases may be publicly owned. Nationalizedenterprisesan increasing presence in sectorslike banking or energy production although nota major factor in electronics--couple industryand government even more tightly.8 Americanbusinessmen increasingly complain of the diffi-culties involved in trying to'l compete with suchventures, which need not make profits, or mayhave unusually long profit horizons.

The variety and complexity exhibited bypresent-day national industrial policiespartic-ularly the difficult questions of when govern-ment support measures should be judged subsi-dies that distort international tradehave ham-pered efforts by international organizationssuch as the General Agreement on Tariffs andTrade (GATT) to fit remedies for many of thepossible means of "unfair" competition into thebody of international trade agreements. As oneresult, bilateral agreements are becoming morecommonexemplified by the Orderly Market-ing Agreements negotiated by the U.S. Govern-ment to control imports of TV receivers fromseveral Far Eastern nations.

For a survey, see R. P. Nielsen, "Government-Owned Busi-nesses: Market Presence, CoMpetitive Advantages and Ra-tionales for Their Support by the State," American Journal ofEconomics and Sociology,"vol. 41, 1982, p. 17.

IssuesAs emphasized above, a vast number of Fed-

eral Government policies in some way affectthe international competitiveness of the U.S.electronics industry. Among the more impor-tant are:

Government support for commercial (asopposed to military) R&D, ranging fromtax policies intended to increase levels ofresearch spending or encourage commer-cialization to direct support;trade policies dealing with exports as wellas importse.g., the ways in which meas-

-tha fin° caiorfrnnirC inch ictry

fit within the overall framework of U.S.foreign economic policy, the meaning of"reciprocity" for an industry like elec-tronics, barriers to investment in foreignelectronics industries;Government policies affecting capital for-mation for the economy as, a whole, and,more directly, the ability of firms in theelectronics industiy to generate and attractcapital for expansion;regulatory policies that may affect thecompetitiveness of the U.S. electronics in-dustrye.g.. antitrust enforcement;

Ch. 2introduction 63

the availability of enough people with ade-quate levels of education and training, par-ticularly engineers and skilled_workerssuch as technicians, as well as the Govern-ment role in supporting technical educa-tion;economic adjustment policies intended toencourage shifts of resources from declin-ing industries to thOse with better pros-pects for future competitiveness, and toaid workers, communities, and regionsthat have suffered because of shifts in in-ternational competitivenesse.g., in con-sumer electronics.

These examples all involve complex issues,with effects that may differ among variousparts of the electronics industry, and from firm

to firm. The remainder of this report attemptsto deal with such complexities; at the sametime, of course, public policies continue toevolve and changewitness the 1981 tax act,or the expiration in July 1982 of the OrderlyMarketing Agreements covering imports of col-or TVs from Korea and Taiwan. The objectiveis not to be exhaustivebut selectiveto try todifferentiate the factors influencing com-petitiveness in electronics that are primarilyunder the control of managements of individ-ual firms from those that are strongly affectedby the Federal Government, and to examine thelatter in the context of a high-technology in-dustry that has been one of the mainstays ofU.S. competitiveness during the postwarperiod.

CHAPTER 3

Electronics Technology

Contents

Overviow

Consumer Electronics. ...........Semiconductors.TransistorsIntegrated CircuitsSystem Design and the MicroprocessorMicroprocessors and Nlemory.Learning Curves anti Yield ...Integrated Circuit DesignManufacturing Integrated CircuitsFut ore Developments

ComputersTypes of Computer-,Computers as SystemComputer Son \,vareGrowth of.Computing PowerApplications of (:oinputers

Summary and ConclusionsAppendix 3A-Integrated

Circuit Technology

Appendix 3B-Computer MemoryAppendix 3C-A Process

Control Example

List of Tables

Tahlo1. Examples of Solid-State Technologies

Used in Information Processing2. Cost Increases for Fine-Line Lithography3. Characteristics of Drferent Categories of

Digital. Computers4. Increase in Computing Power

Over Time5. Comparison of IBM 650 and Fairchild

F-8. Microcomputer6. Characteristics of Generations of

Computer Technology7. Typical Applications of

Computing Capability3A-1. Common Termina'ngy and

Classifications of Integrated Circuits3A -2. Principal Types of Memory Circuits313-1. Computer Memory

l'ago67 List of Figures68 _Figure No.

71727273747fi7778

1. Simplified Diagram of TV ReceiVerComponentry 69

2. Comparative. Feature Sizes ofMicroelectronic Devices Such asIntegrated Circuits 71

3. Controller for a Microwave Oven Basedon Single-Chip Microcomputer 75

4. Increases in IC Integration Level 755. Schematic Learning,,Curve for the

Production of an Integrated Circuit or82 Other Manufactured Item 7683 6. Ranges in Cost and Time for Design and85 Development of Integrated Circuits , 7787 7. Steps in Designing and Manufacturing an90 Integrated Circuit 78

928, Step-and-Repeat Process of

Photolithographic Pattern Formation on,Silicon Wafer 79

93 9. Typical Microcomputer Intended forPersonal and Small-Business Applications 84

99 '10. Typical Minicomputer InstallationIncluding a Pair of Terminals and a

102 Printer 8411. Data Processing Installation Built Around

General- Purpose Mainframe Computer 8412. Simplified Block Diagram of a

Microprocessor System 84Page 13. Elements of a General-Purpose Digital

Computer System 8572 14. Relative Hardware and Software Costs80 Faced by Users of Larger Computer

Systems 8683 15. Minicomputer Price Trends 89

16. Parallel Decreases Illustrate How Costs of88 Computers Depend on Costs of

Semiconductors 8988. 3A-1. Silicon Wafer Showing Patterns

Formed by Photolithography 9790 3A-2. Integrated Circuit Chip Before and

After Final Encapsulation 9891 3B-1. Performance of Computer Memory

Alternatives 1003C-1. Portion of Process Control System for

a Chemical Plant 10394.96

101

CHAPTER 3

Electronics Technology

OverviewThis cllapter outlines the technology on

which the consumer electronics, semicon-ductor, and computer industries depend, cov-ering them with enough depth to provide back-ground for discussions later in the report con-cerning the role of technology as a force oncompetitiveness. Except for occasional ex-a mph competitiveness itself is left to laterchapters.

The primary function of electronic compo-nents and systems is td manipulate and trans-mit information in the form of electrical sig-nalseither analog or digital. The transmissionand utilization of electric power are integralto these processes, but constitute only second-ary functions of electronic equipment. Even inthe case of a 50,000-watt radio broadcasting sta-tion, the high power simply increases the areacoverage of the information in the signal. Theinformation manipulated and conveyed via anelectronic system can range from a simplesequence of numberse.g., a zip code, or thebalance in a checking account to the soundsconveyed by radio, images such as televisionpictures or weather maps, or the informationcontained in radar or sonar signals.'Changes in electrical voltage are the most

common carrier of information in elecAronicsystems. In an analog system, the signal takesthe form of a voltage or some other electricalparameter that varies continuously over arange, while digital information is encoded inthe form of a string of binary "bits." Each bi-nary hit can take on one of a pair of discretevalues, again usually voltages. The magnitudeof these is unimportant, so long as they can bedistinguiShed from one nother.A bit can bevisualized as having values of "0" or "1," or" +" as opposed to " In a digital circuit,the signal normally takes the form of a string

of discrete voltage levelse.g., any voltage be-tween 2 and +1 volts might represent abinary "0," any value from +2 to +5 volts, abinary "1 "

Regardless of the simplicity or complexity ofthe information content in a signal, either ana-log or digital technology can, in general, beemployed. The choice turns on the practicaladvantages and disadvantages for a given ap-plication. A complex system may use analogcircuitry for some tasks, digital for others. Ingeophysical exploration, for instance, an ana-log signalessentially a mechanical pressurepulse or sequence of pulsesis transmitted intoa geological formation. The reflected pulsefrom the subsurface strata is sensed by trans-ducers analogous to microphones. These trans-ducers respond to the mechanical energy of thereflected pulse by generating a proportionalanalog electrical output. Analog-to-digital con-verterstypically integrated circuits (ICs)then convert these signals to digital informa-tion that can be processed and analyzed by adigital computer.

Both analog and digital technologies have aplace in the three sectors of the electronics in-dustry covered in this report. But while mostconsumer electronics equipment is still basedon analog technologyradio and TV receivers,phonograph records, magnetic tape players--virtually all computers process information indigital form. At the same time, computer pe-ripherals such as terminals and printers con-tain analog circuitry, while digitally based con-sumer products are becoming more common.Semiconductor devices come in both analog(often termed "linear") and digital varieties. Afew ICs combine analog and digital circuitryon the same "chip."


Consumer ElectronicsThe most common consumer electronic prod-

ucts are radios, TVs, and audio eqUipment suchas "stereo" systems. Electronic toys and games,electronic watches, pocket calctilators, andhome computers are other familiar examples.These are all "systems" in the sense that theycontain more than a single electronic compo-nent, but some are much more complex thanothers. An electronic watch may, consist of lit-tle beyond a single IC and -a displayitself asolid-state deviceplus a battery. Televisionreceivers contain several hundred componentS.Video cassette recorders (VCRs) are complexinechrmically as well as electronically.

Radio broadcasting provided the foundationfor the development of the consumer elec-

. tronics industry. Despite a real cost muchhigher than today, there were well over 10million radio receivers in use in the UnitedStates by 1930, with annual sales exceeding $1billion.'

Research and development (R&D) on televi-sion began in the 1920's, with limited broad-.casting prior to World War II. Large-scale com-mercialization had to await the end of the war,but by 1949 5 million TV sets were sold in theUnited States all black-and-white. Color tel-evisionfor which most of the early work wasperformed by RCAfollowed the next year, butcolor TV sales in the United States did not passthe 5 Million mark until 1967, and first ex-ceeded black-and-..vhite sales in 1972.2

With more than 11 million color sets sold in1982, and about half as many black-and-whitesets, the TV receiver remains the largest sell-

. ing consumer electronics product, accountingfor nearly half the dollar value of consumerelectronics sales in the United States (ch. 4).Monochrome TV sales have been rather static.for a number of years, with the market for col.or sets expriding only slowly; cable TV anddirect satellite reception may spur future sales,but much of the-growth in consumer electron-

ics markets is now in new generations of prod-ucts, notably VCRs. Still, in many respects.e.g., the relatively standardized design approaches and critical importance of productioncostscolor TV continues to typify consumerelectronic technologies.

Television signals 'ire broadcast via ampli-tude modulation of a high-frequency carriersignal, much like AM radio. But the bandwidthrequirements for TV are far greaterabout 6MHz, versus 10 KHz for AM radio. Bandwidth,which is expressed in terms of frequency-6MHz being equal to 6 x 10° cycles per second,10 KHz to 10 x 103 cycles per second--is ameasure of the rate at which information canbe conveyed, hence must increase with theamount of information in a signal. The pictorialimage in a TV signal has a much higher infor-mation content than sound, hence television'shigh bandwidth requirements. In principle, theanalog information in either a radio or a TVsignal could be conveyed in digital form with-out changing the bandwidth requirementsgreatly.

A home antenna receives the amplitude-mod-ulated TV signal at a microvolt level (1 micro-volt equals 10-8 volts). To produce a visual im-age, this signal is amplified to control an elec-tron beam which scans the front of the picturetubeor cathode-ray tube (CRT)forming anew image 30 times each second (the number-..sf frames per second can vary abroad). The cir-cuitry in a TV receiver is quite complicated (fig.1) and now entirely solid state (the chassis in-cludes both discrete transistors and ICs) exceptfor_the picture tube. The CRT is the most ex-pensive single component in the set, account-ing for about 40 percent of the cost. Produc-ing picture tubes fs a highly specialized activi-ty; smaller firms oiler, buy CRTs from manu-facturers such as Zenith or RCA.

Conventional picture tubes are not onlybulky and expensive, but account for much ofthe power consumed by TV sets; in Japan andFiirrinn nartiridaricr rrin evi mar altar+ rrit,i,c

Ch. 3Electronics Technology 69

Figure 1.Simplified Diagram of TV Receiver Componentry

Antenna,

elalma

Microvoltsignal

Tun17er

1

Amplifiers

Audiosignal

I

Videodetector

Horizontaland

verticalscanningcircuits,

colordecoder

Videosignal

Videoamplifier(s)

Amplifier

Speaker

00

1111.

Powersupply

Picturetube

SOURCE: Adapted from K. (ed.). Radio Engineering Handbook, 5th ed. t Yew York: McGrawHill:1959). ch. 22.

only reduces consumer electrical bills, but canlead to more reliable operationone of the ad-vantages that TVs imported from Japan haveenjoyed over the years (ch. 6). The drawbacksof conventional picture tubesprincipallybulkhave stimulated considerable R&D aimedat flat-screen television displays. Flat screensare not yet practical, but continued progressin solid-state technology will doubtless lead toeventual success.

Before 1960, most of the significant technical

vances have come from Japan as weli.3 Televi-sion technology is now well diffused interna-tionally, and no one country appears to enjoya technological advantage. Product innovationscontinue to come from U.S. firms, but alsofrom other parts of the world. European con-sumers, in particular, are often attracted by

rabid., p. 27; also "International Technological Competitive-,

ness: Television Receivers and Semiconductors," Charles RiverAssociates Inc., Boston, Mass., draft report under National Sci-ence Foundation grant No. PRA 78-20301, July 1979, app. 2A.


new and different product features such asmultiple image displays (several channelsshown simultaneously on one screen). -Thesehave become important to product develop-ment strategies of firms in Western Europe.Continued progress in large-screen projectionTVs has been stimulated by competition be-tween Japanese and American producers forwhat could be a large new market. Japanesefirms have been leaders in reliability, and mayhave tended to emphasize R&D directed atautomationand at rationalization of the man-ufacturing process in generalmore thanEuropean or American producers (ch. 6). Inparticular, Japanese TV makers were leadersin adopting transistorized chassis designs inthe late 1960's and in automating the insertionof discrete components such as transistors, ICs,capacitors, and resistors into printed circuitboards.

Although TVs still account for much of theconsumer electronics market, they are a ma-ture product in the sense that most Americanhomes already have one or more. Thus, thegreat proportion of saleS are now supplementsor replacements. The market for VCRs, in con-trast, is expanding rapidly (ch. 4). U.S:' sales ofVCRs during 1980 Nere less than a millionunitsall importe i; sales nearly doubled in1981.

Video recording on magnetic tape' was pio-neered in the United States by Ampex.4 Al-though Ampex and RCA continue to manufac-

lure video tape recorders for broadcast applica-tions. consumer VCRs were developed largelyby Japanese firms (ch. 5)which now buildabout 95 percent of the world's. V.CRs. InEurope, Philips has a few percent of the mar-ket. hut all the VCRs sold in the United Statescome from Japan.

""Interactions of Science and Technology in the Innovative,Process: Some Case Studies.-. final report. Battelle ColumbusLaboratories. National Science Foundation Contract No. NSF-C667. Mar. ch. 12.

While the commercialization of VCR technol-ogy by Japanese manufacturers is one sign thatJapan may be taking over product leadershipin consumer.electronics, video disks thus farpresent a mixed, perhaps contradictory; pic-ture. The optical video disk systein developedin Europe by Philips reached the consumermarket first. In the Philips system, a laser readsthe digitally encoded signal on a spinning disk;microscopic depressions in the disk representbinary "Os" or "1s." While an elegant technicalachievementand one with potential for high-density digital data storage of other types (e.g.,in conjunction with computer systerns)theoptical video disk sold in the United States byMagnavox has not been a commercial success.RCA's video disk, introduced early in 1981,functions on analog.-principlesmore like aphonograph record. Yet a third system, devel-oped in Japan by JVC, may eventually reach themarketplace. As compared to VCRs, disk sys-tems are cheaper but can only play back, notrecord. While the technology is evidently inhand, it is too early to tell how large the marketfor home video disk players will bee.g.,whether it will rival that for VCRsor whichsystems will survive in the marketplace.

A number of trends in consumer electronicse.g., recent introductions DI "component"TVs analogous to component stereo systems,along with games and low-end home comput-ers that use a conventional television as thedisplaypoint toward the eventual develop-merit of more-or-less integrated home enter-tainment systems. Such systems might in,corporate TV and audio recepti .-'n and repro-duction, including various kinds of informa-tion services, along with applications ,Df com-puting capabilitynot only games, but record-keeping, home security systems, control ofhousehold appliances, and regulation of heat-ing, ventilating, and air-conditioning systems.Such developments do not depend heavily ontechnological advances except as low produc-tion cost is necessary for mass market accept-ance.


SemiconductorsStrictly speaking, the term "semiconductor"'

refers only to the mat;/lids from which semi-conductor devices are made. Such materialshave electrical conductivities intermediate be-tween good conductors like copper and insula-tors such pis glass. Silicon is the most commonsemiconductor materialvirtually all ICs, andmost discrete transistors, are based on silicOn.In this report, the term "semiconductor" willbe low,* applied to the products of the "semi-conductor industry" as well as to the materialsthat are the starting points for these products.The broader designations "microelectronics"or "Microe lectroi tic devices" include Semicon-ductor productswhich have replaced vacuumtubes in nearly all applicationsas well asother types of solid-state devices that process,manipulate, or display information.

The most familiar example of a vacuum tubeapplication that-solidtate technology has notyet been able to match is the CRTnot onlythe common TV picture tube, but the displayscreens of computer terminals. In other cases,solid-state devices are not only much smallerthan vacuum tubes, but cheaper, more rugged;and longer lasting. They also offer higher op-erating speeds; indeed, on almost any measureof performance, microelectronic devices offerorder-of-magnitude improvements over thecomponents they have, replaced. Modern digi-!al computers would be quite impossible with-out semiconductors._ Although a computerotherwise like current models could, in prin-ciple. function with tubes instead of ICs, sucha machine would fill a building and probablynot execute a single program without one or1110L'i! tubes failing. Solid-state circuits havemade practical many electronic systems thatwould earlier have been too big, too costly, orotherwise in fact unthinkable.

Although virtually all commercial microelec-tronic products are now made from semicon-ducting materials. considerable R&D has beendevoted to classes of solid-state technologieswith tmlential for transtrnittina and nrnoccirm

conventional semiconductor physics. Such de.vices might function on magnetic or opticalprinciples, rather than being strictly "electron-ic," although the materials involved are some-times semiconductors.* Boundaries betweenelectronic, magnetic, and optical technologiestend to blur as device technologies move to-ward microstructural and submicrostructuralsize ranges. (Microsauctural sizes are largecompared to interatomic distances but smallcompared to objects that can be easily seen andhandled, like an IC chip itself; the feature sizesof microelectronic devices are currently in therange of 1 to 10 micrometers, or less than atenth the diameter of a human hairfig. 2.) Be-cause solid-state devices based on magnetic oroptical principles are often used as co m-ponents in systems that are broadly electronicin nature, no fine distinctions will be made.

Optical data transmission can give bandwidths much higherthan electronic signals; this, along with the low raw materialcost. is one of the advantages of optical fibers. Systems based'on laser light sources: with optical fibers for signal transmissionand thin-film integrated optical devices for signal processing.could tepla.:e many types of electronic circuits and systems.

Figure 2.Comparative Feature Sizes ofMicroelectronic Devices Such as Integrated Circuits

Representativefeature sizes Dimension in(design rules) Examples micrometers

Human hair 100

1978 5 micrometers .1980 2 micrometers .

1985' 0.5 micrometer

1990'. 0.1 micrometer

`led blood cell 7

1 micrometer = 10 6 meterProjected.

Yeas: cell

-Smallest bacteria 0.2

Polio virus 0.01

0.001

Mom 0.0002

SOURCE: AJaated !'rom G B. Larrabee rharartnriyntinn <".1"

72 International Compotitiveness in Electronics

Among these solid-state devicestable tare:

transistors, ICs, light-emitting diodes(LEDs), all of which are semiconductors;bubble memories; which depend on themagnetic rather than the electronic prop-erties of materials;liquid crystals (as in alphanumeric dis-plays for watches or calculators), chemi-cals that change colors when their temper-ature changes;integrated optics, in which information istransmitted and processed in the form oflight. Integrated optics and prospectivefuture technologies such as organic semi-conductors are not commercially impor-tant at present, but could have impacts onfuture success in internation& r:ompeti-Hon.

Transistors

While most of the fundamentals of semicon-ductor physics were known prior to World WarII, the transistor itself was developed at BellLaboratories after the war, and first demon-strated in late 1947.5

In contrast to passive electronic devices suchas resistors, capacitors, and inductorswhichcan only respond to electrical signalsactivecircuit elements like transistors control and

W. Shockle, ihe Path to the Conception of the-JunctionTransistor." !EEE Transactions on Electron Devices, vol. ED-23,1976. p. 597; E. Braun and S. MacDmald. Revolution in M:nia-ture: TI .0 History and Impact of Semiconductor Devices (Cam-bridge, Mass.: Cambridge !University Press, 1978). ch. 4.

regulate the flow of electricity in a circuit. Asa result, they can amplify electrical signalsafunction that earlier could only be performedby vacuum tubes.

Transistors come in many varieties to servedifferent functions, just as for the vacuumtubes they superceded. To make a transistor,a semiconducting material such as germaniumor silicon is "doped" with small amounts ofelementsarsenic, boron, phosphorusthat lo-cally affect its conductivity. The transistor, tothe naked eye, is then just a small piece of, say,silicon with two or three wires attached. Infact, however, the purity, chemical composi-tion, and perhaps crystal structure have beencarefully tailored on a microscopic level.

Integrated Circuits'An IC is made by fabricating several circuit

elementstransistors, capacitors, and such-on a single substrate. Integrated circuits wereindependently developed in the 1950's at TexasInstruments and Fairchild Camera and Instru-ment, still two of the leading selniconductormanufacturers in the United States (Fairchildis now French-owned). The two companies ap-proached the problem quite differently during1958-59, but their developments shared thecommon characteristic of an ICtwo or moredistinct transistors fabricated on a singlesubstrate.° Thus they were monolithic circuits.

°M. F. Wolff, "The Genesis of the Integrated Circuit," IEEESpectrim. August 1976, p. 45: Braun and MacDonald, op. cit.,ch. 8. The depths of the transistors fabricated on a chip are smallenough that ICs can be considered two-dimensional. Often thesilicon substratea few millimeters on a side and less than amillimeter thickis called a chip, as is the resulting circuit.

Table 1.Examples of Solid-State Technolcgies Used in Information Processing

Technology and examples Description Current status

Semiconductor,,electronics: Depends on electronic properties of semicondectin,s Production

Transistors materials such as silicon, germanium, gallium arsenide.Integrated circuits

Magnetic devices: Depends on magnetic rather than electronic properties ofBubble memories materials.

Bubble memories in limitedproduction.

Solid-state optics: Depends on electro-optical properties of r-aterials, some of LEDs widely used for displays;

(sometimes called which are semiconductors. integrated optics

optoelectronics) experimental.

. Light-emitting diodes LEDs are lighted whan a current passes.

Ch. 3Electronics Technology. 73

Other types of ICs can also be builte.g.,hybrid or thin-film circuitsbut monolithicdevices comprise the bulk of production.

At present, the market for ICs is more thanfour times the size of that for discrete semicon-ductors. Because of this, and because verylarge-scale circuits pace the industry and area major focus of international rivalry, thisreport gives much more attention to ICs thanto other microelectronic devices.

Appendix 3A discusses IC technology insome detail. The significance of the technologyitself for international competitiveness resideslargely in the commercial advantages that canaccrue from innovative and/or widely acceptedchip designs (ch. 5), as well as from masteryof processing technology. Being first on themarket with a new design gives a firm the op-portunity to build market share before competi-tors can offer similar products. The advantageof a particularly well-accepted design is thatit may become a de facto industry standard.A manufacturer whose design becomes sucha standard not only has the assurance of a rela-tively large end stable market, but also the pros-pect of a broad range of licens:ng and/or sec-ond-sourcing agreements. The firm may alsoget a headstart in the competition to design thenext generation.replacement. Processing capa-bility is just as important to competitive suc-cess as design, because advanced circuit de-signs are often limited by what can be built atacceptable yields. (The yield is the fraction of"good" circuits coming off the production line.)In semiconductorkadvances in circuit designand in processing capability are iniArdepend-ent to a greater extent than in almost any otherindustry.

The first of the two major types of ICs to bedeveloped, bipolar, has been replaced for manyapplications by MOS (metal oxide semiconduc-tor, app. 3A). Over the course of the 1970's,MOS technologywhich is denser, cheaper,and consumes less power, but which does notofier speeds as high as bipolar--became domi-nant fnr larop-cralp intporatinn fT.ST1 Vpry

MOS. Finns that were slow to master MOStended to fare poorly in sales growth and prof-itability over the past decade. For the foresee-able future, competition in ICs will continueto center around MOS devices.

System Design and the Microprocessor

In designing a digital system, the engineerhas several options:

1. to assemble a number of standard logic ch.:-cuits like those of the transistor transistorlogic family described in appendix 3A;

2. in cases where large production volumesare anticipated or performance;require-ments are specialized and demanding, tocall for one or more custom ICs;

3. to use a standard microprocessor or micro-computer with software written for theparticular application.

Assemblies of standard logic circuitstypi-cally small- or medium-scale ICsmay be eco-nomical in limited production volumes, despiterelatively high design and development costs.In such cases, the system is implemented inhardwarei.e., its functioning can only bealtered by changing the circuit componentsand/or their interconnections.

Specially designed custom circuitsanalogas well as digital, bipolar as well as MOShavea place in high-volume applications rangingfrom consumer products like TVs and electron-ic watches to telecommunications systems.They are also employed where performance re-quirements such as operating speed cannot bemet in other wayse.g., in some military andaerospace applications, or in mainframe com-puters. Custom circuit design is expensive:hundreds of thousands of dollars, sometimesrunning into millions. Here the designer is alsoworking in hardware, but new custom hard-ware rather than standard ICs.

In contrast, when a system based on a micro-processor or microcomputer is designed, thelogic is implemented largely through software

P a rnmnntpr nrnoram ctnrpri in mpmnry


microcomputers can function as central proc-essing units for general-purpose computer sys-4emssuch as the small machines sold by Ap-ple or Radio Shack---they were originally conceived as replacements for custom ICs to cir-cumvent the high costs of designing, develop-ing, and producing custom parts. The firstcommercial microprocessor was introduced byIntel Corp. in late 1971 to implement the arith-metic functions in an inexpensive calculator.Faced with the request of their Japanese cus-tomer for igroup of custom chips to be usedin a line of calculators, Intel instead proposeda simple 4-bit microprocessor chip.' Ratherthan hard-wiring the opera ions required forIho different calculator modelsaddition, mul-tiplication, printing, and so onsoftware pro-grams permanently stored in memory imple-mented these functions. Money was saved indesign and production compared to the customIC alternative.

Subsequent experience .ias shown that mi-croprocessors may prove the low cost alterna-

'R. N. Novice and M. E. Hoff, Jr.. "A History of MicroprocessorDevelopment at Intel Corporation." IEEE MICRO. February1981, p. 8. Parallel developments took place at Texas Instru-ments.

I

r",

,:-.!it : .

i.

,:r.-:.tvirv-,.':-:-.i-c'c-v-,,.t-,,;,,,,t.,-,-,-ii. ..- . s_, , .

t-vv-?,..f77,,,,r . ___

,-..,,-' t.----.:77,..r:, t-,,,,,,.7754,,,,-,43., , , 1,

.....:.... i-,. ;,..,..:,..-;: irf, k..1 a ; 11121._

..1411iigkaii. rigiej II'

live for systems that could be built with as fewas two or three dozen standard logic circuits.Figure 3 illustrates a typical application of amicrocomputer, control of a microwave oven,where the chip contains memory for programstorage along with a processor. in such an ap-plication, production volumes might be highenough to justify a custom LSI chip designtens of thousands of identical parts, at a mini-mum, are normally called for. But the micro-processor/microcomputer alternative has a bigadvantage in flexibility; design changes aresimple, different models simply need differentprograms. And, in the microwave oven exam-ple, there is no need for high performance.

Before the advent of the Microprocessor, sys-tern designers had only two choices: assembliesof standard parts or custom ICs: The micro-processor/microcomputer introduced a thirdoption, one that proved highly attractive. Asa result, several hundred.different models ofmicroprocessors and single-chip microcom-puters are now marketed (many differ only indetails), and custom microprocessors are some-times designed for special applications.

Microprocessors and MemoryMicroprocessors cannot be used by them-

selves; they must be supported by other chips,at a minimum for program storage. Memorycircuits aredescribed in some detail in appen-dix 34, particularly table 3A-2. Memories, infact, comprise the largest single market cafe-gory'for- ICs; the majority go into general-pur-pose computer systems, but large numbers arealso used in dediCated applications of micro-processors and microcomputers (i.e.. applica-tions where the computing function is invisi-ble to the ,user of the system).

Technological progress is easier to measurefor memory circuitse.g., RAM chips (randomaccess memory)than any other type of IC.Densities have increased steadily over time

/ figure 4while the cost of a chip has remained/' roughly the same. As a result, the_cost per bit

of information stored goes down. This opens


Figure 3.Controller for a Microwave Oven Based on SingleChip Microcomputer

Line power

Display Touch panell

Status hod Displayindicators drivers

8-bitA-D con-

verter

Powersupply

Clock

Alarmdriver

Magnetrontriac driver

Door-interlock

AlarmOven heating]element triac

SOURCE, IR Walker, "Analog/Digital LSI," 1979 Elecfro Professional Program, New York, Apr. 24.26, 1979.

Figure 4.increases in IC integrain:Level

10

tos

c 1040

10.5)

0.1'1960 1965 1970 1975

Year

Oueyssac, "Projecting VLSI's Impact on Microprocessors," IEEE:lpecltum. May 1979, p. 38.

1 chipcaiculator

32-bit microcomputer

256K RAM-64K

16K RAM.

4KRAM 16-bit

RAMmicrocomputmicrocomputer

.8-bitmicroprocessor

-..

Medium-scale integration

Small-scale integration

SOUttcr.

*!980 1985

rely on memory-7-notably microprocessor sys-tems. LikeWise, as semiconductor memory be-comes cheaper it will continue to substitute foralternative storage media such as magneticdisks in general-purpose computer systems. Itwas widely noted in the early 1970's that thecost per bit of memory had fallen below thecost of a jelly bean. According to some esti-mates, a jelly bean (at 10 may buy as many as1,000 -(1K) bits of memory by 1990.

Meinory and microprocessors are the mostvisible products in domestic and internationalcompetition. While it would be wrong to con-sider these,the only important categories, theydo constitute half the 0total market for ICs.Moreover, significant advances:in both devicetechnologies and proCess technologies haveoften found their way into production via cir-.cuits of these types.-

83.

76' International Competitiveness in Electronics

Learning Curves and YieldsMaking ICs is demanding, more so at higher

Itu, els of imegration. Forty or more processingsteps may be required for a VLSI chip, a figurethat will continue to grow. Designing VLSI cir-cuits is also complex, and becoming steadilymore time-consuming and expensive, whileproduct design and process design go hand-in-hand. For consumer electronic products likeTVs, decisions on product features often hingeon the costs of production; Mr a new IC, thefirst question is: Can it be made at all?

Once a semiconductor firm has designed anIC and carried it through the pilot productionstage they can normally assume that produc-tion costs, even if initially high because of lowyields, will decrease over time. Figure 5 is aschematic learning curve, sometimes called anexperience curve, showing cost declines withcumulative production volume. Learningcurves typical of IC manufacture show thatwhen cumulative production doubles, costsdecrease by about 28 p-rcent.9

Learning curves as in figure 5 apply to man-ufactured products of many kinds, but their im-pacts on pricing decisions have been par-ticularly noticeable among sernimnductor

--firms; they are a major factor in forward-pricingsettinp prices below the initial costproduction to gain market share. Be,-:ause firmsfeel confident that costs will decrease as pro-duction experience accumulates, forward-pric-ing has been .a common competitive tactic.

These cost declineswhich can be consid-ered equivalent to increases in productivitystem from much more than simple learning orexperience by the labor force; other causes in-clude better equipment performance and utili-zation, greater understanding and control of

8"Boom Times-Again for Semiconductors," Business Week.Apr. 20, 1974. p. 65: A Report on the U.S. Semiconductor Indus-try (Washington, D.C.: Department of Commerce, September..1979), pp., 48-50. The 28 percent figure is an average from which''the cost experience for a given IC can deviate substantially. Pro-duction volumes typically rise rather slowly at first, because ittakes time for customers to design the new part into theirsystems. In comparison to other types of manufactured prod-Ids, learning curves for semiconductors are not particularly

steep, but continue to fall over very long production rims.

Figure 5. Schematic Learning Curve for theProduction of an Integrated Circuit or Other

Manufactured Item

Cumulative number producedSOURCE: Office of ".::_nnologt Assessment.

the many steps in the production process,smoothing of work flows, and perhaps changesin the design of the part itself. Control of theprocess is particularly important, and often de-pends on subtle variations in parameter, influ-encing phenomena such as diffusion, etching,or polymerization (of photoresistsIC fabrica-tion steps are described .in more detail below).In inany cases, the physics and chemisiry ofsuch phenomena are only poorly understood,and cannot be modeled theoretically; processcontrol models tend to rely on empiricism,hence on accumulated experience. Denser andmore complex ICs call for a better grasp ofprocessing fundamentals.

In general, learning improvements show upas increased yield, the percentage of chips thatpass final test and function satisfactorily. Whena new IC goes into production, the yield is gen-erally low perhaps only a fraction of a per-centbut rises as experience accumulates andprocessing can be better controlled. As a rule-of-thumb, products are seldom marketed, ex-cept for sampling purposes, until yields haverisen to about 10 percentwhich may take asmuch as a year of production-line experiencesFor mature products, yields can rise to wellover 50 percent. Increased yields are a power-

°R. Bernhard, "Rethinking lila 256.kb RAM," IEEE Spectrum,.May 1982, p. 45.


ful force in driving down the costs and pricesof ICs; in effect, doubling the yield hay--is theproduction cost.

Because processing capability is so criticalto commercial success, the specialized equip-ment used in fabricating ICs is one of the keysto competitive ability. Much of this equipmentis designed and built by independent suppliersmany of them American firms =- selling tocustomers throughout the world (ch. 4). Thetech::Jlogical capability and competitivenessof the portion of the U.S. electronics industrythat designs and manufactures equipment hasbeen just as important to the international posi-tion of the United States in semiconductors asthe efforts of semiconductor firms themselves.Because of the interrelations of device technol-ogies and process capability, and the depend-ence of costs and yields on process control, anumber of the more important steps in produc-ing ICs--beginning with circuit designare de-scribed in more detail below.

Integrated Circuit 1114sign

The task cf the circuit designer is to definean arrangement of circuit elementS7---transis-tors, capacitors, logic gates, interconnectionsthat will satisfy the functional requirements ofthe IC. The more complex the circuit and thehigher the level of integration (a 64K RAM con-tains more than 100,000 circuit elemP-Its) themore difficult the designer's job, and . :,;gherthe cost of design and development, .lure 6illustrates ranges of development time and costincluding hardware, software, and peripher-al chipsfor several types of ICs.

Asa rule-of-thumb, circuit design costs his-torically have been rather stable at about $100per gate. Thus, a microprocessor with 10,000gates will have a hardware design cost of per-haps $1 million, and may represent 10 man-years of effort. Sof:ware costs add to this. Itmay well be possibe to make chips with 1 mil-lion gates within a few years, but the costs ofdesigning them will be prohibitive unless thedesign costs on a per-gate basis can be reduced;one estimate has been that an IC with a densi-ty of a million devices would require about 200

Figure 6.--Ranges In Cost and Time for Design andDevelopment of Integrated Circuits

35

30

25

20

15

10

5

16-Oit microprocessors

8-bitmicroprocessors

4- to 3-bitmicrocomputers

Customcircuits $20 million for

Intel 8086(1979)

5100.000 Si million SIO million $100 millionDevelopment cost (dollars)

SOURCE: P. M. Russo, "VLSI Impact on Microprocessor Evolution. Usage, andSystem Design,"-IEEE Transactions on Electron Devices, vol E1127,1980. p. 1339.

man-yearrs for desirna and development usingthe conventional methods of the past decade.10Computer-aided designi.e., the use of special-ized computer programs by the design engi-neersis the principal hope for cost reduction,and is increasingly necessary just to handle thelogical complexity as the number of devices perchip goes up. Designing-a microprocessor-with---100,000 transistors would be irr:,-actical with-out computer aids. R&D aimed a: more regulareven modularchip designs is also under. -

. y, again intended to reduce the time, hencethe cost, of IC design. Modular approaches areparticularly attractive for custom logic circuits.

As pointed out above, the microprocessoritself originated as a way to reduce the costsof custom circuit design; in essence, choosing'a microprocessor means replacing hardwaredesign by software design, .and in many caseslowers costs. But as logic complexity goes

I°C. L. Hogan, cited by F. Ogden, "Audience Gives Mixed F,e-ception," Electronics Weekly, Mar. 28. 1979, p. 5.

85

78 In ternat o nal Competit;veness in Electronics_ .

the software costs for grog: . ning the micro-processor escalate rapidly. in part, this simplyreflects the more sophisticated processorse, g., it 18-bit rather than a 4- or 8-bit deviceneeded for demanding applications. Althoughsoftware production can also be automated, the0.vo basic paths toward implementing logichardware via custom chip design, or softwaretia a standard microprocessor with a programembodying the logicwill continue to com-pete. Design cost, flexibility, and performanceare all factors. But if computerized design aidsfor hardware and software advance sufficientlyfar and in tandem it may eventually make lit-Ile difference. perhaps even to the designer,whether -the-logic is embodied in hardware or

ftwa re:

Although much of the engineer's work re-volves around the logic that the circuit will im-plement, IC design also calls for intimateknowledge of processing and fabrication."fig-tire 7. Designs that can be implemented inn-MOS might be impossible in c-MOS (see app.3A, table 3A-1, for an explanation of the typesof MOS devices). One company might haven -MOS process capabilities beyond the reachof another. The desirn team must consider fac-tors such as the spacing between transistorsand the widths of the lines that interconnectthem. In contrast, when designing a system tobe built from discrete components it was often

11P..W. I. Verhofstadt, "Evaluation of Technology Options forLSI Processing Elements," Procee&r;gs of the IEEE, vol. 64.1976. p. 842: C. Mead anti L. Conway. Introduction to VLSI Sys-tems (ReadinA, Mass.: AddiSon-Wesley, 1980).

enough to be familiar with the performancecharacteristics of off-the-shelf devices. At thesame time, designerstypically electricalengineersmust be at home with the logicalconcepts and ciftl.vare orientation of the com-puter scientist; the need for software skills willarm., ICs come more and more to resembleintegrated syst6ms. This melding of hardware(including process technology) and softwareskills, and the rapidity of technical change,make unusual demands on the people who fillsuch jobsone reason that the electronics in-dustry has been experiencing shortages of qual-ified engineers (ch. 8). Although neither the cir-cuit designer nor the process specialist can befully conversant with all aspects of IC designand manufacture (fig. 7), some knowledge ofeach is needed.

Manufacturing Integrated Circuits12

The design process culminates in a patternor layouta large drawing, several hundredtimes the size of the circuit itselfthat mustbe translated into the "tooling" for producingthe chip. In simple terrils, the procedure formaking an IC resembles a series of photograph-ic processeslithographic patterns are createdin layers on a silicon wafer. Each layer is madeby exposing a polymeric chemical called a pho-toresist to light or other, radiation, the light.

"See, in general, F. W. Vollmer, "Manufacturing Process Tech-nology for MOS VLSI," VLSI Electronics: Microstructure Sci-ence, iol. 1, N. G. Eins.pruch led.) (New York: Academic Press,1981), p. 1.

Figure 7.Steps in Designing and Manufacturing an Integrated Circuit

Science Processingtechnology

KII

off

ManufaCture TestDesign

technologyDesign

(product and proces.$)

Productdefinition

t Designverification

SOIJIICE: G. Moore, "VLSI: Soma, Fundamental Challenge!," (ZEE spect-um, April 1979, p.30:

80


passing through a grid-like mask. as shownschematically in figure 8. More than a dozensuch- masking steps may be needed tc build upa VLSI part. Many other processes besideslithography are involved in IC fabrication, withmore detail given in appendix 3A, but lithog-raphy is critical for future increases in circuitdensity. While advances at many stages in themanufacturing process take place in an inter-dependent vaye.g., laser annealing is replac-ing furnace annealing because it does a betterjob of restoring the crystal structure of thesilo ::on which is disturbed by ion implantationlithography is the major factor in determin-ing how Email individual devices and intercon-nections can on a production as opposedto laboratory basis. Already, transistors can bepacked much more closely in an IC than neu-rons are packed in the human brain; it is the

Lens

Figure 8.Step-andRepeat Processof Photolithographic Pattern Formation

on Silicon Wafer

Filament' Light source

Radiation

Lens

Mask

Photoresist

S'iicoil wafer:

Stage traverseSOURCE: Office of Technology Assessment.

technology of lithographic processing thatmakes this possibie.

I lographic line widths in production ICshave been reduced an order of magnitude overthe past decade, from about 20 micrometers inthe early 1970's to 2 to 4 micrometers current-ly. (A micrometer is about 40 millionths of aninch.) Thinner lines give higher operatingspeeds as well as denser packing. The 16KRAMs designed in the early to mid-1570's werebased on 5 micrometer "design rules, -64KRAMs on 3 micrometer rules (design rules,which are directly related to lithographic linewidths, comprise- the full set of geometric con-straints that designers follow). The next-gener-ation 256K RAMs are based on 1.5 to 2 m.crometer design rules.13 Continued progress inreducing line widthsmore generally, featuresizeis thus a major driving force in movingfurther into VLSI. Forihis reason, a principalR&D target of the Defense Department's VeryHigh-Speed Integrated Circuit (VHSIC) pro-gram has been lithographic technologyan ob-jective paralleling commercial R&D efforts, onere nson the program is likely to have a positiveeffect on nonmilitary portions of the semicon-ductor industry. The VHSIC program goals in-clude two stages of lithographic improvements:the first stage calling for line widths of 1.25micrometers; the second, lines of 1 micrometerand below.

While feature sizes of 1/2-to 1 micrometer arewell above the range for which the physics ofelectron devices will begin to cot:strain per-formance, such feature sizes co demand signif-icant developments in lithographic capability,particularly for mass production," In the past,

""Rethinking tie 256-kb RAM," *n. cit. On design rules, seeMead and Conway, cp. cit., p. 47.

"1. E. Sutherland, C. A. Mead, at- 7 E. Everhart, Basic Limita-tions in Microcircuit Fabrication hnology, Defense AdvancedResearch Projects Agency Report R-1956-ARPA. November 1976;R. W. Keyes, "The Evolution of Digital Electronics TowardsVLSI," IEEE Transactions on Electron Devices, vol. ED-26, 1979,p. 271. The ultimate limit to reductions in the sizes of electrondevices will perhaps be thermal noise, althougi: a variety of prac-tical concerns may intrude first. Feature sizes are likely to de-crease to the range of 0.01 to 0.1 micrometer before fundamentalphysical limitations are encountered.

, 80 Internationa; Competitiveness ,n Electronics

visible lightgenerally ultraviolethas beenused to expose ohotoresists (fig. 8). But evendeer.) ultraviolet, which has a wavelength of

hoot '2 micrometer, cannot produce linewidths much below a micrometer because op-tical considerations limit the lines to abouttwice the tvavelength of the radiation.

ro achieve 1 Ancrometer lines with visiblelight requires very sophisticated lithographicequipment--positioning and layer-to-layer reg-istration of the sequential masking steps must

held to a small fraction of a micrometer. Asingle machine of the direct-step-on-wahr typediagramed in ligure 11 now costs about half amillion dollarstable 2. As the table shows, theearlier generations of equipmen ;placed bydirect-step-on-wafer machines we.e much lessexpensive. Filter patterns will be still morecostlywhether the technology of choice iselectron-beam lithography (table 2), X-rays, orion beams (see app. 3A). The rapidly risingcosts of IC processing equipmentwhether forlithography, for ion implantation, or for testingalong with higher costs for design and devel-opment, are the mo,4 important causes of therapidly increasing capital intensity in the semi-conductor industry (ch. 7). Entry costs are nowroughly $50 million, versus $5 million to $10million in the early 1970's.

Some have already moved to electron-beam lithography for ci itical circuit laYers.Electron-beam lithography is also a routine toolfor making masks. X-rays and electrons havewavelengths much less than light, and so offergreater resolutionat the expense of high firstcost for the equipment, and low p:fedv.',:tionrates.15 Because of its importance drivinl.: iC

,,(;. lt. br,.,r. lio, Resolution Lithograph,..-Tocimningy Aficroolectronic Fabrication. G. R. 1-.et.,-r.(New Vtrk: Acarlun ;r: Precs, 111110), p. I.

Photo credit Ga.! Corp

Electron-beam lithography system

technology, R&D on high-resolution lithograph-it-, techniques (see app. 3A) has been a principaltarget er g,-,vern merit-funded programs in othercountriese.g., Japan's VLSI projectas wellas the VUSIC program funded by the U.S. mil-itary.

Table 2.Cost Increases for Fine-Line Lithography.

Approximate capital requirementsLine width Throughput Approximate cost for production capacity of 1,000

Lithographic,system (mi(:.nmeters) (wafers hotr.) per system water starts per week

LightContact printing 10 60 $15,000 $30,000Projection 2-5 60 $240,000 $400,000Direct -step on-wafer 1-2 30 $480,000 $1.6 million

Electron;beam 0.5-1.0 6 '1)25 million,OUNCE Adapter' tram A J Stein, J Marlay, and R. Ma::r), -Toe Impact of VLSI on the Autcmobile of far,: r).»,

2. N G Eatsprtica fed ) (New York' Academic Press, 1981), p 295.

4.

Elef.:tronic M,crost-t;cturo vo!:


More cost conies with the "clean rooms..needed for VLSI processing. Even micrometer-size dust particles can ruin the lithographic pat-terns: cleanliness is vital to high yields. In aclean room. the air is filtered and people must

. wear special clothing. As circuits become dens-er. and feature sizes smaller, not only is cleanli-

. ness more important, but the whole range ofprocessing equipment used in making ICs be-comes more sophisticated and expensive, add-ing to the capital requirements for semiconduc-tor manufacturing, a matter discussed in chap-ter 7.

Future Developments-;einiconducto,. devices need not be based on

silicon. One alternative is gallium arsenide, amaterial that offers considerable potential forimprovements in packing density and speed(me to two orders of magnitude compared tosiliconhut is still largely a laboratory technol-ogy. Whether gallium arsenide circuits will be-come commercially important depends onrates (*improvement compared to silicon, andalso on developments in other prospective -t.ich-n:..;logies.e.g., Josephson junctions. Josephsondevi,:esalso experimental, and much furtherfrom demonstrated practicality than galliumarsenide ICspromise still better speed

Iso illustrate the importance of speed, as wellas power consumption, consider the technol-ogy embodied in a current-generation main-frame computer. The central processor for onesuch computerthe Amdahl 470-Vemplbys1,680 ICs, with a total of about 100,000 logicgates. As is typical in large .computers, thechips use silicon bipolar technology to givehigh computing speeds. Replacing these bipo-lar chips with .2.-Jlium arsenide may offer thepotential for increasing corn.:.:atat:ional speedsby a factor of 10 to 100, and ra,.iocing the powerconsumed by the processor from about 3,000watts to perhaps 30 wattsless than most lightbulbs.I6 Comparable improvements other ap-plications of ICs carry implications for coMpet-itive trends in many industries if s.;,-,iae com-panies or some countries manage a headstartin reducing such technologies to practice.

"R. C. Eden. H. M. Welch. R. Zur ., and S. I. Lang, -The Pros-pects for Ultrahigh-Speed VLSI GaAs Digital Logic," IEEETransactions thl Electron Devices, vol. ED26. 1979. p. 299. About1 percent of 'he electricity consumed in the United States nowgoes to computers, mosily for .cooling"CBEMA Predic-tion: Say Energy Crunch Could Cut EDF Growth Pate 50%,"Electronic News, Mar. 17, 1980, p. 32. On Josephia... junctions.see J Matisoo, "Overview of Josephson Technology Logic andMemory," HIM Journal of Research and Development, vol. 29.1980, p. 113.

CerriolreeSComputer tecin. ;logy has many roots,

clticli .g militar, needs during the SeconcWirlc VVar for fire control tables and othercomplex a ndior repetit:,fe computations. TheUnitcd States had no great advantage overother nations during the early development ofcomputers; significant innovations also origi-nated in several European countries, particu-larly Great Britain.17 But as computing technol-

r*Gaps in Techou'oey: Electronic Computers (Paris: Organiza-tion for i:ooperation and Development. 1909), p. 61.Ft. a concise :ilimmory of develetnnentsduring thefirst threeizeneratioi. of com;.citing technology, primarily in UnitedStates, see S Rosen, "Electnnnic Computers: A 11 iitorical Sur-vey." Computing Surveys, vol. 1, 1969. p. 1.

cgy progressed, the lead swung decisivelyoverAmerican firms, much as happened over

roughly the same period of time for semicon-ductors.

The Bureau of the C".7.' =us was an early non-.military custorner American computers,census ds, procssing requirements remain-ing a typical example of computer ru:ilications.When a Univac I NAP,s delivered to the Bureauof the Census in It some observers pre-dicted that the market for digital computersmight eventually total a dozen; a few yearslater, when :;ales to private industry began, theestimatez were tle.t the potential market in the


United States consisted of perhaps 50 corpora-tions."

Needless to say, as computer technology ad-vanced many flu': > firms became customers.and the ranks at computer manufacturersswelled. Among the entrants were a numbeeof companies that had become established inthe office. equipment market. InternationalBueiness Machines Corp., Burroughs, and Na-tional Cash Register (now NCR) joined firmslike Univac that had been set up specificallyto manufacture digital computers. By the endof the 1960's.. computer applications hadspread well beyond numerical computationsand data processing. The great part of com-puting power is still devoted to data processingfor business and governmentaccounting,sales, production, inventories, recordkeepingof all kindsand to scientific and engineeringcalculations. in addition, many individual com-puters, mostly microprocessors and microcom-puiers, now perform "invisible" functions inapplications ranging from alto raft flight con-trol systems to the microwave oven exampleshown in figure 3.

The spread of computing power has some-times been technology-driven, sometimes driv-en by user demands. Technology-driven devel-opments arose when more computing capabil-ity was available than people knew how to useproductivelyi.e.. before the eiiplications werewellefined. Under these circumstances, theavailability of more powerful machines orgreater performance per dollar tends to gener-ate new applications, or. more broadly, serveneeds earlier unmet. As in many instances oftechnological change, what the technologycould do. at any given time and for a given cost,evolved in conjunction with applications, withone or the other ter-poi arily in the lead. Muchtlie same has been true for ICs. In the period'when demand from military and 'space pro-grams in the United States was high, themarket deeve the technology; but leaders in the"semiconductor industry heve periodically wor-ried tbh.t applications for the full capabilities

"I.. M. liransco111, .'f.',1ect conics and Ccanputf.,, An 0:ivr-.ccience, VOL I?, 1qn2, p. 755

of new circuits containing greater numbers ofdevices, now VLSI, might not appear. Suchfears seem to have vanished from the computerbusiness, though not the perennial questionsof which firms will get the largest share of thenew markets.

Types of ComputersAs pointed out in the earlier section on "Sys-

tem Design and the Microprocessor," the essential elements of a small digital computer (ex-clusive of power supply and input/output de-vices) can be placed on one IC to create a sin-gle-chip microcomputer. A microprocessor ismore limited in function, but when combinedwith the necessary memory and peripheralchips cn a printed circuit board becomes asingle-board microcomputer. From such prod-ectse---selling for around MO without cabinetsand other auxiliariesdigital computers rangeupwards in size, speed, and cost to "supercom-puters." Intended for complex scientiLe andtechnical calculationse.g., modeling theEarth's atmosphere, designing airfoilsnuclear weaponssupercomputers are mhdeby only a few manufacturers, and cost in thevicinity of $10 million each.

In between board-level microcomputers andsupercomputers come a number of broad cate-gories of machines: personal and small busi-ness computers like those made by Apple; min-icomputers of various types; general-purposemainframes. The latter, typified by many ofIBM's larger models, can handle many differ-ent tasks at once. Table 3 outlines some of theconventional distinctions among eeese cate-gories. The differences are not always clea:-cut and will blur even moE2 as microcomputersbecome more powerful, competing power stillcheaper, and computers of all types more ver-satile.

The central processing unit (CPU) for - im-j computerin fact, just a -ticroproce:is shown schematieally in fif:et.e 12. A micro-prcressor functions like the CPU-a: any com-puterit brings information (in the form of bi-nary bits) into an arithmetie logic unit, manipu-lating the bits in accordaree with instructions.

Ch. 3 Electronics -Technology 83

Table 3.Charac!,1ristics of Different Categories of Digital Computers

Family" Distinguishing features

Microcomputer. The central processing unit (CPU) consistsof a microprocessor or single-chip n.,:.focomputer, some-times several. The most common microcomputers use an8-bit word and sell, without peripherals but otherwise com-plete and ready to operate, for under a thousand to a fewthousand dollars. They will typically fit on a desk topfig. 9and do not require special training to operate. Ex-amples include popular models sold by Apple and RadioShack, along with the IBM Personal Computer. -

Machines based on microprocessors or microcomputerchips with 16-bit word iengths are beginning to appear, par-ticularly for business applications. These are nearer in per-formance to low-end rninic-3mputers than to the 8-bit micro-computers originally ,oped for the hobbyist and per-sonal computer markets.

Minicomputer: Microcomputers, by the definition above,could not have existed before the development of themicroprocessori.e., before the early to mid-1970's. Mini-computers, in contrast, stem from the 1960's. A popularearly mini introduced in 1905 the PDP-8, built by DigitalEquipment Corp.was the first tow-cost, mass-producedcomputer of any type. It was designed around a 12-bit wordand discrete transistors.

Minicomputers are small compared to mainframes,which can fill a room. As fig. 10 indicates, minicomputersare often about the size of a desk. Although not as portableas micros, many minicomputer models can be mt.:led rela-tively easily within an office or factory environment.

Minis found much of their market as dedicated proces-sors designed into more complex systems, or in specializ-ad data processing taskse.g., industrial controllers,data acquic.iiiion systems for laboratory research, inventorymanagement in factories. While such applications remaincommon, minicomputers are also dely used for general.

SOURCE 014,ce of Technology Assessment

purpose data processing. Often mainframes were neededin such application's only a few years ago; minicomputerstend to supplement rather than displace them.

At the lower end, it is increasingly difficult to distinguishminicomputers from the more powerful microcomputers.Many less expensive minis now rely on a single-chip proc-essor. However, he smaller minicomputers typically-use16-bit wordse.g.; the currently popular POP-11 modelsmade by Digital Equipment, or the Nova series of Data Gen-eral. Larger, more powerful machinessometimes called"superminis"normally have a 32-bit word length. Exam-ples of superminis are the Data General Eclipse series orthe VAX models of Digital Equipment. In the 1960's, 32-bitwords were found only in mainframes.

Most minicomputers carry prices in the $10,000 to$100,000 range. A principal distinction between minicom-puters and mainframes is that minis seldom. require eitheroperators with a great deal of training or specially con-structed facilities. Mainframes, in rnt!'.?. usuallybe permanently installed; some large computers dissipateso much heat that air-conditioning is needed even in mid-winter.

Mainframes: The CPU for a mainframe typically contains sev-eral thousand logic chips, usually bipolar for speed. Wordlengths are commonly 32 to 64 bits. Mainframes often sup-port multiple terminals and peripherals fig. 11 and gen-erally require trained personnel onsite.

While i BM is the world's largest producer of mainframecomputers, more than a dozen other firms build machinescomparable in computing power. Mainframes can sell for$10 trillion or moreexclusive of peripheralsbut themore popular general-purpose machines typically costunder $5 million.

from a program stored in memory, and endsinformation back to the memory or to an out-put device. The stored program, whichmadethe modern digital computer possible by pro-viding a means for telling the computer whattry cic without the need for hardware cl .;nges,accounts for a good deal of the information thatenter:; and leaves the CPU. Even rather simplecomputer systemsfigure 0typically are sev-eral types of memory. .wh' are desci bed inmore detail in appendix 3B.

Computers as Systemsedition to CPU and memory, a computerkteeds and outp,q devices (fig. 13),

srnal comput==rs, all the components ma}he integrated inV, a siugie' self-donteined unit.

Peripherals such as disk and tape drives, ter-minals, and print7-3 are made by large num-bers of independent vendors, as well as by com-puter manufacturers. Nei:11;y 90 Americanfirms were producing terminals as of 1980,about half of these "smart" by virtue of em-bedded microprocessors, while nearly 30 hadannounced their intention to build 8-inch Win-.chester disk drives, a product just beginningto reach the n;larketplace at that time.19 Themarket dynamics associated with the computerindustryrapid growth, intense competition,new entrants with new productscharacl,rizeperipherals and &ware as well as processors.

""The Digithl Age," Electronics. Apr. 17, 19:A. p,.:487; G. Sluf..sker, "28 Rivals Eye 8-Inch Disks But None Lands F. OEMYet,' Fitic::r!,:nic News, Jan. 21. 1960, p. .1C.


12)

re S.Typical Microcomputer Intended forPersonal and Small-Business Applications

Photo credit: Apple Computer, Inc.

Figure 10.Typical Minicomputer InstallationIncluding a Pair of Terminals and a Prinier

1.110f0 credit: Digital Equipment Corp.

Figure 11.Data Processing InstallatioL atilt AroundGeneral-Purpose Mainframe CompAer

Photo credit. Control Data Corp.,

Figure 12.Simplified Block Diagram of aMicroprocessor System

Addiess bus(to callinstructions anddata)

r

Power I

ROM 0 t*-

RAM}

CPU

unit :

Registbus I

Databus(to and frommemory)

--Controi-logic

Clock

--I

1

1

1

1

1JI/0 (input/output) bus(external control,D /A. A/D)

D/A = Digital/analog conversionA/D = Analog/digital conversionROM = Read-only memoryRAM = Random access memory

SOURCE Office of Technology Assessment.

Cn. 3Electronics Technology a 35

Figure 13.Elements of a Gonoiiil.Putpose DigitalComp, Aar ::,ystorn

I

Output

Terminal(

Tape. disk drives

Centralprocessing

unitic..;PU)

Cachememory

Mainmemory

Auxiliarystorage

Pr -.3r

Tt,,h119:oy,y

Technological change in computing has beenrapid since the }ieginning of commercial pro-duction in the 1950's, but now the industry isperhaps facing the most comprehensive set ofchanges yet. These stem from "distributed in-telligence," the dispersal of computing powerto many farflung locations. In some respects,this trend began with the development of time-sharing in the early, 1960's. Time-sharing per-mits users at i'emote terminals to interactdi-rectly-with a central processor, extending thecapabilities of a powerful computer to manypeople simultaneously. It also uses the proc-essor more efficiently. Even during big jobs theCPU maybe idle much of the time; with time-sharing, system software keeps the CPU busyby dividing its processing poWer among manypeople, each of whom is unaware of the others.

Conceptually, the next step beyond time-shcring---for which each user needs only a

"dunib' terminal (an input/output device withno fu .coon other Liao. to communicate withthe central processor)is to link a central com-puter to satellite machines which can share theprocessing load. Many sir h distributed proc-essing schemes are possibie, among the morecommon being a mainframe -',.i.pported byminicomputers. A mainfranie or mini car alsocommunici,::- with "smart" terminals ..:atcarry out computations, compile pro-grams, and u'inerwise relieve the central or hostcomputer of some of the work. Point-of-sale ter-minals found in retail stores often function asparts of distributed systems. The terminal notonly acts as a cash register, but sends data onpurchases to a central computer that can

,aman-

age ventory, compare sales volume by brands,and provide other information to managers.Automatic banking machines are another fa-miliar example; each automatic teller functionsLs a smart term; linked to the bank's cen-tral computer(s). These systems nay inch: fiehundreds of machines spread over severalStates.

Networking is a related term, referring to dis-persed machines that communicate with oneanother but are each autonomous. Any one ma-chine can transmit data to any other; controlof the network may be distributed over the sys-tem or may reside in a designated processor.In some but not all cases, netwo7'.ed computersnot only communicate and share control, butalso share the processing load. Local networksserve a limited group of users, such as a singleoffice. At the other extreme, a multinationalcorporation might link computers located inmany countries to form a worldwide network.

Computer Software

Physical equipment, o. .rd'.vare7--rangingfrom ICs, to disk drives, to netNti orks--has beenthe p 'Inary subject above. But modern com-puters. depend just as heavily on software. Theprograms /hat stand between 'user and CPU tell

hardware what to do. Arrayed in severallevels, they range from applications softwarewritten in langua!!os such as Fortran or Cobolthe only type of program that the typical userever sees--to operating systems that supervise

9?

56 in!er.lat;onal Com,-)etitiveiess rr, Electronics-

arid iinate both hardware' and soft.vare;Aim', !us, It is the softwarearchitecture, oper-at in e system, compilersthat allows complexnetworks- of computer and communicationscomponents to control steel mills, regulate airtraffic, determine the path of a guided missile.distribute secial security checks. Hardwareand software in conjunction determine systemperformance, and customers weigh both as-pects when making purchase decisions. Insome cases this may entail buying software andhardware from different v endors and assem-tiiing a unique system. The spread of distrib-uted intelligence, new applications of cqmput-ers in liouws and offices, shopfloor automation,computer-aided engineering analysis and de-signall depend more heavily on versatile, reli-able. user-friendly software than on hardware.

Since the begins tugs of large-scale commer-cial production, computer hardware has be-come steadily cheaper relative to software,Costs for hardware have decreased by a fac-tor of at least 1.000, holding processing powerconstant, over the past 25 years.20 In markedcontrast, software costs have not decreased ap-preciably, and may even have risen in realterms. A single line of programing, as a . :le-a f-thu mb, costs in the range of SIO to $50in terinflation, about the same now as in 1955.21 As

computt.r Business tVee.k Sept. 1, 1980,p 41, ,Lf: iMprGvement depends on the type

ur -e'-asioned.pry-inc.:I.:it:: in programii g--;is measured in lines of

le per in of timehas pc,,babiy increased, the rata of in.ease has barn orders of magfAtUde slower than for hardware

is fa uanan«,. fur instance, points out that cchile pro-gre-rn r pro,luctivity has increased by about a factor of 3 since

- y,,-emp,Tforinaiva, to-cost ratios have gone up by roughly2r the -,arne brut; ind--J. S. Birbaum, -Computers: A

!i,ir%;'. :rends and Limitations,"Science, Feb. 12, 1382. p. 760.In s, i!ie uniru tasks. productivity has probably remained

rather stable -perhaps CI,V11 tit'Utt!atied. Applications program-may now be sornowhat more efficient because of improve-

news in higher level languages. Productivity in systems pro-gra Ming, or developing software for dedicated microprocessors.tnicrocomputers, and minicomputers, has probably not improvedas rapidly; when ,,sterns become more complicatcAl, many ofIht in program developmentfrom conceptual design to(101,411..;golg---!)Icomo man! arduous. Even a relatively simple pro-};I:t111 MO' have of the orde r of 1020 different execution paths,depending on the number e, r loops, branches, and subroutines.Cost,: per line cuts escalate program size and complexity in-

; ease. Another common . ale-ofahumb is that a man-month ofeffort is required to demonstrate that 100 lines of coy .1 is. forpractical purposes, errur"Hsr and functionally correct,

9 ,1

a result, the largest part of the total cost to theuser of a large computer system. is now soft-ware, rather than hardwarefigure 14. Thechart appli is to both purchased softwarefromcomputer manufacturers or independent ven-dorsand to user-developed programs; soft-ware maintenance is also included. At onetime, many computer manufacturers providedsystem software such as control programs,language processors, and utilities free to hard-ware purchasers. Now, separate charges arethe rule. For example, IBM currently sellsabout Si billion worth of software per year, ac-counting for a little over 5 percent of the firm'stotal revenues; in newer systems such as theIBM_ 4300 series, nearly half the price of a typ-ical installation is for software.22 Similarly,more than half of the R&D commitment of atypicil computer firmmeasured either interms of total expenditures or in terms of man-powergoes toward software.23

Cost trends for the deyelopmt,nt of 'imE.warefor dedicated applicationse.g., the logic foran embedded microprocessorare sirni r.Even the simplest such application will requr edebugging and testing of the program to verithat it functions as desired. Software develop-ment for a microprocessor application may

31Missing Computer Software," op. cit.""Computer Technology Shifts Emphasis to Software; A Spe-

cial Report," Electronics. May 8, 1680, p. 142.

Figure 14.Relative Hardware and Software CostsFaced by Users of Larger Computer Systems

100

8

Hard,', are

60

AO

Software20

0

1960 1980 2000

Yam

SOURCE- Office of T1 ,,notogy Assessment.

Ch. 3L-7f:ctronics Technology 87

Photo crecla TIktronlx, rric

Microprocessor development system

cost several hundred million dollars, withestimates for 1985 running to S3 million or

The rising relative costs of software havebeen one factor in the rapid growth of inde-pendent firms that develop and market pro-grams of 111 types. Many conioter manufac-turers have iraditionally been re.= ter liz:rdwa re-oriented, leaving an attractive market for ven-dors who concentrate. on software (see app. C,-Computers: A Machine for Smaller Busi-nesses,- on the role of "systems houses" in thedevelopment of the minicomputer market).Even IBMwhich has built its market domi-nance in larger machines on software as much.as hardwarehas turned to independent soft-ware firms to supply programs for its personalcomputer. Independent software vendors sell

,""Missing Computer Software," op. cit. One supplier of rnicro-procasso, s and microcomputers has estimated that a typicalrnich1970"!'. applumm III carried a software development cost ofabout 520.000 ($2t) per line of code), but by 1980 the cost was5100,000 to half a dollars ($35 per line, of code,but also many :non, lines). Meanwhile the b "rdware costs haveremained r.bout the sans, -in the vicinity of $100 per unit. SeeJ. G. rasa, "Inb.; Takes Aim at the '80s." Electronics, Feb. 28,1980. p.

perhaps Si billion in ,ff-the-shelf programs per .year, and twice that amount in custom pro-graming.25

GroWth of Computing Power

14 gave one picture of the rapidity ofcl in computer technology, and in theco. !ter industry in general. But change ex-tends far beyond the relative costs of hardwareand software, the rapid growth in the micro-computer market (now about 50 percent peryear), or continued increase in performance/cost ratios for computer systems. And, whiledistributed intelligence ma eventually havebroader and deeper effects on the way peoplelive and work than big machines, the absoluterise in computing power delineated in table 4illustrates simply :Jut dramatically how rapid-ly the capabilities of the most powerful digitalcomputers have increasednine orders ofmagnitude since the close of the Second WorldWar, six orders of magnit-Ide in the 30 yearssince the introduction of uae first commercialmachine, the Univac I. All the computers listedin table 4 would be classed as mainframes, Andthose of recent years as supercomputersrep-;resenting the maximum in computing poweravailable at a given time.*

While the biggest computers have been grow-ing in speed, smaller machineslike all com-putershave been growing in performance perdollar. Table 5 compares an 8-bit single-boardmicrocomputer r,Tresentative of 1970's tech-nology to the.IBM 650a first-generation vt,c-uum tube processor of the mid-1950's. The twomachines are roughly comparable in comput-ing power, but the modern microcomputc:. isorders of magnitude Smaller and cheaper,

z,ty. 0. Gardner, "The Key to Greater Productivity." Dun'sReview, August 1980, p. 74. The total value of computer soil-ware in Use worldwide probably exceeds $200 billion,

'Arithmetic operations per socond, the measure used in thetable for comparing computi71g power, is not a perfect yardstickbecause many data-processing programs are limited by opera-tions other than arithmetic.a.g., inverting matrices. More so-phisticated comparisons employ -benchmark" programs basedon repraso..tative tasks. Arithmetic operations as used in thetable have the advantage of beim; easy to understand and ap-plicable to early Model computers. some of which could not ex-ecute modern benchmarking pi.ograms,

6'3 `^ti;-,,,-,,,,Vional Competrt, i.±_-,ness in Electra-rics

Tab

Year

1944

-1.Increase in Comp Ming Power Over Time

ModelHa^:ard Mark I

ielectro

2omputational speed(arithmetic operations

per second)

0.41946 451951 Univac I 2701953 IBM 701 61519t1,1 IBM 7074 33.7001963 Ci:/C 360C 156,0001965 IBM 360/75 1,440,000

1971972

CDC Cyber 176Cray 1

9.100,00080.000.000

1981 CDC Cyber 205 800.000.000fi;,!. Technologv Forecasting Literature Emergenne arid Im-;r Inrcvatton." P. Kelly and M. I:ranz.trri; 'ids I.

inlo.sfol. A Critical Review of Current ftii,ow,,:lgo-ar.c.sco San Francisco Preis. 1978i. p. 300; "Tne CI:g.tal Age."

Frecricanics, -Apr 17, 1980. p 382. P. J Schuyiten. "Me. Ba rite in Super-Krea Yor-ii Times. July 22, 1980. p. Dl

andat least as significantvastly rnoreNote that these computers are s1parated

circuits 'l;. exhibit reliabilitiesmeasuredas mean timer., beRveen fa. , Nof the order of 10'' hours!gate. Thus. a type al ink. .,wiessor containing 10,000 gatesmight have a mean time between failures of about 10 million

cr ;moo years, In cont :.ast, mean times between failuresfur discr,!to tran,INtors are :moot 108 hours, for vacuum tubes,leas than I0" huufs. Sue, S. Midde1hoek, B. Angell, and D. 1.

"%fin:oprocessors Get Integrated Sensors." IEEE.covtrinn. February 19B0. p. 42.

in time by only two decades. Figurc 15 givesan alternatic9 p:-.:ture of growth in perfor-mance per dollar. The plot shows the declinein price for a minicomputer family after 1965the pioneering PDP-8, although the first severalyears apply to an earlier modelthe rapid fallstemming in part from learning curve phenom-.-na as for semict:1ductor devic=:.!s. Drops inprices for the semiconductors a machine.con-tainsfigure 16also lead to cost reductions.Digital Equipment Corp. introduced the PDF-8at $18,000; by the early 1970's some versionswere price.. as low as $2,500.20

"Generations" of computers can be distin-guished based on advances in the technology.For example, the IBM 650 (table 5) representsa first - generation machine, the F-8 microcom-puter third generation. Zeroth - generation sys-tems were similar to the 650 in using vacuumtubes, but early computers such as Er.iaclacked the ability to execute stored programs,the hallmark of the modern digital computer.To change &program in Eniac mean: altering

"C. Lewis. "Small Computers," Ei;rtronic News (an. 25, 1902,sec. II, p. 70.

Table 5.Comparison of IBM 650 (1955) and Fairchild F-8 Microcomputer (19/0's)

Physical volume (ft') .....

Weight (pounds)Power consumption (watts)

IBM 650270

5,65017,700

F-8 Remarks

0.01 F-8 about 30,000 timessmaller'

2.5

Memory (bits) 3K main, 16K ROM, EK RAM101.)K secondary

CPU 2,000 vacuum tubes 20,000 transistors

Time for adding twonumbers (microseconds).. 750 150

Reliability (iitean timebetween failures) Hours Years (3 million to 10 million hours F-8 at least 10,001) times

is a typical mean time betweenfailures for a current microproc-essor--More than 300 years--butthe srbsystems with which themicroprocesscr communizatese.g., terminals, printersmay bemucl, less reliable)

Cost $200,00. (1955 dollars) ti!ider $1,000 with terminalsouncr 18M 650 mlormation horn... "1978 First Quarter and Shareholders Meeting Report," Texaa Instruments, Inc ,trcnild F-8 infoTrha.tIon fron1, G. Linvnll and

C L iogan, -Intelkctua1 and Economic Fuel for the Electronics Revolution," Science, Mar, 18, 1977, p. O.

F-8 consumes 7,000 timesless power

650 alsn needed manydisgrE,te resistors a.nocapacitors

more reliable

Ch. 5Competitiveness in the International Electronics Industry 187

development programs, and now have chosento sell Japanese VCRs under their own brandnames.

The point is that, once having lost productleadershipas has occurred with VCRsAmerican firms find it increasingly dif-ficult to compete in new technologies, and mayeventually find themselves importing or adapt

ing other products as well. Because U.S. man-ufacturers cannot expect cost advantages, theymay be left with only their distribution systemsand brand recognition as prime competitiveweapons. To a considerable extent, Japanesefirms have already countered these advantages:thus, the long-term prospects for Americanfirms in consumer electronics do not appearbright.

SemiconductorsTechnological forces have dictated the mar-

keting strategies of semiconductor companiesin all parts of the world since the inception ofthe industry. The case study on 4K RAMs inappendix C points to the importance of engi-neering capability for U.S. merchant firmssuch as Mostek or Intel. Technology is no lessimportant now than a decade ago, when the4K RAM was being developedbut as late asthe mid-1970's the business strategies of foreignsemiconductor manufacturers were of little in-terest to Americans. As the 4K RAM case dem-onstrates, U.S. firms appeared to have little tofear from producers in Japancertainly notfrom those, in Western Europe. But from aminor position in 4K chips, Japanese firmswent on to cllim about 40 percent of the worldmarket for r.ollowing generation of 16KRAMs. By 1982, the perception was wide-spread that U.S. firms had "lost" the marketfor dynamic RAMs. Certainly this is an over-dramatization, and the RAM market can by nomeans stand for the industry in microcosm; butthe picture has changed. How did it change sofast?

During the 1970's, Japanese companies ex-ported considerable numbers of electroniccomponentsincluding transistorsto thiscountry, but the major growth segment, ICs,was dominated by 'American suppliers. Eventhough Japan's Government protected the localindustry, U.S. shipments took a substantial partof the expanding Japanese IC market. Custom-ers in Japan depended on American firms fordevices that domestic manufacturers could notprovidehigh performance or large-scalechips, custom-parts, even some types of linear

circuits needed for consumer products. As thetechnological level of Japan's semiconductorindustry caught up with that of the UnitedStates, many of these imports were replacedby indigenous production. The phenomenon,termed import displacement, has been charac-teristic of Japan's computer industry as well.Displaced items quickly become potential ex-ports for Japanese firms.

During the 1970's, awareness of the possibleconsequences of foreign competition grewwithin U.S. industry and Government, al-though the production and trade data showedlittle cause for concern. The. Federal TradeCommission, reporting on interviews con-ducted in 1976, stated: ". . . a number of com-pany executives expressed the opinion thatcompetition from foreign companies would bemuch tougher to handle than competition fromother U.S. companies in the next 5 or 10 years.In contrast, some other executives felt that U.S.companies would not have a difficult timemaintaining their technological lead overforeign companies."27 Hindsight shows thoseof the first persuasion closer to the mark.

One sign that patterns of international com-petition would change came from subsidies,and promotional efforts adopted by foreigngovernments with the aim of fostering in-digenous production. Japan, France, West Ger-many, the United Kingdomall in one way oranother marked the semiconductor industry ascritical to continued economic vitality, an in-

27Staff Report on the Semiconductor Industry: A Survey ofStructure. Conduct and Performance (Washington, D.C.: FederalTrade Commission, Bureau of Economics, January 1977), p. 130.

193


dustry not to be given over to foreign interests.Since the United States was far ahead in bothtechnological expertise and production, vol-ume. the implicit targets were American com-panies, not excluding those that had investedin local production facilities. These govern-ment-led attempts to build competitive semi-conductor industries have had Mixed result.s.Joint projects involving public and private sec-tors in Japan were quite successfulin semi-conductors as in earlier Japanese industrialpolicy initiatives. European attempts have beenfar less fruitful, for reasons that may have asmuch to do with the characteristics of the in-dustry and marketplace on the continent aswith the policies pursued.

United States

Applications of semiconductors reflect ongo-ing synergistic relationships among merchantsuppliers and their customers. Purchasers out-side electronics have lately presented growingmarket opportunitiese.g., in automobiles.NonethelesS, from a technological viewpoint,firms building computer- or microprocessor-based systems remain the most influential cus-tomers (fig. 34). Manufacturers of consumerelectronics, communications systems, instru-ments and controls, and office equipment haveconsiderable impact as well. While most of theattention below goes to merchant firms,/cap-live operations have played a vital role in thetechnological development of the U.S. ,indus-try. Furthermore, production decisions' by thelarger integrated manufacturers sometimeshave major consequences for the merchantmarket.

Figure 34 shows that the phenomenal expan-sion in semiconductor output during the 1970'swas accompanied by a major shift from de-fense purchases to consumer and industrial ap-plications; competitive success in the mostrapidly growing market segments depended onthe ability to make the transition from special-ized military requirements to the demands ofprivate sector customers. Some companies that

"fared quite well in the military market couldnot compete effectively for commercial sales,

Figure 34.Distribution of U.S. SemiconductorSales by End Market

M !nary Consumer . Computer andindustrial

1960 1968

Year

1980

SOURCES 1960. 1968: "innorafion. Corrpctitu;. !.16 t,,...,:entilental Policy inthe Semiconductor Industry," l..;,,ert Art Associates, inc.,final report for Experimental Tei.hr..,:l!'iT, eAl^oires Program,Department of Commerce. Mi,rch.1130,

1980: Status '80. A Report o f the Intck.vted :nOustryiScottsda:e, Ariz: Integrated Circuit En;.,ep. 34.

where the needs of customers are nr ire div.arse,and nontechnical dimensions like price .0 de-livery schedules more important.

Factori in Strategic DecisionsCompetitive strategies adopted by merchant

semiconductor firms revolve around factorssuch as size, market power and technologicalcapability, internal need for devices (if any),and stage of development relative to others inthe industry. A company's technical strengthsshape its product line. Process technologywhether a manufactu-re,f is strong in bipolar orMOS (metal-oxide semiconductor), which va-rieties of MOS a-firm knows bestis one as-pect, design capability another. Some com-panies are known for innovative circuit de-signs, others for prowess at mass produc-,tionsome for both. Smaller entrants tend tospecialize; only a few merchant suppliers havebroad product lines (the world's semiconduc-tor manufacturers supply perhaps 50 billiondevices a yearof 100,000 different typestoseveral hundred thousand customers).

Ch. 5Ccrruc,..1.tine!...:-..7 Pille.-iilational Electronics Industry 189

Pho,;. Electric Co

Technician loading r.i'i.taii:6,,i7:;::;,c)semiconduclf;:r

A number of U.S. mi:.1-..::,!:.ani..::..e.7.;ufacturershave integrated to SGic:l_:7 i'ato systems.A few, such as Texas IL-it.ri,,,7,.t,sAtfi. have alwaysbeen diversified. Others htM1 put chasedby larger enterprises but sZ;;Ii the great bulkof their production on the open ffiarhe't fch. 4,table 24). As merchant supplf...ers expand, sodoes the range of their product offerings.Smaller companies with limited resources aimat niche markets. Newer entrants set out to de-velop specialized or custom devices of less in-terest to larger corporations; tho 1981 startupLinear Technology - -a spinoff from NationalSemiconductorspecializes in linear. ICs, inwhich the founders have expertise.

Captive semiconductor producers have dif-ferent strategic aims. While most of the largercomputer firms make some of their own logicchips, IBM has traditionally produced most ofits memory circuits as well. The company con-sumes so many that, for products such asRAMs, on occasions when it chooses to pur-chase from outside vendors it can account fora sizable proportion of total demand (the com-pany is probably the largest single purchaserin the merchant market, as well as the world'slargest producer of semiconductors). This thenaffects the business decisions of merchant sup-pliers: II3M's external purchases were a power-ful and rather unpredictable force on the mar-

-%,67K at the close of the .1970's,c'or.:*,; -fuy's Q,::;-le.i.:pected entry contributing

.,..;AoriF.,gPs of ih,''se devices. Capacity short-,._by firmsprimarily stemming

firoe:n invest in additional productionis: ,.pities .tn the of the recession of 1974-75

an open c"...)or for Japanese IC suppliers_ , in. this c.:;:ultr,.

Captive maotUat- Curers contribute in a majorway to the Otrec-'1:, f-ti-ength of the U.S. position.n tnicroelectr.-i'cs through their R&D activ-ities;: particu'ici-..1.BM and AT&T's Bell Lab-orazolies have !.,7;en.respons:ble for much of the

research ,..-ncieriying_tne semiconductorik-,.clustry in thi-7, country, indeed around the

M'4?rcharit firmsbecause of the pace.f.nteristy.or.:pr '1:tuct. development, the con-

cycles tr, improvement in design andI characterize succeeding'.Cs---must set different

-t .,,YIT also hava more limited resources

Ir.aes.rt(A 11:-.1.! rsa new device must

UF44/.;111 their design is at thethr. of the art it will be

supercedt.:1 sooner. Timing is crit-Tt.i: . sometimes open,

voviding opportilnities for leapfrogging thecGaipetition. Companies .that quickly masteredproduction of dynamic RAMs, or concentratedon microprocessor design architectures in at-tempts to tie up large portions of that market,were aiming at such advantages. Needless tosay, some firms have better records at ex-ploiting these opportunities than others. Anumber of companies that had been strong inbipolar technologyincluding Fairchild andTexas Instrumentsdid not move as rapidlyinto MOS as the competition; Texas Instru-ments staged a quick recovery,, while Fairchild'has continued to lag. Mostek, as its name con-notes, was founded with the intention of spe-cializing in MOS; the company has emphasizedmemories, designing their own RAMsthe defacto industry standard 4K RAM was a Mostekdesign (see app. C)while serving as an alter-nate source for microprocessors. Electronic

19J

190 internacional Competitiveness in Electronics

Arrays, now ol.,:med by Nippon Electric, hadspecialized in read-only memories (ROMs).Other firms seeking to exploit particular tech-nological paths have had less success: Amer-ican Micipsystems' work on v-MOS ICs is oneexample, RCA's pursuit of silicon-on-sapphirec -MOS another. Internationally, Japanese firmsmoved into MOS ICs much more rapidly thanthe European most of whom are still wellbehind their competitors in the United Statesor Japan and relying on technology imports totry to catch up; this has been one of the objec-tives of Le Plan des Composantsa major in-dustrial policy effort by the French Govern-ment (ch. 10). In the United States, some firmsSignetics, Monolithic Memoriescontinue tospecialize in bipolar devices. IBM likewise re-mains relatively stronger in bipolar than MOS;the speed advantages of bipolar circuits have`!d many computer manufacturers to continueemphasizing the older technology.

Quality and reliability comprise another com-petitive realm where strategies depend both oncircuit design and manufacturing practices (ch.6). While Japanese firms have zealously pub-licized the quality and reliability of their ICsin much the same way that Japanese con-sumer electronics firms used reliability as awedge into the American TV marketdomes-tic producers like. Adva need Micro Degiceshave also pursued an 'age of quality andreliability as a marketing tool.

Products and Prices

One of the attractions of memory circuitsin addition to the vast marketis the relative-ly orderly aid predictable progress of the tech-nology itself; circuit design is vitalalong withexcellent process capabilitybut more straight-forward than for logic or microprocessors.Everyone in the industry knows that the nextgenerations of dynamic RAMs will be 256Kchips, followed by 1 megabit; circuits offeredby various firms are much more similar thanthe designs of competing 16-bit microproc-essors. One result is the fierce price competi-tion that has often seemed the dominant char-acteristic of the memory market. Progress instatic RAMs, and in the various types of ROMs,

is likewise rather easy to predict. Under suchcircumstances, Japanese suppliers quite nat-urally emphasize memory products.

In contrast, market acceptance of logic cir-cuits or microprocessors is less predictable. In-vesting in a new microprocessor design=the32-bit Intel iAPX 432 cost more than 5100 mil-lion to developis risky, but the potential re-wards are great; designs with an edge over thecompetitionin performance, ease of program-ing, adaptability to a wide range of applica-tionssell for premium prices. -'3 Furthermore,microprocessorsbest thought of as familiesof related ICs rather than unique deviceshavelonger product cycles, extending the periodover which investments can be recouped.Memory circuits are manufactured as long asdemand holds up, but sales tend to peak anddecline more rapidly than for other devicetypes. Five or six years elapsed between theonset of high-volume production for 8-bit mi-croprocessors and mass production of thesuc-ceeding generation of 16-bit parts, while life-cycles for RAM chipsthough slowly length-eninghave been perhaps 3 years, onmetimesless.29

Abbreviated product cycles dictate strategiesaimed at profitability within a narrow timewindow, along with continuous efforts to de-velop new or differentiated offerings. The lat-ter can be original designs but need not; sec-ond-sourcing has been widespread for manyyears, in part because customers often insiston more than one supplier before they willdesign an IC into their end products. Thus, sec-ond-sourcing can accelerate market expansionfor everyone. Semiconductor firms choose tobecome alternate sources for chips developed

za0n the costs of microprocessor design, see R. N. Noyce andM. E. Ho:f, Jr., "A History of Microprocessor Development atIntel," IEEE MICRO. February 1981. p. 8. Intel's first micro-processor, a 4-bit device, was designed in 9 months by a singleengineer; 100 man-years went into the iAPX 432.

"Inters 8080 familyintroduced in 1974 and the largest sell-ing 8-bit microprocessorwill no doubt remain in productionfor many more years. Worldwide, more than 10 companies stillproduce 8080 chips. MOStekan alternate source for anotherpopular 8-bit processor, the Z-80for a number of years pro-duced more of the devices than Zilog, the originator. See TheAntenna." Electronic News, Mar. 12, 1979, p. 8; "Eight-BitLevel." Electronic News, July 5, 1982, p. 12.

19u

Ch. 5Cornpetitiverriss in the International Electronics Industry 191

by other companies to beef up their own prod-uct lines, perhaps by complementing circuitsthey already build, as well as to reduce marketrisks and save on R&D expenses. From theviewpoint oi the originator, it may be more sen-sible to settle for a smaller piece of a rapidlyexpanding market than to try to keep othersfrom duplicating an IC design. Attempts to pre-vent duplication are virtually impossible if acircuit finds an enthusiastic reception in themarketplace. As one consequenc-e, formalizedalternate sourcing agreements have largely re-placed the copying that was once common-place. Sometimes alternate source manufac-lurers acquire the originator's technologye.g.,mask sets for lithography. Other times onlydrawings or specifications are provided. Therecent agreement between National Semicon-ductor and Fairchild, the latter acquiring theright to build National's model 16000 micro-processor in exchange for developing a com-plementary line of peripheral chips, is an in-creasingly popular route.

Cost reductions via the learning curve (ch.3; fig. 5) help shape competitive strategies. Asproduction volumes increase, yields rise andmanufacturing costs drop. Pricing decisionshave often been based on projections of ex-pected cost reductions into the future. For afirm early to market with a new design, costadvantages over potential rivals can build rap-idly, increasing with leadtime. Firms that arelate to market face a dilemma; they may haveto choose between foregoing participation orpressing on with their own design in the hopethat it too will win acceptance. In early 1982,with six Japanese entrants mass-producing 64KRAMs, versus only two American manufactur-ers; a number of U.S. firms were confrontedwith such decisions; Advanced Micro Devices,for one, decided not to build a 64K chip.

In different circumstances, then, firmsassume different strategic postures. Companiesentering the market with a new device, par-ticularly one incorporating proprietary technol-ogyproduct or procesSmay have several ad-vantages over competitors, that follow. An earlyentrant will:normally try to remain ahead onthe learning curve, keeping production costs

below those of its rivals, margins above. Com-panies with proprietary designs often licensealternate sources, but at least at the outsetsecond-source suppliers will be at a cost disad-vantage. If the initiator decides to follow a pric-ing strategy keyed to anticipated cost improve-ments, follo-vv-on firms may find it difficult tomake a prosily. Texas Instruments, for instance,has had-the reputation of practicing advancepricing whenever possible. In a very real sense,then, later entrants can be at the mercy of in-novatori thc latter choose to cut pricesand exercise the cost advantages of being far-ther down th.f: learning curve. On the otherhand, an innovating firm might choose to in-crease margins by holding price levels high.Under such circumstances, an alternate sourcemay itself be able to carve out a place throughprice. One facet of Intel's corporate strategyhas been to choose products where it couldenter the market first., reap high profits, thenmove onleaving later sales, at lower margins,to others. Nonetheless, in many cases, especial-ly where the innovating company is small, lin-ing up an established supplier as a secondsource may be a preret,:lisite to sales in any vol-ume. .

A further strategic choice, increasingly crit-ical for American firms, is whether to designand produce commoe!ty-like chips or to con-centrate on custom or semicustom devices. Thefirst option entails high-volume production of.ICs, that are, or may become, shelf itemsstandard circuits serving the needs of cus-tomers who design them into end products.The alternative, customizing, can be ac-complished in a variety of ways; semicustomchips such as gate arrays or programmablelogic arrays are specialized' only at the laststage of processing. Regardless of the techno-logical approach, firms in the custom or semi-custom business create specialized ICs meetingthe needs of one or a few, rather than many,purchasers. Because circuit design is expen-sive, prospective order quantities must be largeenough to cover engineering costs; alternative-ly, the buyer must be willing to pay a higherprice. Custom chips for automotive applica-tions are an example of the high-volt, me case,

197

792 fntemationaY Competitiveness in Electronics

military systems of custom chip markets whereproduction volumes tend to be small. Occa-sionally. end users design their own ICs andcontract out production.

At the center of competitive strategies insemiconductorsas for many industriesthenlies the choice of products. Firms with broadproduct lines may offer devices based on avariety of technologies while attempting to stayat the technical frontier with only some ofthese. Otherssuch as Mostek in the mid-1970 'soperate within narrower boundarieswhere they attempt to be leaders. Some en-trants are content to-follow the obvious trends,offering unique designs infrequerely whilerelying on other strengthsperhaps low pricesor a reputation for qualityto attract custom-ers. In its early years, Advanced Micro Devicestook such an approach (see app. C).

Beyond 'these common themes, mostly hing-ing on aspects of the technology, companiesplan their strategies according to the strengthsand weaknesses the.; perceive in their own po-sitions compared to those of their rivals. Nosingle company has the resources to manufac-ture and sell all the tens of. thousands ofsemiconductor products now marketed in theUnited Statesone of the reasons for theperiodic emergence of startups. Extensiveproduct lines can confer advantages wherecustomers prefer to deal with only a few ven-dors; broad-line manufacturers may also beable to achieve economies by spreading mar-keting costs over many items. Nonetheless,such factors are sev.ondaly compared to choiceof product and process technology, along witha variety of ingredients that could be labeled"entrepreneurship." Successful new compa-nies hale frequently been esta&ished to exploita particular product, often one that larger com-panies liave failed to pursueperhaps becauseof limited resources, or simply a judgment thatthe pi.kaatial market was too small.

No rnA,..,tter the decisions they themselves take,of semiconductor firins can be cer-

4.11 of Jr,,L H-,ilkires of their market: 1) competi-:'on .P.,511,;ventually drive prices downward (it

bopn extremely difficult to capture signift-

cant monopolistic profits Yri in new technolo-gies), and 2) just as inevitably the pace of tech-nical advance will render new product offer-ings obsolete within a few years at most. Tht-,price history of the 64K RAM illustrates thefirst point: offered in sample quantities at 5100each in 1-59, and $2n tr, during 1980,prices dropped to the :5 range in early1981 and $5 to $7 a year later; during 1981, 16KRAM prices, were driven down from S4 toabout Sl, largely as a result of price declinesfor 64K parts.3° Such pricing trends have meantthat all firms, U.S. and foreign, have had towork continuously at cost reduction. In (-en-trast to numerous other industries, passive: orreactive pricing policies are hardly possible;although product differentiation is a viablealternative under some circumstances, pricecompetition in semiconductors is a constantforceenough by itself to set this industry offfrom many others. The second market charac-teristiE, rapid technological change, has forcedmanagers and technical personnel alike toadapt constantly; firms that have rwedded to older technologies have faltered ordisappeared from the marketplace.

International Dimensions

From an international perspective, the largerU.S. merchant firms have shared three majorstrategic thrusts: 1) offshore manufacturing toreduce labor costs, 2) foreign investment toserve overseas markets, and 3) attempts to dobusiness in Japan. This last efforton whicha number of companies are just embarking. orreembarking after past rebii fr -it-ical to.the continuing abilifirms to compete with lai , Akir ,Ikt:U Japa-nese manufacturers, particulariy in commodityproducts like memory.

Offshore manufacturing investments havebeen concentrated in developing Asian na-tions. Generally, the more labor-intenSiveassembly operationse.g., wire bonding andencapsulationh il.,lbeen moved. In the semi-

,,A dr '..1; i'erm 64K Pacts UntilU.S Fire; ':. I :; FterfroaP: P.Ti.u.s. Feb. 8, 1982. m. p. C.f I 7 ;II ape (114) Studied." New Yu::: limes,Mar. p ; for 64F. chips eventually fell ;;r, lowsof about S3 duriag the 1982 slump.

1 96

Ch. 5--;a2mpe.titeness rn the interrtalionat Electronics industry T93

conductor industry, the :stimulus for thesetransfers has no bee-n. import competition, asin consumer electronics, but domestic rivalry.Transfers offshore began in the early 1960's,long before Japan appeared a significant threatin semiconductor production. By 1970, Ameri-can companies operated more than 39 subsidi-aries in such locales as Hong Kong.. Singapore,Malaysia, the Philippines, and Mexico." Re-locating labor-intensive production operationshas been especially attractive because transpor-tation costs are low; chips are often .,hippedby air. A cost comparison illustrating the ad-vantages of offshore assembly is included inappendix B (table B-2).

A second international involvement of U.S.semiconductor companies has been point-of-sale production to serve developed countrymarkets. Investments in point -of -salt plantsbegan about the same time as offshore

but from the standpoint of industry strat-egy the motives were quite different. Thesehave- been twofold. First, foreign governmentshave ofte:, taken steps to make local produc-tion attractive, or, conversely, to make export-ing from the United Slates onerous. Europeancountries, in particular, have relied on incen-tives combined with tariff and nontariff bar-riers to attract U.S. high-technology in-vestments. Second, point-of-sale productioncan become a competitive necessity to the ex-tent that other firms have already made suchmoves.

Efforts to establish sales, production, and/orR&D 'at:jiffies in Japannow a bigger marketthe, ,ill of Europecomprise the most recentoverseas thrust by American firms. WhileTexas Instruments had been able to establisha wholly owned subsidiary in Japan, otherAmerican firms were kept out until recently.Semiconductor manufacturers attempting toinvest in Japan suffered much the same fateother U.S. companies; the Japanese Govern-ment, through the Foreign Investment Coun-

31W. F Finan, The !mei-national Transfer of SemiconductorTechnology Through U.S.-Based Firms, National Bureau ofEconomic Research Working Paper No. 118. December 1975,p..57. This figure excludes point-of-sale facilities in industrialnations.

cii, MITI, and other agencies, controlled in-ward investment flows and for the most partprevented the establishment cf manufacturingfacilities under foreign ownership:32 Joint -aen-tures in which a Japanese. company held themaiority interest met a more favorable re-sponse. The purpose was obvious: to pro-aide:shelter for 'Japanese companies which a;. thetime were well behind in semiconductor tech-nology. MITI believedwith good reason, ifthe European case is indicativethat allowingAmerican firms to r.roduce in Japan would sti-fle the domestic industry, particularly when it.carne to more advanced device types. In actingthis way, the Japanese Government was behav-ing much like others that haVe sought to pro-tect infant industf.es, but ,Japan has often beenaccused of maintaining rizotectionist mea;ureslong past the point at,Which her industries havebeen able to fend .fir themselves.

In any event, as a consequence of protec-tionism in Japan, American suppliers wereforced to adopt business tactics differ ii. fromthose Pursued on the continent. Most re-sponded to MITI's entreaties and entered intolicensing agreements with Japanese produc-ers.33 Such steps were entirely rational, pro-vided the U.S. firm could be reasonably cer-tain the technologies transferred would notfind their way into products they would faceat home or in third-country markets. With thisproviso, it would pay to sell technical knowl-edge, the proceeds from which could then atleast partially offset the costs of generatine :hatknowledge. The outton. of licensingmelds with Japanese firms have led to manysecond thoughts within the American industry.Nonetheless, clear-cut, cases in which U.S.technology was an irreplaceable ingredient inthe growing ,capability of Japanese semicon-ductor manufacturers are ripe, particularly illlater year:, ble etxr,'optioL being perhaps de-velopments ilowinp: from Bell Laboratories.

"M. Y. Yeshino, "Japanese Foreign .'Direc. Investment," TheJapanese Economy in lnternationall'etrspective, I. Frank (ed.)(Baltimore; Johns Hopkins University Press, 1975), p. 248.

"See, in particular, W. F. Finan, "The Exchange of Semicon-ductor Technology Between Japan and the United States," FirstU.S.-Japan Technological Exchange Symposium, Washington.D.C.. Oct. 21, 1981. Finan points out that American firm); general-ly did not transfer proprietary information to licensees (p. 9).

199

194 b-i!ernational Competitiveness i,n il-i'ectranics

Texas Instruments became the single excap-tion to MITI's licensing ruleri reality onlya partial exception. Because it held a series offundamental patents covering ICs, Texas in-struments.was in stronger position than otherAmerican companies. As a condition for li-censing its patents to Japanese firms, Texas In-struments demanded that it be allowed to estab-lish manufacturing operations there. WhenMITI refusedconsistent with its decisions re-gardinggarding other electronics firmsthe stage was

-et for prolonged negotiations." Texas In-struments and .MITI eventually compromisedin a 1968 agreement permitting a joint semi-conductor manufacturing venture with Sony.Four years later, Sony sold its share to theAmerican company.

Thus, Texas Instruments, although reported-ly subject to a production ceiling, became theonly U.S. semiconductor firm to both manufac-ture and sell its devices in Japan, just 6s IBMhada few years earlierbecome the onlyAmerican company to build computers there.(IBM was also able to gain entry by taking ad-vantage of its patent position.) Only recently,as the Japanese have gained confidence in theirown technical abilities, has MITI softened itsattitude toward foreign investment;; a growingnumber of U.S. electrcnics manufacturers arenow contemplating wholly owned subsidiariesin Japan. While the longer term consequencesof these decisions are not yet clear, investmentwithin Japan couldgiven the examples ofothci industriesprove a vital support forAmerican firms seeking to compete'With Japa-nese rivals in third-country markets as well asin japan.

Current TrendsThe competitive strategies of American semi-

conductor firms have been aimed first andforemost at survival in a highly competitive,rapidly changing market. Companies big andsmall have had to stay abreast of and adapt totechnological change. Flexible managementand organizational structures have been a

341. E. 'Mon, In!ornational Diffusion of Ton.iinology: Thu Cast,of Semiconductors (Washington, P.(;., Brookings Institu-tion, 1971), pp. 141147.

necessity. The usual explanations for the exitsof a number of large corporations during theearlier years of the industry revolve aroundrigid decisionmaking styles.

More recently, the character of the markethas been shifting; American companies havebeen forced to alter their thinking. In somerespects the changes are- a continuation of-familiar patterns: more complex ICslarge andvery large-scale integration (VLSI)make stillmore applications co ;t- effective, creating newand different puzzles for chip-makers. Morefundamentally, VLSI has altered the cost struc-ture of the industry in at least two ways. First,production is growing more capital-intensive;new sources of financing are needed to pur-chase more expensive manufacturing equip-ment (ch. 7). Some of the capital has come viamergers, which have changed the industry'sstructure. The second way in which VLSI isaffecting the structure of the industry stemsfrom shifts in product design. What had beena hardware-oriented business is now systems-and software-based as well. ICs are becomingmore than components. To tap the vast poten-tial markets made possible by micr.i.processorscoupled with cheap memory, semiconductormanufacturers must commit substantial re-sources to computer-aided design and softwaredevelopment. This comes at a time or in' !n

Eying internatio 011--) . orv. 'I ..out the Japanese in the picture, the prob-lem facing U.S. merchant firms is one of lo-cating sources of new capital in substantialamounts while battling to ---'serve even theirc :isting pr-Ait mar 1:s, A companies deviset eir respt,..ises, se, ,ids are emerging.

Greater vertical integration will probablyhave the farthest reaching consequences. Larg-er merchant companiese.g., Texas Instru-ments, which has entered a variety of con-sumer markets, including that for personalcomputers- -are taking advantage of broad-based positions in microelectronics to integrate;lownstream into the manufacture and sale offinal products. The reasoning behind such de-cisions is straightfo ward. If much of the tech-nology of data procw..,',og and other electronicsystems is incorporated in ICs, why not make

20(1

Ch. 5Competitivene;s in the International Electronics Ind-..i.stry 195

end products too, increasing value-added andprofitability? To this strategyessentially an of-fensive onecould be added a defensive ele-ment. For semiconductor manufacturers withthe resources to contemplate entry into systemsmarkets, greater vertical integration reducesvulnerability in the event that customers beginto inte.z.-4-rate backward-into lee pi iuilLiiT::.

where it is becoming common for such agree-ments to spell out in considerable detail theR&D and/or circuit design obligations of eachpartner.35

Arrangements in which two or more com-panies independently develop different mem-bers of a family of chips fall at one end of theR&D spectrumcomplementary product devel-opment. At the other end, closer to basic re-search, industry groups are moving towardcooperative rather than independent but com-plementary projects. The Semiconductor In-dustry Association and the American Elec-tronics Association have each established pro-Grams that channel contributions frommember firms to university projects. The Semi-conductor Research Cooperative is funding re-search directly, while the Electronics Educa-tion Foundation aims to improve training inelectrical and computer engineering, primari-ly through fellowships and faculty support."Another effort, Microelectronics & ComputerTechnolr Corp., will be an independent prof-it-seeking iacEi organization capitalized by theparticipating firms.37 At least six universitiesare also establishing centers for R&D in semi-conductor technology and/or systems applica-tions of inicroelectronics.38 Whether all these,efforts will survive and flourish remains to beseen.

The emerging strategic picture in the UnitedStates, therefore, is fluid and uncertain. Semi-conductor manufacture; along with other por-tions of electronics, is undergoing far-reachingrestructuring, with outcomes that are hardlyobvious. Given settlements in the IBM andAT&T antitrust cases, the way also seems clear

The fact is that backward integration is on theupswing. as manufacturers of computers, of-fice equipment, consumer durables, and a hostof.other products sense the need to develop in-house capability in microelectronics. One pathis purchase or merger with a semiconductorcompany. Merger activity in the industry hasbeen high since the latter part of the 1970's; by1983 only a few of the larger, broad-line mer-chant suppliers remained independent.

Mergers have been of several types: somesemiconductor firms have been absorbed intoconglomeratesone example is United Tech-nologies' purchase of Mostek. Other acquisi-tions have been more directly motivated by in-ternal needs, as in General Electric's acquisi-tion of Intersil. Foreign t L .13 beenprominentSchlumberger cchase of Fair-child. Sometimes the appare. motive on thepart of the semiconductor company is the needfor new financinc.; this no doubt a factorwith Mostek, explicitly so with IBM''of a substir..!tial interest in Intel. The motivesof foreign i-vestors have varied: buying an

-nerican firm can be a quick way to get tech-nology as well as a convenient entrance intothe U.S. market.

In a related development, many Americansemiconductor companies are seeking alter-natives to "going it alone" in the developmentof new technology largely because of risingcosts. New variations on accepted technotfogyshanig arrangements have been devised. Somesemi :inductor manufacturers have prevailedon customers for assistance in developing spe-cialized chips and software. Both GeneralMotors and Ford have supported such efforts.Semiconductor firms have also .sought newways to share product development costs,among themselves, sometimes through exten-sions of past practices in second-sourcing;

"For examples, see S. RuLisell and S. Zipper. "Intel. MotorolaTighten Fold on General-Purpose MPUs: See Peripherals KeyMarket for Niche Suppliers," Electronic News. Mar. 8, 1982.p. 1. U.S. merchant firms are also negotiating such agreementswith Japanese menufacturers-5'. Russell. "Zilog. Toshiba toSwap MPU, CMOS Technology." Electronic News. Apr. 19,1982, p. 53; "National Semiconductor Sets Venture With lap.nose Firm," Wall Street loth-nal Jan. 23. 1983. p. 22 (the Japaneseparticipant is Oki).

"S. Russell, "SIA Eyes $5M Funding for Research Coopera-tive." Electronic News, Dec. 21. 1981. p. 6.

"C. Barney, "R&D Co-op Gets S6t To Open Up Shop." Elec-tronics, Mar. 24. 1ry83. p. 89.

'8C. Norman. "Electronics Firms Plug Into the Universities,"Science, Aug. 6; 1982. p. 511.

2 01

1.96 international Competitiveness in Electronics

for continued expansion by these two giantsemiconductoricommunicationicomputer com-panies. AT&T's manufacturing arm, WesternElectric. plans to be the first American firm todeliver 256K RAMsa rather spectacular en-trance into the merchant market. Other firmsXerox is oneare also contemplating broad.systems-pi ienteu strategies. Nicctim n e, Sinder companies continue to seek specializedproduct niches that will prove lucrative whilenot attracting large and powerful competitors.And in the background are the Japanese, add-ing another dimension that will continue to in-fluence the strategies of American firms bothdomestically and internationally.

Japan

Until half-a-dozen years ago, few in the U.S.semiconductor industry gave much thought toJapan as a serious competitive threat. Japanesemanufacturersalmost exclusively divisions oflarge corporationsmostly produced devicesfor consumer products; even today, nearly halfof Japan's semiconductor output goes ,:o con-sumer appiications.39 During the 1970's,Japan's budding computer manufacturers de-pended on American suppliers for advancedICs. While Japanese companies were clearly onthe way to the skill levels needed for mace ad-vanced devices, the prevailing belief in theUnited States was that they could not reallyhope to catch up. The primary concern was theclosed Japanese market. American companieshad been prevented from establishing a pres-enCe remotely comparable to that which theyhad achieved in Europe; customers in Japanbought only those devices that were not pro-duced locally.

Today the situation seems quite different.Japanese firms have emerged as viable globalcompetitors in VLSI devices. Although theirprowess has centered on memory chips, theyhave made up a great deal of ground in logiccircuits and other device types as well. By 1980,the gravity of the threat had become obvious;quite suddenly, Japanese firms captured near-

"Japan Electronics Almanac 1982 (Tokyo: Dempa Publications,Inc.. 1982). pp. 142. 143.

ly half the American market for 16K RANIs.Two years later, Japan's manufacturers seemedwell on their way to comparable levels ofpenetration in the next-generation 64K RAMs;indeed, as sales began to build, the Japineseshare soared toward 70 percent. While thereis considerable feeling that ultimately they willltui be able to 11 1

01 11,311 dbuut 11(111 tilt,U.S. market for 64K chips, any temptation tounderestimate the capabilities of Japan'ssemiconductor manufacturers has long sincepassed. Seemingly countless studies recountthe strategic attack, tracing the targeting prac-tices of government and industry.

From Linear to Digital CircuitsFirms in Japan had long since become major

producers of linear semiconductors, a main-stay in consumer electronics. By the early1970's, some American companies began toabandon this part of the market, especially asdomestic sales seemed to be drying up. LeadingU.S. producers put their resources into rapid-ly expanding digital IC technologies. Mean-while, for tine Japanese, strength in consumerdevices was both a blessing and a curse. Whilegiving their engineering staffs experience incircuit design andmore important in high-volume production, the concentration on linearcircuits did little to raise overall levels-of competence. At the time, the primary customers fordigital ICscomputer manufacturerswere arelatively minor component of Japanese de-mand.

This was the situation when, in line with itslongstanding policy of fostering internationallycompetitive industries, MITI acted to break theimpasse created by the focus on consumerproducts. The agency helped fashion an R&Dprogram intended to increase Japan's capabili-ties in large-scale digital ICs, particularly MOSdevices, and accelerate movement towardVLSI. The organization of the program, whichbegan in iate 1975, is described in chapter 10;five companies and three separate laboratorieswere involved in the 4-year effort. Fundingtotaling about $300 millionwas provided part-ly by the participants and partly by govern-ment. The program has had far-reaching im-

Ch. 5Competitiveness in the International Electronics l:Idustry 197

pactsas much through diffusing technologyand training people as through the technologydeveloped. A parallel government-sponsoredV LSI program---Cc.is one focused on telecom-municationswas carried out in the labora-tories of Nippon Telegraph and Telephone(NTT). which had the most capable microelec-fz:rtpi,-,-; Rk-n !Ppan

MITI's objective was not only to aid Japan'ssemiconductor manufacturers: the VLSI pro-gram was part of a much more extensive ef-fort to move the nation toward knowledge-intensive products with high export potential.Like its counterparts within the governmentsof carer industrialized countries, MITI recog-nized that semiconductors would be funda-mental building blocks for many sectors of theJapanese economy. Supporting the computerand information industries was the first step.MITI was fully aware that technical compe-tence in digital ICs would be essential, and thatwithout some form of stimulus private compa-nies would find it difficult to shift rapidly fromlinear to digital devices. From MITI's perspec-tive, support for "cooperative" R&D was a nat-ural extension of past efforts in other indus-tries; the VLSI program itself has been followedby related work in computers and robotics, aswell as further microelectronics projects.

Still, by American standards, MITI's sub-sidies were not large. Individual U.S. firms likeTexas instruments had R&D budgets that camecl,-Ise to matching tho yearly outlays of the VLSIproject; IBM's corporate R&D spending was anorder of magnitude larger. Of course, partici-pating Japanese companies continued their in-ternally funded R&D programs; MITI spendingthus gave a substantial incremental boost toJapanese semiconductor research, reducingrisks and supporting longer term work. Evenso, total expenditures in Japan remained wellbelow those here. Nor did the VLSI project re-sult in large and direct benefits to _Japanesefirms, at least in terms of product offerings. Agreat deal of attention in the United States hascentered on the thousand or SG patents asso-ciated with the program, but it is not clear whatvalue these have. There are no signs of majorinnovations. Primary attention went to process

rather than product technologies; one-third ofthe funds were spent in the United States sim-ply on purchases of state-of-the-art manufac-turing equipment. This suggests that the majorthrust was to develop skills in low-cost produc-tion of commodity-like devices such as RAMchips.

Two aspects of MITI's approach deserve par-ticular emphasis. First, subsidization of micro-electronics-R&D was only the opening movein a broader strategy for building a competitivecomputer and telecommunications sector.Hindsight provides ample corroboration ofwhat was in fact an explicit goal: MITI's subse-quent support of computer and software devel-opment, as well as the Japanese Government'sreluctance to allow open competition for NTTprocurements. NTT was a principalthoughindependentactor in the VLSI program; thegovernment evidently hoped to restrict its high-volume purchases to domestic manufacturers(the company does not produce its own semi-conductors), helping generate the economiesof scale so necessary for international correti-tiveness.

The second pointsuggested by MITI's levelof support, generous for a government R&Dprogram but certainly not enough by itself to.boost the Japanese industry past Americanfirmsis that the VLSI project was never con-ceived purely as an exercise in technology de-velopment. Consistent the usual Japaneseapproach to govern:: '-ant- supported R&D, itwas intended to focus industry efforts, helptrain engineers from private firms in state-of-the-art technologies, diffuse these technologieswithin the Japanese industryin other words,to overcome weaknesses in Japan's technologi-.cal infrastructure created in part by the lackof personnel mobility (ch. 8).

This makes it doubly difficult to assess thecontribution of the VLSI project. While sep-arating what might have happened from whatdid occur is impossible, pieces of evidence doexist: for instance, MITI excluded Oki Electricfrom participation and subsidies, yet Oki man-aged with NTT's help to develop a 64K RAMthat the company now exports in considerable

193 International Corn;;etitiveness in Electronics

volume. Many of the events of the past fewyearsthe ups-urge in Japanese production andexports of RAMswould probably have oc-curred in any event. albeit at a slower pace;memory chips were obvious targets for Jap-anese firms confident of their :abilities tomass-produce relatively straightforward de-

:, sl:Indards.Thcy wan; alsoneeded for the computers that these same firmswere determined to make in greater numbers.

In the United States, the impact of the VLSIproject has been exaggerated. It has come tosymbolize not only direct subsidization of com-mercially oriented R&D but also interfirm co-operation that might he illegal under Americanantitrust. law. In fact, as pointed out in chapter10, cooperation diming Japanese companieshas been rather limited evidence of thestrength of the harriers within the Japanese in-dustry that MITI was trying to overcome; thisaspect has been overplayed by American man-ufacturers understandably distressed at in-roads by iverseas competitors. While the VLSIproject makes a convenient target, by itself itis a far-from-adequate explanation for penetra-tions of what had been traditional Americanmarkets. Indeed, government policies in sup-port of Japan's information industries haveranged far beyond R&D subsidies. Among theother poliCy tools have been:

preferential -governineA procurement;favorable credit allocations, especially dur-ing the formative years:special depreciation and other tax meas-ures; andgrants for training and education.

Domestic firms have been effectively protectedfrom import competition as well as from pro-duction within japan by foreign-owned con-cerns. Protection of growing industries throughgovernment action is hardly unique, but canonly be judged to have succeeded if the pro-tected companies eventually emerge as viablecorn-petitors. In mi,-r-caectronics, the- "infantindustry" approach has been attempted else-where, most notably in several European na-tions, but only Japan has achieved success. Jap-anese industrial policy is discussed in detail inchapter 10; here the point is that none of thepolicy measures adopted by Japan's Govern-ment, taken separately, appear to have beenmajor forces in the ultimate growth and ma-turation of the semiconductor industry. Takentogether, they paint a different picture--one inwhich industrial policy provided vital guidanceand support for the development of an inde-pendent capability in semiconductor designand manufacture. Cumulatively, the policies ofJapan's Government have had a major impact.

Strategy and Structure

Despite MITI's pervasiva influence, the com-petitive strategies of individual Japanese semi-conductor manufacturers are governed first bythe basic structure of the industry, which ispopulated by companies for whom microelec-tronics comprises a relatively small part oftheir business. Most of these companiesOkiand Nippon Electric being partial exceptions'are large, integrated firms whose sales consistpredominately of final products such as com-puters, consumer electronics, and telecommu-nications systems. Table 45 shows that only for

Table 45.Proportion of Sales Accounted for by Semiconductor Products°

Japi.nese firms (1981) U.S. firms (1979)Nippon Electric Co. (NEC) 19.8°4 Mostek 93%Fujitsu 12.9 Advanced Micro Devices 89Oki Electric 9.8 Intel 75Toshiba 8.1 Fairchild 69Hitachi 5.3 Texas Instruments 36Mitsubishi 5.3 Motorola 31

5Including internal consumption.SOURCES: Japanese FirmsTable 29 (ch. 4).

U.S. Firm "U.S. and Japanese Semiconductor IniluLt,les: A Financial Comparison," Chase Financial Policy to,the SernicOnductor Industry Association, Juno 9, 1980, p. 1.5.


NEG and Fujitsuthe latter Japan's largest'computer manufacturer, with heavy internalconsumptionhave semiconductors contrib-ye(' a proportion of total sales even half asgreat as for those U.S. merchant firms that areleast dependent on their semiconductor divi-sionsMotorola and Texas Instruments. Semi-conductors account for less than one-tenth thesales of the other Japanese manufacturers. Inth.s,, they are closer to American companieslike Rockwell or RCA, which nonetheless dif-fer in being primarily suppliers of specialtyrather than mass:.market circuits.

The fact that the major semiconductor sup-plitrs in Japan build end products createspotential intracorporate synergisms absent incompanies that are primarily chipmakers.While some U.S. managers view integration asdyrtfuncticinal, likely to sap entrepreneurialdrive and retard innovation, it has been an ad-vantage for Japanese companieswhich havedi rerent sets of strengths and weaknesses thanAmerican firms. To begin with, the same half-do: ,,en corporations ,that produce most of Ja-.pair's ICs account for perhaps tvvo-thirds of de-inE;nd; given the focus on vertical integrationand internal production, it should be no sur-prise that U.S. suppliers have had difficultyselling in Japan. Second, as the next chapterpoints out, the quality of Japanese ICs has beenhigh again, thi might be foreseen. given thatfit-ins producing for internal consumption willfind themselves bearing high downstream costsvt:'hen quality lags. A further synergism asso-oCrated ith size and diversification stems from(le ability to tap cash flows generated in otherlines of business; these funds can be channeledto R&D or added production capacity, matterssimplified on in chapter 7. Further, diversifiedt."::ompanies can more easily tolerate short-termlosses resulting from price-led penetration ofhew markets. Diversified Japanese companieshave combined such tactics with an emphasison qualityboth image' and realityto driveboldly into markets once the province ofAmerican firms. Indeed, few other strategiescould have worked. Unfortunately, from thestandpoint Of smaller and less diversified U.S.merchant manufacturers, unrelenting price

competition in products representing a sub-stantial part of their total business leaves fewoptions for counterattacks.

The Japanese strategyprotecting domesticsemiconductor manufacturers from overseasrivals while providing modest R&D subsidiesacid at the same time fostering domestic com-petitionparallels that in television. It hasyielded 'equally impressive results: deep pen-etration in targeted markets based on lowprices and quality levels above previous norms.There has been a fortuitous element as well;unexpected demand swamped U.S. suppliersduring 1979 and 1980. As a result of continuedcapacity expansions during the preceding mar-ket slump, Japan's producers were ready to fillthe void.

Sbme spokesmen for the American industryfind other familiar features: claims have beenrepeatedly voiced that the Japanese practiceprice discrimination, maintaining high mar-gins in protected home markets while slashingprices in the United States and Europe. Suchtactics would imply either explicit or implicitmonopolistic agreement among Japanese man-ufacturerse.g., tacit acceptance of existingmarket positions at home, with price cuttingconfined to foreign markets. Even so, questionsof dumping are problematic for integratedfirms; companies making ICs for both internalconsumption and open-market sales have agood deal of latitude in allocating costs and set-ting prices. Dumping, as defined under GATTrules and the laws of most countries, would bedifficult to prove; nor would the usual,4.ation.ales for prohibiting dumping necessarily bevery relevant.

MITI's push toward ICs for computers andcommunications has contributed to Japan'sstrength in world markets for memory chips.At the same timeone legacy of the industry'sroots in devices for consumer applicationsJapanese product lines remain more narrow-ly based than those of the leading Americansuppliers. Microprocessors are a case in point;the major Japanese firms all continue to pro-duce American designs. NEC, Toshiba, Mit-subishi, and Oki sell members of the Intel 8080

205


family; at least three Japanese firms are build-ing Intel's 16-bit 8086.4° Although several Japa-nese manufacturers have designed microproc-essors For internal use, these have not foundother markets. And if made-in-Japan dynamicRAMS now claim a major share of worldwidesales, the overall Japanese presence in theUnited States remains modest. In 1982, importsfrom Japan accounted for about 51/2 percent byvalue of total U.S. integrated circuit consump-tion; although increasing rapidly (fig. 26, ch.4), Japanese imports remain small in absoluteterms. Still, the inroads have come in a marketsegment that American manufacturers right-

"R. I I. Sil in, The Japanese Semiconductor Industry: An Oyer-viet.s' (Hong Kong: Bank of America Asia, Ltd., January 1979),p. 148 "Background of VI,SI War With United States Reviewed,"japan Report, Joint Publications Research Service JPRS L/10662,July 16, 1982, p. 43.

ly view as critical; U.S. firms heavily depend-ent on memory products have been severely af-fected. In other product categories, Japanesecompetition is also stiffening; a major effectwas to further depress prices and profits dur-ing 1981 and 1982, when domestic firms weretroubled by a deep recession.

In the longer term, American semiconduc-tor manufacturers have every reason to bewary of continued pressure from powerfulmultinationals with headquarters in Japanfirms that have already demonstrated their abil-ity to compete successfully in major world mar-kets for other technically demanding products..The U.S. merchant manufacturers have theirown advantagesthey do well some things thatJapanese firms do poorlybut they cannot ex-pect an easy time of it in the future.

ComputersIf American manufacturers were for many

years unchallenged in world markets for semi-conductors. the United States has been stillmore preeminent in computers. Even in Japan,American-owned firms continue to account forover 40 percent of mainframe sales; the U.S.share of the Japanese market for small systemsis lower, but such firms as Data General andHewlett-Packard have recently established pro-duction facilities there. In Europe, U.S. com-panies are far out front except in'the UnitedKingdom, where the government has activelysupported ICL through procurements and sub-sidies. American-owned enterprises accountfor nearly three-quarters of all computer salesin Europe.

This section again concentrates on theUnited States and Japan. While Japanese com-puter manufacturers have not yet proven no-tably effective competitors outside their homemarket, they are at present uniquely situatedto launch a campaign aimed at the U.S. posi-tionin part because of their newly acquiredstrength in microelectronics, in part becauseof their active. pursuit of joint venture ties with

suppliers in Europe and the United States. Itis too early to predict the extent to which theJapanese strategy may succeed, but structuralchanges in the world computer industry arecreating new opportunities for firms every-where. The Japanese will probably be able toexploit at least some of these, certainly betterthan European producers.

The Environment for U.S. Suppliers

By virtually any standard, the United Slateshas far and away the most computer-intensiveeconomy in the world, a position it can expectto maintain indefinitely. From the early daysof the industry, the number of computers in -.stalled in the United States mounted at a pacethat kept the total about an order of magnitudegreater than for all of Western Europe.'" By1981 there was a computer terminal for every48 people employed in the United States; by

"Gaps in Technology: Electronic Computers (Paris: Organiza-tion for Economic Cooperation and Development, 1969), p. 16.

200

1

Ch. 5Competitiveness in the International Electronics Industry/ 201

1986 there should be 1 for every 10.42 Americanleadership in design, production, and salesas tell as utilizationis reflected both in tradedata, where the computer industry continuesto be a prodigious net exporter, and in theprominence of U.S.-owned subsidiaries inother parts of the world.

Strategic Patterns

For many years the story of American su-premacy in the gtobal computer industry wasthe story of one companyInternational Busi-ness Machines. Although IBM trailed Reming-ton Rand, builders of the Univac, in marketingearly computer models, by the late 1950's IBMhad gained the huge lead it still enjoys. For theother old-line firmsincluding Burroughs,NCR, Honeywellcompetition has mostlymeant jockeying for places in the residual mar-ket left by IBM; these companies have foundit difficult to reach the scale needed to offera full product line and to support a sales/serv-ice organization competitive with IBM's. Ingeneral-purpose mainframes, IBM has ac-counted for 60 to 70 percent of the worldmarket over the years, with lower figures insuch countries as Japan and the United King-dom balanced by even higher percentages else-where. To 1- .; sure, numerous entrantsmostlyAmericanhave attempted to carve out com-petitive positions against IBM, with muchmore success in rapidly growing markets forsmall systems than in mainframes. Companiesranging from Digital Equipment Corp. or Con-trol Data in the United States to Fujitsu inJapan and Nixdorf in West Germany have es-tablished themselves solidly in some portionsof the market. But none has come close toIBM's overall sales, despite the rapid shifts inoverall market structure described in the pre-vious chapter.

Most of IBM's American competitors havetaken a straightforward approach to their situa-tion: following IBM's lead in the developmentof faster and larger systems, trying to maintain

421- M. Branscomb, "Computer Communications in the Eight-iesTime To Put It All Together," Computer Networks, vol. 5,1981, p. 3.

99-111 0 - 83 - 14

product lines that match up reasonably, wellwhile at the same time staking out their ownterritory.e.g., Control Data in high-perform-ance scientific machines, Burroughs in smallbusiness systems (on the latter, see the casestudy in app. C). In these efforts, Americancomputer firms have been aided by the tech-nological lead of the U.S. semiconductor indus-try. Although IBM has relied heavily on inter-nal semiconductor design and manufacture,other firmswhether or not maintaining cap-tive production facilities--have been able totake advantage of components available on themerchant market that were often superior byconventional yardsticks to IBM's devices. Thisis one of the reasons IBM has itself begun topurchase ICs on the outside. A major elementin the strategies of other mainframe suppliers(excluding those making plug-compatible ma-chines) has been to expand into new applica-tions while tying their installed base to propri-etary softwarethus keeping old customers;None have had more than limited success;other mainframe-oriented firms have general =,ly been a good deal less profitable than IBM,and have made little headway in eroding IBM'Smarket share. Several have done better abroad;Honeywell's joint venture in France has beena greater force in the European market thanthe parent has been in the United States.

The Impacts of Microelectronics andReliability Improvement

Because the market has enlarged and changedso radically, focusing on the older mainframecompanies hardly gives a fair sampling of cur-rent strategies. As figure 35 indicates, marketgrowth for general-purpose mainframesthemainstay of the industry just a few years ago=is now much slower than for other types of sys-tems. As sales of minicomputers, small busi-ness installations, and desktop and personalmachines exploded, competitive dynamicsaltered fundamentally. Newer entrants havestaked out major shares of markets for productslike word processors. These structural shiftsare continuingindeed accelerating.

What are the implications for internationalcompetitiveness? As in semiconductors, prod-

207

?02 International Competitiveness in Electronics

..'&p,i,ocessqrs

. .

Z.

Figure 35.Market Segmentation of U.S. Computer Sales by Value

'

1975

Wordprocessors

10%

.1 .

businesssysterris...

Minicomputers.

13%21% .

1985(projected)

1980

SOURCE: "Moving Away From Mainframes: The Large Computer Makers' Strategy for Survival," Business Week, Feb. 15, 1982, p. 78.

208'

"'

Ch. 5 Competitiveness in the International Electronics Industry 203

uct. planning decisions in the computer in-dustry are shaped by technical possibilities.Specialized machines of all sizes ave long hadtheir place, but the turn toward small and ver-satile computers is a direct consequence oftwin driving forces: advances in software fornetworking and distributed rocessing, plusadvances in microelectronics. Computer sys-tems need no longer be structured around asingle processor. A central unit can be sur-rounded by a number of satellites, or the en-tire processing load can be shared throughouta network. Given cheap microprocessors Findsingle-chip microcomputers, designers can put"intelligence" where they want it. Computerfirms that fall behind in such developmentsmore broadly, manufacturers of systems incor-porating machine intelligence, not all of whomthink of themselves as part of the computer in-dustrywill be poorly placed to compete in fu-ture markets.

Improvements in system reliability--flowingpartially but not wholly from the growingreliability of microelectronic devicesyieldanother powerful driving force. Mean times be-tween failure for computer systems have beenincreasing steadily despite greater complexi-ty. Today's computers are orders ofmagnitudemore dependable than those of even a decadeago; this not only cuts operating and main-tenance costs but helps expand applications.Computers can now be used in a host of ap-plications where the dangers of a failure wereformerly too greatreal-time air traffic control,electric, utility load management, critical in-dustrial processes.

Ever-greater reliability combined with ever-greater computing power per dollar has eatenaway at IBM's traditional strengths-7-cuatomersuppOrt and service, plus the ability to lock incustomers via a product line broad enough tosatisfy virtually every need. Now, so many ap-plications have opened up that no one com-pany can hope to cover them all; in many ofthese, IBM's mcrket powerso valuable in sell-ing general-purpose mainframeshas been, ifnot irrelevant, al least a far smaller advantage.New entrants can specialiie in systems for

banking or distributed word processing. Start-ups of earlier yearsData General, Prime, Tan-demhave become substantial multinationalsin their own right. Independent software ven-dors are creating a whole new industry.

Better reliabilityin addition to broadeningthe applications of computershas had a sec-ond, equally important, impact. As in con-sumer electronics, it has allowed ,manufac-turers to skirt traditional distribution channelsand reach customers through new outlets. Thistrendwhich began as early as the 1960's withsystems houses that purchased minicomputersand peripherals in quantity, assembling them,together with software, to supply turnkey in-stallationsalso promises to continue and per-haps accelerate. Greater reliability has reducedthe need for onsite maintenance and repair;where field service was once a vital element .

in any marketing strategy, smaller manufactur-ers are now less constrained by the need to fi-nance service networks. Moreoverwhile hob-byists, engineers and scientists, and many bus-inesses could be reached through specializeddistribution channelsselling personal or desk-top computers to the general public requiresretail distribution. This, in turn, is realistic onlyif the need for aftersales service is modest. Cur-rently, the personal computer market is mov-ing through'a transition paralleling the earliershift in color TVdesktop machines are be-coming off-the-shelf items rather than productssold and serviced by specialty outlets. The per-sonal computer is a product in which IBM hadno great advantages beyond name-recognitionand abundant internal resources for productdevelopment. While these are far from trivialassets, IBM will probably not be abletO dupli-cate its position in mainframes in the far morediverse and competitive desktop market, justas the company has not been able to do as wellin small business systems, supercomputers, orword processors. The general point is: to be inthe computer business no longer necessarilymeans confrontation with IBM; it need not en-tail attempting to cut into the installed base ofany mainframe manufacturer, much less try-ing to match IBM's hardware or softwareacross a broad spectrum of products.

204 International Competitiveness i-

.Even in small systems, only a few companieshave been able to span a major portion of themarket. Somee.g., Hewlett-Packardhavespecialized in powerful machines for sophis-ticated customers. Others, like Wang and Data-point, have aimed at business applications.Digital Equipment Corp. (DEC)which startedas an OEM suppliernow has a relativelybroad product line, including personal com-puters. As in the semiconductor industry, man-agers have had to search for market opportuni-ties that match their organizations' strengths.Their choices and decisions, constrained by re-source lir stations and conditioned by govern-ment Ellilicies, will determine the future com-petitive posture of the U.S. industry. Mistakeswill be made, and weaker companiesmostlikely those making peripherals, office automa-tion equipment,fand desktop computerswillfind their ..,esibeing absorbed or merged withcompetitors.

Product Strategies: Hardware and Software

The mainframe-oriented companies do retainadvantages in structuring complex and far-flung inforthation systems. Designs for suchsystem's are often shaped by existing softwareinventories. The original supplier has an easiertime achieVing compatibility; indeed, computerfirmshave had a good deal to gain by makingit difficult for competitors.to reverse-engineertheir software and develop compatible systems.Some have-gone so far as to replace portionsof the system software with "firmware" storedin ROM chips which can be changed from timetntime. Generally, such efforts have been in-tended to thwart plug-compatible manufac-tvi ers.

iniportanc'e of software extends far be-yond the system level. Machines capable ofdata processing for business needs are nowwithin the financial reach of even the smallestfirmsand, of equal significance, can be.usedby people with little special training. Softwarethat is user-friendly, as well as reliable, is a keyelement in selling to those without previousdata processing experience. As the case studyin appendix' C points out, credit for the suc-cess of IBM's System/32 small business corn-

tl

puter goes in large part to the specialized ap-plications programs that were available. Evenmore, as hardware costs fall, specialized soft-ware becomes the pacing factor in applicationsranging from office automationwhere muchof the competition for word processor salesrevolves around softwareto industrial robots.Limited growth in software productivity andhigh associated costs (ch. 3) are problems thatnow confront all firms in the industry, here andoverseas; among the possible solutions are iiiul-tinationalized software generation. In thefuture, the importance of software comparedto hardware can only increase; the exceptionsare perhaps at the very high and very low ends,where supercomputers remain hardware-inten-sive and small machines selling for less thana thousand dollars compete on the basis ofprice.

21 u

From a slightly different perspective, soft-ware can become a constraint: switching tonew software, particularly system software, istime-consuming and expensive. Customerswith extensive data processing installationsand large software inventories become lockedin because of the high costs of transferring.This is a constraint on the system manufactur-er as well, who may be burdened with obsoles-cent software that cuts into potential perform-ance. The picture is somewhat different forcomputers sold to purchasers who are techno-logically adept---e.g., OEMs who integrate com-puters into their own products, or those withneeds in engineering or science. Such custom-ers commonly have the internal resources forsolving their own software problems, and findshifting to new systems, though a difficult task,not an insurmountable one. Still, given theirsoftware investments, virtually all customershave strong motives for replacing or augment-ing their equipment with new models from thesame manufacturerand manufacturersstrong motives for ensuring software compati-bility within their product lines. Therefore,once markets begin to mature, a manufactur-er's share of the installed base becomes a goodindicator of future prospects; competitors needhardware that is substantially better or cheaperto stand much chance of convincing customersto switch allegiance. Brand loyalty has been

Ch. 5Cornpetitiveness in the International Electronics Industry 205

high in general-purpose data processing mar-kets, largely for such reasons.

Plug-compatibility is aimed at breaking thiscycle. Originally referring to peripherals suchas disk drives, plug-compatible manufacturers(PCMs) later moved into mainframes that can.operate on IBM software; now some buildequipment compatible with DEC minicom-puters or IBM Personal Computers. Basically,the PCM strategy has been to make equipmentthat can be used interchangeably with IBM's,while undercutting the latter's price/perform-ance ratio5.43

The forces outlined above shape the strat-egids of companies striving to keep up in themarketplace. New approaches to product de:veloprnent are appearing throughout thedustry; even IBM has begun to purchase morehardware outside, as well as software. Inanother new departure for the company, IBMhas 'started selling disk drives on an OEM basis.As in semiconductors, technology-sharugagreements have become more commoncross-licensing of patents, direct purchases oftechnology, joint developmentas firms con-serve resources through specialization. This isthe idea behind Microelectronics & ComputerTechnology Corp.spearheaded by ControlData and presumably aimed not only at oncom-ing competition with the Japanese but the con-tinuing struggle of smaller entrants with 113M.Movement toward technology purchases andtechnology sharing appears to have even moremomentum in Europe, despite earlier failuresof joint efforts like .Unidata. Siemens, ICL, andOlivetti are among the companies now market-ing Japanese-built mainframes in Europe.

International Aspects

The picture that emerges in the U.S. com-puter industry is one in which the long-dom-inant leader is being challenged on all sides.Structural change has been driven largely by

',The founder of Amdahl. the leading supplier of PCM main-frames. has said that surviving in competition '9M requiresCosts that are 15 percent lower or performance that is 20 per-cent better. See "Makeshift Marriage." The Economist, Aug. 11,1979. p. 78.'

the technologyalthough occasionally marketdemand outstrips what the industry can sup-ply, as happened with word processors andit is difficult to predict where it May lead. Someobservers ly.11ieve that IBM's market power willcontinue to deteriorate, even in areas wherethe firm's position has heretofore seemed un-assailable. Others think the future lies withlarge and powerful companies able to combinefar-flung communications and information net-works into vast integrated systems. In fact, bothviews are probably correct, given the fragmen-tation and specialization brought by cheaphardware.

American computer manufacturers, livingnervously with rapid technical change at home,face another series of choices in foreign mar-kets. Governments in industrialized nationswhere American subsidiaries have long beendominant continue to follow policies trans-,parently intended to reduce that dominance.Such policies are nothing new: France's PlanCalcul was set forth more than 15 years ago,and the Governrrients of Great Britain and Ja-pan have, over the years, found many ways tosupport local computer manufacturers. Whilemost such policies have had only limited ef-fects in the past, certainly in Japan the tech-nological fervor is now intrmse.

If competition from Japanese computer firmsis on the rise, American entrants are them.selves fashioning new international strategies.Already DEC operates six plants in Europe andthree more in the Far East. A partial list cifother American minicomputer manufacturerswith foreign production facilities would In-clude Hewlett-Packard, Wang, Data General,Datapoint, and Texas Instruments. U.S.-basedmultinationals specializing in desktop ma-chines include Apple, with plants in Irelandand Singapore, and Tandy. Manufacturers pro-ducing plug-compatible mainframeshave alsobegun to expand abroad: Amdahl has openedan Irish facility intended in part to supply theCommon Market, as has Trilogy Systems.

The rules of the competitive game are in par-ticular flux in lesser developed parts of theworld. Developing countries are putting in-

21i

206 International Cornpetit,'Ynness in Electronics

dustrial policies to work attracting technologyand foStering local computer manufacturing.Mexico's appriach has been to restrict lInports,limiting sales to companies that agree to estab-lish production facilities. With an annual mar-ke: now approaching $500 million; Mexico hasbeen able to attract a pair of U.S. minicomputerfirms willing to live with these rules. Brazil'sGovernment has reserved the domestic mini-computer market for locally controlled enter-.prises; transfers of technology have been en-couraged, but foreign investments are limitedto minority interests. While American compa-nies have generally chosen to stay out, severalEuropean and Japanese firms have agreed toparticipateno doubt hoping for benefits sim-ilar to those now flowing to the Japanese,con-SU mer electronics manufacturers that acceptedsuch conditions in earlier years.

How' will the onset of local production indeveloping countries affect international com-petition in computers? While any answer' re-mains conjectural, it would be foolish to dis-miss the possibility that some of these nationsmay evolve. into viable forces -in the market-place. Although their ability to compete willprobably be restricted to simpler products over;'.';: foreseeable future, developing economieswill begin by building equipment such as ter--minals, printers, and disk drives, where laboris a major cost element. It is not a big step frommaking TV, receivers to producing the simplertypes of computer terminalsindeed a stepthat countries like Korea and Taiwan have al-ready taken. With the experience. gained insuch products,. and with protected '.marketscontributing to scale economies, a number ofthe newly industrializing countries could movefairly quiCkly into world markets.

As a final point, again consider .software de-velOpment. By its nature, programing hOF', beenlabor-intensivetherefore increasingly costlyin high-wage nations. Software generation de-pends on people with ability and experienceincluding an understanding of the problemsfaced by users. Such faCtors have prevented thetransfer of this work to developing countries,even those like. India where the raw program -ing skills might be available. Nonetheless, sev-

eral industrializing nations are attempting toimprove the capabilities of their labor forcesso that they can produce software needed inadvanced economies. Countries like Singapore,Hong Kong, and Taiwan are seeking to create"software centers" where Western computermanufacturers could establish subsidiaries thatwould transfer skills and provide training forthe local work force while also producingmuch-needed software. Once the people wereavailable, locally owned companies could takeover at least some of the work.

Japan

_Objectives announced by Japan's Govern-ment over the past few years herald 'a com-petitive onslaught directed at the U.S. com-puter industry. MITI is sponsoring a pair oflong-term R&D projects dealing with computersystems, plus several related efforts." The fifthgeneration computer projectthe origins andorganization of which are described in chapter10is software-intensive, directed at artificialintelligence, information organization andmanag mt, and natural language input and

t l :e ond project, MITI is helpinglune one development of a supercomputer in-tended to surpass the most powerful offeringsof American companies like Control Data andCray. A related 10-year project will supportdevelopment of the high-speed microelectron-ic devices needed to implement the softwareconcepts of both fifth-generation machines andsupercomputers. The goal is nothing less thanto thrust Japanese companies into the forefront_of world computer technology, to leapfrog theUnited States in the design and marketing ofboth hardware and software. The objectives of

"Outline of Research and Development Plans for Fifth Genera-tion Computer Systems (Tokyo: Japan Information ProcessingDevelopment Center. Institute for New Gederation Computer

echnology. May 1982): Computer White Paper: 1981 Edition(Tokyo; Japan Information Processing Development Center.1982). pp. 59-75: "Machinery. Information Industries '81 Pro-grams Outlined." Japan Report, Joint PubliCations Research Serv-ice JPRS L/10086. Nov. 2. 1981. p. 21: "Archety" e of Fifth Genera-tion Computer Described." Japan Report. Jo:, , Publications Re-search SerVice JPRS L111007 Dec. 14. 1982. p. "MITI Proj-ect To Develop Supercomputer Starts in January." Japan Report.Joint Publications Research Service JPRS L110348. F h. 23. 1982p. 34.

Ch. 5Competitiveness in the international ElectrclnIcs Industry 207

these programs are by no means unique toJapanese manufacturers; they are squarely inthe mainstream of the evolution of computing.It is the strength of Japan's commitment thebacking by MITI and other government agen-cies, the I0 -year schedulesthat differentiatethem from efforts in other countries.

As the rhetoric associated 'ith such pro-grams makes clear, Japanese !I ,ms, with thehelp of their government. hope within 10 or 15years to lead the world in computer technol-ogy. Despite Japan's 'alatively successful ex-perience with previous government-sponsoredR&D effortsthe Pattern Information Process-ing System (PIPS) project, the VLSI project---:this is a tall order. At present, the market pcsi-tion of Japanese manufacturers is modest; asof 1981, American-owned companies heldmore than three-quarters of the world com-puter market in value. terms, Japanese-ownedcompanies only about 7 percont;Still, Japan isnow the second largest supplier of gener,a12purpose computers lo the world market, will,a very high rate. of export growth (ch. 4, fig.30). The country is also second only to theUnited States in intensity.of compuxer utiliza-tion. After the experiences of consumer elecironies, automobiles, and semiconductor mem-ory chips, few in the American industry wouldtake Japan's goals lightly.

Nevertheless, because of ,lie role that factorssuch as installed base and s )ft ware inventoriesplay in the marketing of computer systems,Japanese manufacturers must begin with theknowledge thatno matter how good their

\ technologythey cannot hOpe to come closeto the United States for many years. In thisSense, the computer market is not at alilike thatfCir semiconductors, where purchasers quick-ly switch suppliers to take advantage of lowpriCes, quick delivery. or new device types.Success in niche markets. for computer systemsis quite possible, indeed a necessary first step,but breadth in an industry expanding in asmany different directions as information proc-essing can only be a long-term undertaking.The U.S. position, both in technology andmarket share, is simply too strong. Leaders inJapanese\Government and industry recognize

\their weaknesses, 4nd have made plans accord-ingly.

Technology

Carefully targeted R&D isle central strand inthe Japanese computer strategy, as in earlierventures into other industries. Japanese pro-ducers and their government realize, just asthey did in microelectronics, that international"competitiveness in computers cannot be at-tained so long as they rely\\on technology fromthe United States. The reasons are twofold.First, American firms are far ess likely tolicense technologies than in past. The Jap-anese know that computer Jnanufacturers inthe United States, unlike at least some of theirpredecessors in other sectors, are acutelyaware that technical leadership is a primarysource of competitivij strength, and that tomake 11,eir !nclinology too easily availablewould weaken their own poSition. The secondreason is even more fundamental. In the basicbuilding blocks of computer hardware, semi-conductors, Japanese firms are near parity withAmerican companies; in some areas they maybe ahead. Japan can hardly depend on im-ported technology; rapid progress toward aneventual goal of leadership in information /processing requires extensive indigenous capa-/bility of the sort that Japanese firms now havein high-density memory chips.

The Japanese also recognize that their short-comings in the marketplace are not so muchmatters of hardware as of software and relatedapplications-based constraints. Several japa-nese firms now offer computer hardware aspowerful as any. However, IBM's Mige in-stalled base and vast catalog of applacationsprograms have forced Japanese competitors,as those elsewhere striving to break` into themainframe market, to build plug-compatiblemachines that run on IBM softv4re. To getaround this impediment is perhapS the majorreason for the fifth-generation project. Whilecompanies like Amdahl have demonstratedthat a comfortable business can be built sup-plying PCM mainframes, markets tied toanother manufacturer's software are inevitablylimited. Japan's gamble is that it can jump

21J


ahead of American entrants with families ofcomputer systems having performance capa-bilities that will render present-day softwareinventories obsolete. This goal has shaped.thehardware and software R&D planned for thefifth; - generation project: to take full advantageof emerging microelectronics technologies inmore closely linking the needs and abilities ofpeople with the capabilities of the system. Ifthese objectives are met, individuals--eventhose with little trainingwill be able to com-municate with fifth-generation machinesthrough ordinary language in spoken or writ-ten form, as well as through graphical ();tonal bnages, Such systetn: would not onlybe user-friendly, but might,ultimately displaysomething of the independent decisionmakingcapability associated with human reasoning.*If the technical objectives of the fifth-gen-eration projectand similar efforts in -othercountriesare achieved, even novice userswould be able to harness enormous computingpower. The commercial potential is immense.

Government AssistanceThe Japanese Goverriment has supported

R&D activities in information. processing overMany years. MITI has been selective in finan-cial aid, directing funds to potential bottle-necks, exemplified by the VLSI program's sup-port for digital ICs, or to R&D that could helpJapan's industries leapfrog the competition, theintent of the fifth-generation computer project.Funding for the latter is projected at about.$500million over a 10-year span (1.981-91); the super-computer project is expected to get another

An example from the field of artificial intelligencethe areaknown as expert systemswill illustrate. Research in expert sys-tems aims at computer programs that mimic attributes of peo-ple who are "experts" in some realm of knowledgee.g.. medi-cine. where such programs might help automate diagnosis. Theobjective would generally be to augment or complement ratherthan supplant human skills: an expert system would not havethe judgement of a physician, but could offei, for example. per-fect recall of vast amounts of information. Expert systems typ-ically depend on complex software and large data bases; thus.advances in hardware as well as software may be needed if theyare to be widely implemented.

$100 million over roughly the same period.**A parallel miercelectronics projectwhichgoes' by names such as "R&D on New Func-tion Elements"has a budget ef about $150million and is scheduled to run from 1980 to1990. Money v,-;t1 go to three major develop-ment efforts:45

Three-dimensional circuit elementswhich can be visualized as more-or-lessconventional ICs stacked atop oneanother, increasing the density.High clef trc ri mobility transistors!HE'N,`Tsi, obe \,..iriety of which consists ofvery thin layers of semiconducting mate-rials such as gallium arsenide or galliumaluminum arsenide; HEMTs offer poten-tially higher switching speeds, hence fastercomputers.Radiation-hardened devices suitable foruse in extreme envirumntsrits such as nu-clear powerplants or outer space (resist-ance to heat and vibration is a related ob-jective).

The first two especially will support bothsupercomputer and fifth-generation projects.Among related government-sponsored pro-grams, another of major significance for thecorporate strategies of Japan's computer man-ufacturers has aimed at the development ofsoftware and peripheral devices with Japaneselanguage input-output capability. Scheduledover the period 1979-83, nearly $200 millionwas allocated to this effort."

As in microelectronics, R&D is but one ofmany ways in which Japan's Government as-sists the computer industry. The Japan Devel-opment Bank loans money to the Japan Elec-tronic Computer Corp. (JECC), a jointly heldfirm which purchases computers from partic-ipating manufacturers and leases them to

**Planning for the fifth-generation program began severalyears earlier, as outlined in ch. 10. A variety of funding levelsfor both projects have been reported: spending plans and sched-ules will no doubt shift as they progress.

""FY82 Government Projects in Electronics Listed." JapanReport, Joint Publications Research Service JPRS 1110676, July22. 1982. p. 55.

"Computer White Paper: 1981 Edition. op. cit., pp. 4ff.

21

Ch. 5.Competitiveness in the International Electronics Industry 209

Photo credit: IBM Corp

Memory cells in experimental Josephsonjunction integrated circuit chip

users. Manufacturers can set up tax-free re-serves to offset losses incurred when lease con-tracts with ) ECC are canceled and equipmentmust be repurchased. Since 1979, tax-free re-serves have been permitted for up to half theincome associated with some categories of soft-ware. Purchasers of certain types of computerscan write off 13 percent of tliFf value, beyondnormal depreciatiOn, in the first year. The gov-ernment has also established special deprecia-tion schedules for high-performance remotedata processing equipment.

A panopoly of support measuresof whichmany mor- examples could be citedhas thusbeen desigaed to help Japanese companiesachieve technological superiority and commer-cial success in the 1990's. At first glance, thesums of money involved may seem large; infact, when viewed in the context of the worldcomputer industry, they are modest; as chapter10 stresses, it is the consistent support providedby many individual measures acting in concertthat gives Japanese industrial policy its impact.

To place the expenditures of the JapaneseGovernment in perspective, table 46 lists R&Dspending by a number of U.S.-based computerfirms. On an annualized basis, subsidies pro-vided by Japan's Government for R&D in in-formation processing come to less than the ex-penditures of any one of these American com-panies. (Total subsidies for the information in-

Table 46.Resoarch and Development Expendituresby Several U.S. Computer Manufacturers, 1981

R&D spending(millions of dollars)

Burroughs $220Control Data Corp. 202Digital Equipment Corp. 251Hewlett-Packard 347Honeywell 369IBM 1,600SOURCES: Annual reports

dustries in Japanincluding indirect supportthrough tax preferencescould only be esti-mated by making a large number of essential-ly arbitrary assumptions; see ch. 10.) MITI'sR&D subsidies are also modest in comparisonto the research budgets of Japanese companies.Fujitsu spent $260 million on R&D in 1981,while Hitachi and NEC spent $610 million and$230 million, respectively.'" The governmentmoney does have an important function: help-ing with the kinds of long-term R&D that -in-dividual companies might otherwise have dif-ficulty in justifying. In addition, MITI-spon-sored projectsthough not cooperative in theusual senseattempt to stimulate creativethinking, technology interchange, and the com-plex of synergies so vital to engineering re-search. The Japanese electronics industry prob-ably benefits more from these factorswhichtend to be lacking within the laboratories of in-dividual corporationsthan a strict compari-son of funding levels would suggest.

Of course, ether governments also provideassistance to their computer industries, not ex-cluding the United StateS. European nationsroutinely channel direct financial aid to localcompanies, along with indirect subsidiesthrough procurement and tax benefits. Hand-some incentives designed to attract in-vestments and technology have been dangledbefore the European subsidiaries of U.S. andJapanese companies. In the United States,funding by the Department of Defense throughthe Very High-Speed Integrated Circuit' pro-gram and this country's own supercomputerprojectstill in the planning stageswill have

"Annual reports.

2 1 J


commercial spillovers. The lesson is that no in-t ustrial ized nation has l,,,3en content to accepta !au:011(1m.y position in technologies and mar -ker m considered essential to future ec;onoinicdeve!opment..The concern,is that the Japanesemay bo more successful in implementing theirpolicies than other countries.

Marketing Strategies andMultinational Operations

Individual computer manufacturers in Japanhave begun to formulate product strategiesbased on technologies expected to flow fromgovernment-supported R&D projects,, as wellas their internal activities. Marketing c.om-puters presents special difficulties because thechief competitors are already well entrenched;only in peripheral equipment such -as ter-:ainals, printers, and disk drives have Japanesemanufacturers made a significant impact out-side their home market. In Europe and theUnited States, where nearly two-thirds of the,world's computer systems have been installed,Japanese companies are inconsequential as in-dependent suppliers.

In an industry where salas depend on a thor-ough grasp of user needs at a technical levelsoftware as well as hardwarelate entry is amajor handicap. American suppliersinclud-ing newer participants like DEC and Hewlett-Packardhave built networks of sales and serv-ice centers staffed by engineers and techni-cians who now have longstanding ties withcustomers; IBM has such a network within Ja-pan. Even those.Japanese firms with strong in-ternational positions in microelectronics ortelecommunications cannot match the distribu-tion systems of U.S. computer manufacturers.To make much progress, Japanese entrants willhave to invest substantial sums over manyyears without the expectation of immediatereturns. The history of fields like consumerelectronics indicates that at least some lap-anese companies will be willing to make thiscommitment.

Perhaps surprisingly, given the global per-spectives of consumer-oriented firms like Mat-sushita, the Japanese computer industry as awhole suffers from a pronounced lack of in-

ternational business; experience compared toAmerican firms of even suite modest size.Until recently, most (but not all) Japanese com-panies have preferred to manufacture at homefor export; this could prove a major weakness111 computers. While a few Japanese electronicscompaniesToshiba is anotherhave ex-panded aggressively via overseas investment,most of Japan's past international successeshave come in products where integrated man-ufacturing and marketing in foreign countrieshas not been essential. The examples includesteel, automobiles, and semiconductors; con-sumer electronics is only a partial exception.In each of these cases, Japanese companies, atleaSt in the beginning, concentrated on exportsales. Generally able to take advantage of es-tablished distribution systems, they investedoverseas only when import restrictions com-pelled local manufacturing. The competitivepressures that led U.S. semiconductor or com-puter firms to invest in Europe and elsewherehave on!.y recently begun to impinge on theJapanese. By now, American computer firmsnot only operate wholly owned sales and serv-ice networks in many parts of the world, theyhave established internationally dispersed andintegrated manufacturing operations partly inresponse to governmental demands and part-ly due to the nature of the market. Japanesefirms, on the other hand, have been largely un-willing or unable to make the enormous in-vestments required to participate in the worldmarketplace for computers.

Managers of Japanese firms, along with bu-reaucrats within the government, recognizetheir lack of background and experience, andare seeking remedies. The international (as op-posed to R&D) strategy appears to be an in-cremental one, geared to minimizing the re-sourct, at risk and taking advantage of existingstrengths. As part of this strategy, Japaneseelectronics firms, with the.encouragement ofMITI, are beginning to establish manufactur-ing plants in other industrialized countries.Following investments by Japanese consumerelectronics suppliers in the United States andelsewhere, tentative steps have been taken insemiconductors, a market in which the Japa-nese have already become well entrenched


through exports. These same semiconductormanufacturers are of course the major com-puter firms. Experience gained from invest-ments in semiconductor production will helpin structuring multinational computer opera-tions.

As a parallel step, Japanese manufacturershave established marketing links with a num-ber of foreign firms. Fujitsu now furnishesSiemens (West Germany) and ICL (Britain) withlarge mainframes, while Hitachi has similar ar-rangements with BASF (West Germany) andOlivetti (Italy). Fujitsu has taken a minority in-terest in SECOINSA of Spain, while agreeingto a technology transfer tie with a companypartially owned by the Brazilian Government.In the United States, Fujitsu holds a minorityinterest in Anidah:, the PCM pioneer to whichit exports large machines; for several years, Fu-jitsu distributed its smaller systems within theUnited States through a joint venture withTRW. National Advanced Systems, a subSidi-lay of National Semiconductor, sells Hitachicomputers here.

Such arrangements build from the fact ofJapanese parity in hardware for large com-puters, parity which does not extend to soft-war,3; both Fujitsu's and HitaChi's systems areIBM compatible. European firms have beenunable to attain the economies of scale that'.Japanese manufacturers get in their home mar-ket, and have chosen to compete with Ameri-can producers by importing from Japan. Fromthe collective viewpoint of Japanese firms,these ventureseven where the equipment islabeled with some other brand-nameincreasemarket exposure and add to production scale.

. For'some time, such relationships will continueto be essential elements in the marketing strat-egies of at least several of Japan's computermanufacturers. Even so, they link companiesnone of which has more than a minor shareof the global computer market. Siemens, ICL,and the other partners of Japanese firms to-gether do not account for even 5 percent ofworld computer sales. Wth the possible excep-tion of ICL, none has a ;Gale of operation anddistribution approaching that of the competinglocal subsidiary of IBM. None is strong in

minicomputers or small systems. Moreover,the Japanese participants remain a critical stepremoved from the customers whose applica-tions their equipment is intended to servejoint ventures will provide limited help at bestin remedying past weaknesses of Japanesefirms in software or customer support andservice. To become viable international com-petitors, Japan's computer companies will needto accumulate experience in dealing directlywith the requirements of customers inparketsWhere they hope to sell.

Computer manufacturers in Japan do notshare these problems in equal measure; the in-dustry is far from monolithic. Fujitsu, at themoment in a clear leadership position (ch. 4),has, along with Hitachi, chosen to stake its in-ternational position on supplying IBM-com-patible equipmentdecisions that will limitboth companies' options for many years tocome. NEC has taken a different route, devel-oping its own system software (although deriv-ative of U.S. technology), Nor has. NEC yetentered into marketing arrangements with for-eign concerns. Instead, the company's manage-ment appears to be shaping a strategy intended,to take advantage of the overlap and mergerof computer and communications technolo-gies, areas where the company is already prom-inent. Despite its relatively small size comparedto other Japanese electronics firms, much lessIBM or AT&T, Nippon Electric's managers areattempting to position their organization forwhat they see as an eventual competitive 'strug-gle with these two American giants for domi-nance of the international information indus-try.

At several points above, the entry barrierscreated by the well-established sales and serv-ice networks of American firms have been de-scribed. This aspect of the market for com-puters effectively turns one of the supposed ad-vantages of the Japanese system on its head.Barriers erected by government to keep outforeign firms have given Japan's manufacturersadvantages in a number of industries, partlythrough scale economies. Closed markets cre-ated by import restrictions and foreign invest-ment controls have been reinforced by corn-

2 1

212 . International Competitiveness- in Electronics

plea distribution structures and a deeply in-grained "Buy Japanese" attitude. In computers,however, the longstanding customer ties main-tained b' I ;.S. firms combine with technologi-cal strengths to create formidable entry barriersfor Japanese companiesindeed, new entrantsfrom any part of the world. Windows do openbecause of technological advance; throughthese vindows newcomers have moved intomarkets for microcomputers, small businesssystems, and other specialized products. Thusfar, most -of these entrants have been Americanfirms -in Fart because the U.S. market is solarge, but also bectiuse American companiescontrol the distribution apparatus in most partsof the world. The going will be difficult forJapanese manufacturers, although they are be-ginning to find niche productsdesktop corn-.paters may be onesuited to their strengths.

In medium and large systems, Japanese corn-.panies can choose from a number of alternative(or complementary) cotOses of action.One isto continue to build joint relationships withforeign enterprises. As noted above, suth astrategy will require, first, deeper involvemen'::with -end users by the Japanese participants,.and, second, movement into markets in moreparts of the world. If firms such as Burroughs,Control Data, or Honeywell were to be en-ticedeach has a relatively small but well-established market sharethe prospects forJapanese firms would look a good deal better.The constant pressure of trying to achieve costs.comparable to IBM's could well force one ormore American companies to accept such ties.

As an adjunct to joint marketing-ventures,Japanese manufacturers will probably seekother ways of incrementally expanding sales,while awaiting the fruits of the fifth-generationcomputer project. If Japan succeeds in pioneer-ing a new generation of hardware and soft-ware, companies with multinational produc-

_,

tion and marketing experience will be able toexploit the new technologies most effectively.In this context, present efforts would not beso important in themselves; rather they wouldbe preparatory steps for rapid growth in the-1990's. Another path, one that some firms willcertainly pursue, is to concentrate on sellingsmaller systems and personal machines. Herethe now-traditional Japanese entry strategy isfeasible because distribution networks are opento all comers. Thus far, attempts to challengeAmerican companies like Apple or Tandy inpersonal computers, or the many U.S. entrantsin the market for small business systems, havenot been notably successfulin the UnitedStates or elsewhere. Still, if and when suchproducts become more nearly -standardizedand interchangeable, Japanese companiescould expect an easier time. But even if com-panies based in Japan were to expand intothese .ma;:kets, it is not at all obvious that thiswould help them in ether types of systems.

Japanese producers of computers are thustaking what seem the only paths available intheir attempts to break into the world market:independent technology development coupledwith joint marketing relationships. That the.marketing ties involve firms that are them-selves weak-and in need of partners is hardlysurprising, but makes the establishment of aviable international presence that much moredifficult. At this point, the Japanese have hadonly marginal impacts on global markets; athome, IBM-Japan remains a formidable com-petitor. Whether or not technical developmentsin ,microe.l.ectronics and software will thrustJapan into a position nearer the forefront re-mains to be seen. If Japan's computer manufac-turers do begin to increase their market sharessignificantly, the most likely victims will besmaller competitors first in Europe, then per-haps in the United States.

Summary and ConclusionsWhile international competitivenessin any

industrydepends on many factors, the busi-ness strategies pursued by private corporations

are central. Costs of labor and capital, techno-logical resources, government policies, humanresource endowmentsall can, at least in prin-

218

Ch. 5Competitiveness in the Internaii:nal Electronics Industry 213

ciple, be looked on as forces impinging on man-agement decisions, as features of the landscapefor business tactics and strategies.

While a useful perspective, the strategic viewof competitiveness is nonetheless an imperfectsubstitute for more quantitative indicators. Un-fortunately, the swiftness of technical changein electronics precludes useful quantitativemeasures. Productivity trends mean littlewhere the standard products of todaywheth-er semiconductors or mainframe computershave capabilities that may be orders of magni-tude beyond those of a decade past. Compara-tive manufacturing costs carry weight in somecases, but not where one company can buildproducts exceeding the reach of competitors.Little meaning attaches to patent statistics assurrogates for technical ability when incentivesfor acquiring patents vary widely among coun-tries and nowhere correlate very closely withqualitative aspects of technology.

If shifts in international competitiveness can-not be extracted from statistical series, carefulexamination of business activities can yield in-sights into future prespects as well as pasttrends. In semiconductors and computers, notto mention consumer electronics, Americanfirmsonce undisputed leaders in technology:as in sales in their home markets and virtliallyaround the worldface much stronger compet-itive pressures. Foreign enterprises, most13,Japanese but also entrants with headquarter;in other Far Eastern countries, are sellinglarger volumes of electronics products withinthe, United States; American corporations arehaving a more diificult time in foreign markets.The sources of these shifts are many. By-and-large, they are not due to mistakes or faultystrategies by American firms or by the 'U.S.Government. First and foremost, rising foreigncompetition flows from continued rebuildingof the electronics industries of Europe andJapan in the aftermath of World War II. It isnot a new phenomenon. By the mid--1050's,when much of the basic reconstruction of over-seas economies was complete, companies inJapan and much of Europe found themselvesstill well behind the United States in their abili-ty to design, develop, and produce electronics

products. But they were in a good position tocatch up. The first signs of success came early,when Japanese manufacturers like Sony cre-ated new families of transistor radios smallerand lighter than those offered by Americanfirms. The transistor was invented in theUnited States, the first transistor radios alsomade here, but Japanese firms pushed theirproduct development efforts vigorously andoutstripped their U.S. rivals within a few years.Now that Japan is in the lead with new genera-tions of consumer products it will be difficultfor American or European manufacturers toregain the lost ground.

In computers and semiconductors, WesternEurope came out of the war well ahead ofJapan. The Europeans had good fundamentaltechnology, but were stymied by small andfragmented markets, as well as by manage-ments that had neither the resources nor thevision of their counterparts here. Subsidiariesof U.S. corporations became the backbone ofthe European computer industrythey stillareand took the lead in microelectronics. Inthe Japanese market, American Prrn5 could notmatch the'ir accomplishments in Europe because of the protective policies of Japan's Gov-ernment. Still; if not dominant, the UnitedStates wasand remainsa major force in Jap-anese computer sales, particularly for largemachines, as well as in some types of semicon-ductor devices. Continued efforts by Americanfirmsbacked if necessary by the U.S. Govern-mentto participate on equitable terms withinJapan, whether by exporting or by direct in vest -ment, appear vital for maintaining U.S. com-petitiveness in electronics. The Japanese elec-tronics market is large and still expandingrapidly; it is now more important than Europe.

Japanese industrial policy has been a moresignificant source of support in semiconduc-tors and computers than in consumer elec-tronics. The MITI-sponsored VLSI researchprogramWhile not as important as someAmericans have claimeddid help Japanesefirms master process technokigies for verylarge-scale digital ICs. In standard devicefamilies like memory chipswhere the path oftechnological evolution is clear for all to see,

.214 International Competitiveness in Electronics

and technological success a function more ofpainstaking development and detail designthan highly creative engineeringthe Japanesehave excelled. While they cannot as yet matchthe breadth of American product lines, theywill certainly continue to improve their capa-bilities in circuit design as well as processing.If U.S. semiconductor manufacturers can ex-pect intense competition, they too have theiradvantagesa different set than those of Japa-nese firms. If American companies continueto capitalize c.i these strengthsthe besttrained engineers in the world, quick recogni-tion and response to market needs, innovationsin circuit design, applications of computer-aided techniques, specialized products pursuedwith entrepreneurial zealthe United Statesshould be able to maintain a leadership posi-tion. Still, American companies will not be ableto monopolize world sales as they did a decadeago.

Competitive pressures, evolving technology,and growing capital intensityalong with thecontinuing expansion of captive production byintegrated firmsare changing the structure ofthe U.S. semiconductor industry. New struc-tures bring new strategies. Structure is chang-ing in the computer industry as well, drivenby the technology of computing, itself depend-ing heavily on microelectronics. As computingpower becomes ever cheaper, more and moreapplications become cost effective. These at-tract new firms, designing and developing notonly peripherals and software, but specializedprocessorsminicomputers, personal anddesktop units, business systems. While themainframe is hardly a dinosaur, a "computer"can now be a great many thingsmany neverenvisioned by the designers of the general-purpose machines sold two decades ago by asmall number of companies such as Univacand IBM. Computing power is now cheap andwidely dispersed, often invisible to users. Asdistributed processing and data communica-tions continue to spread, new firms will try toestablish themselves, entering through win-dows of technological or market opportunity.Some of the older firms will find themselves

hard-pressed to keep up. even survive; theirmanagers will face hard choices in allocatinglimited resources. Few companieseven in-cluding the largest, here or in Japanwill beable to cover more than a small fraction ofproduct markets.

While no one can foretell competitive out-comes in the world computer industry, it is ob-vious for all to see that Japan has made a seriesof explicit decisionsgoing back as far as the1960's and involving both government and in-dustryaimed at claiming a major share ofsales and applications. Based on past perform-ance in other sectors of electronic.: (he prob-ability of continued expansion by the majorJapanese computer manufacturers is high. Be-cause the characteristics of the market for dataprocessing equipment differ from those forsemiconductors, the United States remains ina stronger relative position, There is no reasonwhy the United States cannot continue to holdan overall lead in both technology and sales.

As events. in all three portions of the elec-tronics industry demonstrate, competitive posi-tions in global markets have shifted, more-or-less continually over time. Some firms in someparts of the world rise, others decline. Of thosethat decline, a few may eventually revive,others disappear. No country can expect all itsindustries to thrive in international competi-tion; any nation that trades will be more com-petitive in some industries than others, theleaders in competitiveness shifting over time.That U.S. competitiveness has slackened inconsumer electronics does not imply that sim-ilar events will follow in other sectors. Thiscould happen, but there is no reason to expectdeclines in microelectronics or computers par-alleling those in color TVparticularly so longas the technology continues: its rapid evolutionand markets expand at high rates. These areconditions under which American firms havetraditirmally prospered. When the pace ofevents slows, other sectors of the Nation'seconomy might begin to find themselves far-ing better than electronics in interactional

acompetition.

CHAPTER 6

Manufacturing: Quality,Reliability, and Automation

221

Contents

PageOt.ervicr.s. 217

Qua lit:. and Reliability 219, and N1easortinienl. 219

;Ind \lonaging le,. :)uality 221in:i..0:.fanc,.! of Design 222

Appri.ycli 294E,.]..dbility of Intgraf,,!,..i 926

. i:eilability of Color TVs 231

Automation 233anti Flexible :Automation 234

Llectronics NI.anufacture 235239

:summary and Conclusions 246

Appendix GAQuality and Reliability Comparisons for Integrated Circuits 247

List of Tables

T-Ildt. NO. Page17. t:auses of Field Service Failures in Color TV Receivers 22348. Typical Costs of Detecting and Repairing Faulty Components in an

Electronic System 22949. RankingS by Repair Shops of TV Receivers for Quality and Reliability 232i0. Reasons Given by Japanese Electronics Firms for Automating 2366A-1. Hewlett-Packard Data on 16K Random Access Memory Circuits 2486A-2. Quality Levels of Japanese and U.S. Random Access Memory Circuits 2496A-3. Reliability Levels of Japanese and U.S. Random Access Memory Circuits . 249

List of FiguresFigure Pago36. Data for MOS RAMS Showing Constant Failure Rates per Circuit as

Integration Levels Increase, Decreasing Failure Rates per Bit 22037. Reliability Improvement With Cumulative Production Experience

for a Microprocessor 22138. "Typical Trends in Failures Attributed to Design and Production 22339. Typical Failure History for Semiconductor Devices 22740. Testing Sequence for Point-of-Sale Terminals 22941. Reliability Trend for Analog Integrated Circuits 23042. AutoMatic Installation of Integrated Circuits Onto Ceramic Substrate 23743. Average Labor Hours for Assembly of 21-Inch Color. TV Receivers - 23844. Two Approaches to the Design of Industrial Robots 23945. Manufacturing Costs for Robots, Hard Automation, and Human

Operators as a Function of Production Volume 24146: Cost Comparison for Human Operator and RcbotAssuming One-for-One

Replacement and Two-Shift Operation yi-the. Automobile Industry 24147. WOrldwide Annual Sales of Robots,-Past and Projected 24248. Robot Sales in Japan by Application 24349. Estimated Numbers of..Robots in Use, 1980 244

222

CHAPTER 6

Manufacturing: Quality,Reliability, and Automation

OverviewAssuming comparable productsand a lack

of subsidies or other strong exogenous influ-encescosts are a primary determinant of in-ternational competitiveness, in electronics asin any industry. Comparable products may notbe identical, and small differences in perform-ance or specifications can override small dif-ferences in costs in the eyes of customers. Buteven for military systems, manufacturing costswhich depend on both the design of the prod-uct and the design of the production systemare almost always a major consideration.

In electronics, costs are much more criticalto the successful marketing of some types ofproducts than others. Intense price competi-tion in consumer electronicstelevisions, vid-eo cassette recorders, stereo equipmentmakes low manufacturing costs a vital com-petitive weapon. Much the same is true forstandardized semiconductor products rangingfrom discrete transistors to random accessmemory chips; price cutting is the rule, costshighly sensitive to the yields of the productionprocess (ch. 3). For other types of semiconduc-tor devices, low manufacturing costs and lowprices are less vital; if only a single firm makesa particular integrated circuit (IC)perhapsone that meets unusual or demanding perform-ance requirements such as a high-speed, highresolution analog-digital converterit willprobably set prices to maximize profits, giventhe lack of competition. Leading-edge comput-er hardware and software falls in much thesame category. Even so, electronics firms areseldom able to establish and maintain techno-logical advantages so large that manufacturingcosts are of little relevance.

-In addition to direct and indirect manufac-turing costs, prices charged to purchasersreflect expenses associated with research, de-

99-111 0.- 83 - 15

sign, and development, as well as marketingand distribution. While accounting proceduresvary, such costs are generally treated as in-direct expensesi.e., a percentage is added tothe direct manufacturing cast of each item pro-duced, as for other overhead. Depreciation ofplant and equipment is handled the same way.Direct manufacturing cost then consistsprimarily of parts, materials, and labor. Re-search, design, and development costs aremuch higher for products such as computersor large-scale ICs than for consumer itemswhere technical change is slow and incremen-tal, major redesigns infrequent. In the produc-tion of semiconductor devices and computers,research and development tends to account for

"'a considerably greater percentage of costs; de-preciation charges are also likely to be greaterbecause new production equipment must bepurchased as the technology advances.

But costs are not the only way in which man-ufacturing operations affect competitiveness.Beyond production costswhich depend onwage rates, prices of materials, supplies, andcomponents, capital charges, and related fac-torslie dimensions such as the quality and thereliability of the goods produced. While moresophisticated purchasers are most interestedin the quantifiable dimensions of quality andreliability, in markets for consumer productsperceptionswhether or not well foundedin-fluence the decisions of prospective customers.Along with other qualitative aspects, such asappearance, purchasers base their assessmentsof value for money ,on perceptions of qualityand reliability.

These attributes bath the reality and the per-ceptiondepend on factors such as engineer-ing design, how the pebple in the work forceare trained, organized, and managed, and on

9 0 0217


the capabilities, indeed the quality, of themanufacturing equipment. Automation can itn-proi e quality by reducing the probability ofhuman error or simply improving the con-siste ._y of the production process. In other in-stances, there will be no effect. In some cases,quality may be degraded; people are better atsome jobs than machines, and vice versahu-man skills far exceed those of machines fortasks involving pattern recognition, or wherejudgments must be made based on partial orimperfect information. Regardless of specifics,the quality of a firm's products will ultimatelydepend on the stress top management placeson quality as a goal of the production process.

By the end of the 1970's, issues of productquality were in the public eye for industries asdisparate as nuclear power, automobiles, andsemiconductors. Perceptions were widespreadthat the quality of American goods had de-clined compared to those from foreign coun-tries.1 Some observers speculated that Ameri-can firms and American labor had slipped,others that consumers had become more de-manding and were no longer satisfied by qual-ity standards that had once been acceptable inthe U.S. market. Either way, a "quality gap"with respect to imports, extending even to com-modity items such as steel, has frequently beenadvanced as a contributing factor in the declin-ing international competitiveness of Americanfirms and industries. To take an example fromelectronics, the reputation of RCA's color TVshad slipped badly by the end of the 1960'5.2 Not

'OTA's contractor on quality and reliability notes that about80 percent of the attendees at an April 1980 American Manage-ment Association seminar on Japanese techniques for qualitycontrol and productivity improvement felt that the products oftheir own firms were surpassed by products from Japan. Thosesurveyed were at the seminar to learn from the experience ofJapanese companiesnot a random sample. See J. MihalaskYand A. B. Mundel, "Quality and Reliability of Semiconductorsand CTVs: United States v. Japan," report No. C972, preparedfor OTA by Consultant Services Institute, Inc.. under contractNo. 033-1170.0, p. 6.

'R. A, Joseph, "Automation Helps RCA and Zenith Keep Color-TV Leadership in Face Of Imports," Wall Street journal. MaY5, 1981, p. 56.

only did this hurt the company's sale: , RCAalso lost some of its dealer base. Automatedproduction was at the heart of the company'seffort to improve the quality and reliability ofits TV line.

Despite the importance of direct costs of pro-duction for competitive success in electronics,OTA has not attempted to estimate or comparemanufacturing costs. Companies guard costdata closely. More important, the dynamics ofshifting cost structures, rather than costs at agiven point in time, are central to ch-mgingcompetitive fortunes. To some extent thesedynamics can be inferred without the need forproprietary data. This chapter then focuses onaspects of manufacturing such as quality andreliability, plus automated production technol-ogies.

Product quality is treated primarily from ahardware perspective: What are the relativelevels of quality in the United States and Japan?(The comparison is limited to these two coun-tries.) How do product design and the applica-tion of production engineering and quality as-surance techniques affect quality? How doproducts fail?3 Less tangible but equally impor-tant matters of the human element in man-ufacturing and quality controlincluding ques-tions of management and organization, as wellas the education and training of the work forceare also discussed in chapter 8.

'Much of the material on quality and reliability assembled be-low is drawn from "Quality and Reliability of Semiconductorsand CTVs.: United States v. japan," op. cit. This report is basedin part on a series of questionnaires and surveys-200 coveringboth manufacturers and purchasers of ICs, 60 coveringindependent TV service shopsplus 42 visits to facilities of U.S.and Japanese firms that make ICs or semiconductor manufac-turing equipment.

While comparisons between products of American and Japa-nese firms were of primary interest, some of those surveyed alsocommented on the West European electronics industry. In gen-eral, the feeling was that European firms had been behind bothAmerican and Japanese manufacturers in the quality and relia-bility of their ICs and TV receivers. While European producersmay recently have caught up to the United States in the qualityand reliability of certain types of semiconductor devices, overallthey probably still lag both the United States and. Japan,

2.2.1

Ch. 6Manufacturing: Quality, Reliability, and Automation 219

Quality and ReliabilityMeanings and Measurement

Quality, meaning fitness for functionthe ex-tent to which a product meets the specifica-tions of its designers and manufacturers, theexpectations of users can be treated subjec-tively or objectively. Consumers typically makesubjective judgments concerning the quality ofcompeting products. Manufacturers attempt todefine quality in terms of parameters that canbe measured quantitativelye.g., the ability ofa TV set to receive weak signals. In addition,they may adopt rating scales which trained in-spectors apply to characteristics that are notinherently quantitative. Often the indices arebased on comparisons with samples or stand-ards; an example would be the appearance ofthe cabinet for a TV setwhether the trim fitproperly and colors matched, whether thepanels were free of waviness, the number ofvisible flaws and blemishes.

Through measures like these, manufacturerstry to satisfy consumers' perceptions of quali-ty as well as ensure that their products func-tion properlyall at a reasonable cost. Forproducts like ICs or computers, the quality im-age that a firm establishes is likely to be morenearly consistent with quantitative measuresthan for consumer goods; indeed, some elec-tronic components are sold to specifications

I


Probes for testing integrated circuit chips

written by the purchaser. Nevertheless, theperceptions and subjective judgments of cus-tomers sell many computers, and ICs are in-spected to be sure that logotypes and partnumbers are properly printed and convey thedesired image.

Reliability is a measure of continuing fitnessfor function once a product is placed in serv-ice. While quality is determined at a singlepoint in timegenerally the end of the manu-facturing process or the beginning of servicelifereliability is measured over time, as afailure rate or similar parameter.

The most common indicators of reliabilityare mean time to failure or mean time betweenfailuresthe interval between disabling fail-ures averaged over a large number of items,usually in terms of actual hours of operation:Failures that average one per million hours canbe expressed as a mean time between failuresof 106 hours or as a failure rate of 10-6 perhour. The graphical presentation in figure 36shows the number of ICs (from a much largergroup) expected to fail in 109 (1 billion) hoursof operation. A billion hours is 114 centuries;such plots are constructed from short-term datausing statistical techniques. A failure rate ofone per billion hours means that the expectedor most likely lifetime for a single item chosenat random is 109 hours.

Definitions of reliability based only on fail-ures that prevent the product from function-ing are straightforward. Measurement cannonetheless be time-consuming, as well as pre-senting difficult statistical problems. Still, alight bulb works until it burns outtesting alarge enough sample. of nominally, identicalbulbs will yield a statistically valid mean timeto failure. Partial failure, or gradual degrada-tion in performance, is more difficult to quan-tify. A 10-year-old TV set may still function, butnot' as well as,when new; there are no simplemeasures of "reliability" that apply to suchphenomena.

223


_Figure 36.Data for MOS RAMs Showing ConstantFailure Rates per Circuit as Integration Levels

Increase, Decreasing Failure Rates per Bit

100 Failure rates per circuit

10

1

10-2

to-3

Failure rates per bit

1970 1972 1975 1977(256) .(1K) 14K) (16K)

Year(bits)

SOURCE T. Goto and N. Manabe. low Japanese Manufacturers AchieveHigh IC Reliability." Electronics, Mar. 13, 1980, p. 140.

The reliability of computer software presentsanother type of problem. Software does not"fail" or "wear out" from physical causes asdoes hardwarealthough the media that storethe software may suffer such failures. But soft-ware still needs continual maintenance; com-plex programs are altered and updated, period-icallysometimes to correct errors that arecaught only after the software has been placedin service, other times to improve performance.The reliability of a piece of software then de-pends on the frequency of modifications re-quired to correct programing errors that couldcause the system to malfunction," As a result,

-J. D. Musa, "The Measurement and Management of SoftwareReliability," Proceedings of the IEEE, vol. 68, 1980, p. 1131. Moregenerally, see R. Dunn and R. Ullman, Quality t:ssurance forComputer Software (New York: McGraw-Hill, 1982). New com-puter programs tend to have of'the order of one mistake per hun-dred lines; some but not all of these will be found before theprogram is placed in service M. Lipow, "Number of Faults perLine of Code," IEEE Transactions on Software Engineering, vol.SE-8, July 1982, p. 437.

the reliability of an entire computer systemdepends on both hardware reliability and soft-ware reliabilityfailures of the first type hav-ing physical causes (although ultimately de-pending on design and manufacturing prac-tices), failures of the second type dependingwholly on the design of the software.

Exhaustive engineering efforts are directedat ensuring the reliability of new and complexsystems of all types, particularly where failurescan be costly or dangerouse.g., airplanes ornuclear powerplants. To improve reliability,designers apply techniques such as failuremode analysisestimating the probabilities ofdifferent typeS of failures and attempting tominimize the more serious. A common prac-tice is to add redundancy to the system, pro-viding functional alternatives so that the failureof one part does not compromise the whole.A wire rope has a great deql of redundancy be-cause one strand, or many strands in a largeenough cable, can break without impairing thestrength significantly. A chain, in contrast, hasno redundancy. Complex computer systemsoften include redundant processors and otherhardware components, as well as fault-tolerantsoftware that can reconfigure the system fol-low:ng hardware failures. In any type of sys-tem, degraded performance will normally bepreferable to sudden and total failure. For ex-ample, elertr nnic cot 'rel systems for auto-mobile en ,Ines are designed so that componentfailuresperhaps of a sensor or a memory chipwill not cause the engine to suddenly stoprunning. Instead, the engineers aim for "soft"failure modes, or "limp-home" capability.

While quality and reliability are related, theyare by no means synonomous. Reliability cle7pends more heavily on design engineering,qualityon control of the manufacturing proc--.ess. In general, as experience in making a prod-uct accumulates, levels of quality and reliabilityboth increase, Note the similarity with yield in-creases for semiconductors, as discussed inchapter 3. Figure 37 illustrates the reliabilityimprovement over time of the Motorola 6800microprocessor, a popular 8-bit circuit that hasbeen in production since 1974. _

22


Figure 37.Reliability Improvement WithCumulative Production Experience for a

Microprocessor (Motorola 6800)

10

1974 :1976 1976 1977

YearSOUPCE D Projeciing %%SU s Impact on Microprocessors, IEEE

.47 1919 p 38

Statistical methods can be applied to quali-ty and reliability problems during both thedesign and manufacturing stages, but the spe-cialized discipline of statistical quality controlis largely a tool of the production process.5 Asan example, defects in ICs can be monitoredover time to give insight into the effects of proc-essing variables. In contrast, reliability analysistechniques are applied, not to process variablesbut to tests conducted on finished products andto field service experience. Steps take: to im-prove quality sometimes but not always im-prove reliability.

1978 1979

Organizing and Managing for QualityManaging the interface between design en-

gineering and manufacturing engineering pre-sents a classic set of problems that affect pro-duction costs, as well as quality and reliability.Designers specify the characteristics of prod-ucts in great detail, while manufacturing en-gineers must determine how lo make the prod-uct so that it will have those characteristics.Sometimes the same people are involved inboth functions, but more commonly the re-sponsibilities fall on different parts of an or-ganization.

'See. for example. J. M. Juran, F. M. Gryna, Jr., and R. S. Bing-ham, Jr. (eds.), Quality Control Handbook 3d ed. (New York:McGraw-Hill. 1974), especially secs. 22-27.

Separation of responsibility for design, pro-duction, and quality control characterize man-ufacturing enterprises all over the world, butperhaps more so in the United States than else-where (e.g., in Japan). One reason for the prev-alence here appears to be the heritage of scien-tific management, an approach to job methodsand the organization of production originatingin the work of an American engineer, Freder-ick Taylor, during the early part of the century.6

Production engineering includes all the tech-nical aspects of the manufacturing process:plant layout, process design, work methods, se-lection of equipment, quality assu-ance. Inlarger firms some of these functions may notonly fall in different departments, but be fur-ther subdivided. Still, regardless of organiza-tion charts that isolate the design, manufactur-ing, and quality functions from one another,these activities are closely related functionally.?Product design affects the choice of manufac-turing technology. The equipment that a firmhas on hand, together with the costs of invest-ing in new equipment, can severely constrainthe design of its products. Inspection and test-ing, quality and reliability, depend not only onthe choice of manufacturing technologies, buton the overall control of the process. Applica-tion of statistical quality control techniques toindividual steps in manufacturing may bestraightforward, but overall integration andcontrol of a complex production process ismuch more than the sum of control cf the in-dividual steps.

Although design and production are inher-ently interdependent, in some cases even sim-ple communication is lost. Stories of designand production supervisors who are not onspeaking termsor the commonplace of the de-sign group "tossing the drawings over the

"See. in particular. Quality Control Handbook op. cit., sec.48 on "Quality Control and the National Culture," which pointsout that the sharp divisions of responsibility typical of largerorganizations in the United Statese.g., separate departmentsfor quality control or inspactioncreate reservoirs of specializedexpertise, but at the same time may hinder the widespread ap-plication of this expertise. Scientific management is discussedin more detail in ch. 8.

'J. A. Alic, "Manufacturing Management: Effects on Produc-tivity and Quality," Efficiency of Manufacturing Systems (NewYork: Plenum Publishing Corp., 1983), p. 281.

227


wall" to the manufacturing departmentarerife. There is at least anecdotal evidence thatforeign firms may handle, not only the prob-lems of training design and manufacturing per-sonnel, but of managing the interface betweendesign and production, better than manyAmerican companies. One approach is to makethe same individuals or groups responsible forboth design and production, or at least extendmanagement responsibility for integrating de-sign and manufacturing farther down into theorganizatior.al structure.8 In Japan, for exam-ple, companies often rotate design engineersthrough production departments early in theircareers.8 Not only do Japanese electronicsfirms tend to stress integration of product andprocess design within their organizations, butthey frequently involVe vendors, distributors,and customers in the wc r;s of their manufactur-ing er gineers.

V..ie some American firms have grappledwith such problems more successfully thanothers, companies here begin with a fundamen-tal handicap: low, prestige and low pay tend tobe associated with white-collar jobs in manu-facturing relative to other categories of engi-neering or management; the best people are sel-dom attracted to such jobs. Manufacturing car-ries higher status in European or Japanese cor-porations. And, on the manufacturing side ofan American firm, quality control tends to beal the bottom of the pecking order. Too oftenit seems that manufacturing managers see qual-ity control only as an obstacle to production.

American management has -been criticizedfor overemphasizing the costs of quality,whereas some quality control professionalsargue that a comprehensive program for de-

, signing and building quality (and reliability)into a product at all stages can save money.Again, there seems to be a contrast with thetypical attitude in Japanese companiesdis-

°R. E. Cole, "The Japanese Lesson in Quality," TechnologyReview, July 1981, p. 29. See also "Sources of Japan's Interna-tional Competitiveness in the Consumer Electronics Industry:An Examination of Selected Issues," report prepared for OTAby Developing World Industry and Technology, Inc., under con-tract No. 033-1010.0, pp. 103-104.

°J. M. Juran, "Japanese and Western QualityA Contrast,"Quality Progress, December 1978, p. 10.

2 2 h

cussed in more detail belowwhere preventionof defects is emphasized more strongly than de-tection through inspectiun.

One reason for the lo.. status of manufactur-ing in the United States is simply the low priori-ty that industry places on it, as indicated bylow pay scales in manufacturing relative toother parts of the firm; engineers employed inmanufacturing and quality control get salariesnear the bottom of the range for their age andexperience groups at all points during theircareers; engineers doing administrative workearn 50 percent more than those involved inproduction."

Another indication of lack of attention tomanufacturing is that only 4 percent of grad-uates of engineering technology programs inthe United States specialize in the "manufac-turing, quality control, industrial" category."Engineering technology is a relatively new fieldintended to provide practically oriented train-ing meeting the needs of industry (see ch. 8 forfurther discussion of technology education),this it is particularly surprising that such asmall fraction of graduates are oriented towardcareers in manufacturing. In engineering pro-grams, so few U.S. graduates receive degreesin manufacturing that they are not separatelythulated. Although students who have studiedmechanical or industrial engineering often findmanufacturing jobs, many programs in thesefields have dropped the once common requiredcourses in such topics as manufacturing pipc-esses and plant layout.

The Importance of Design

Figure 38 contrasts schematically the effectsof design and manufacturing on reliability. Re-liability tends to improve with production ex-perience, but failures stemming from designweaknesses sometimes show up only after longperiods in service, hence may even increaseover time. Such behavior is typical of many

10R. Connolly, "Career Outlook," Electronics, June 16, 1981,p. 266.

1,P. J. Sheridan, "Eugineen es and Technology Degrees, 1982,"Engineering Education, April 1983, p. 715. The percentage isthe total for associate and bachelor levels.


Figure 38.Typical Trends in Failures Attributedto Design and Production

Attlbutedto design

Attributedto manufacturing

Time from product introductionSOURCE Quat,%' and Re/43bil,ty. of Semiconductors and CTVs. United States

Ja;,an ,epc1 No C972. prepared for ON= by Consultant Sentces In.51,tute, Inc, under contract No 03711700. p. 18

types of manufactured products, not just elec-tronics.

Table 47 indicates the extent to which the re-liability of color TVs depends on design ascompared to production. According to thetable, .a greater percentage of service failureshave their sources in the design and develop-ment process than in assembly. One of the rea-sons that Japanese TVs achieved better relia-bility than American-made sets during the

,Af197 ' appears to have been more conservativedes gn practice. For example, Japanese setswere designed to draw less power. Picturetubes operated at lower voltages, with somesacrifice to picture quality but lower internaltempe..atures and less stress on components.In some contrast, the vice president for engi-neering of an American TV manufacturer, nowtaken over by a Japanese firm, has been quoted

Table 47.Causes of Field Service Failures inColor TV Deceivers

AttributiOnPercentage offield failures

Design (and development) 20-40%Quality of components 40.65%Final assembly 15-20%SOURCE: J. M Juran, "Japanese and Western QualityA Contrast,- Quality

Progress. December 1978, p. 10.

as saying, "At Warwick, much of the designwork happened after the product was intro-duced. We relied on field failure informationto tell us where we had a problem."12

According to the estimates in table 47, abouthalf the failures in TVs are due to defectivecomponents. Some of these may be purchased,others manufactured internally---some compo-nents fail because they themselves suffer fromdesign problems. Many of the components ina television receiver are transistors or ICs. Asillustrated by figure 36, failure rates per chiptend to remain about the same as circuit den-sity increases. If so, going to higher levels ofintegration and increasing the number of cir-cuit functions per chip will have two impor-tant consequences. First, it will cut assemblycosts because the total number of componentswill decrease. Second, there will he fewer com-ponents to fail, hence reliability should im-prove. The cost and quality/reliability advan-tages of chassis designs based on fewer butmore complex ICs have led to rapid reductionsin the numbers of components in TV receivers.In 1977, Zenith's 25-inch color TV contained685 components. Less than 2 years later, thenumber had been reduced to 441.'3

As part of their strategic thrust into the U.S.market, Japanese consumer electronics firmsset nut to create an image of high-quality, re-liable products (ch. 5). They needed trouble-freeproducts in reality as well as appearance inorder to exploit the distribution channels avail-able to them. Reductions in parts counts wereone of the techniques adopted. Likewise, by theend of the 1970's Japanese semiconductorproducts had attained enviable reputations for

12-American Manufacturers Strive for QualityJapaneseStyle." Business Week Mar. 12, 1979. p. 32B.

"Ibid. Over roughly the same time period, the Japanese-ownedQuasar firm reduced its parts count from 516 to 406. whileToshiba claims a 60-percent deCrease in parts count between1971 and 1979. Other Japanese firms have reported similar re-_ductions. typically coming earlier than for American TV man,ufacturers. For example, the number of parts in a particular Pan-asonic color TV model went from 1023 in 1972 to 488 in 1976-L-see "Quality and Reliability of Semiconductors and CTVs: UnitedStates v. Japan." op. cit., p. 47. Japanese TV manufacturers oftenpursued simpler chassis designs in parallel with the develop-ment of automated production facilities, as discussed later inthe chapter;

22 ad


quality and reliability. But manufacturers inJapan have not relied on design improvementsalone; emplorees of the large, integrated Japa-nese electronics companies tend to have con-siderably more training in quality control andproduction technologies than their counter-parts in the United States.

The Japanese Approach

Managements of Japanese electronics firmsprofess to believe that improvements in quali-ty and reliability will automatically cut costsand increase productivity, as well as aidingtheir marketing strategies. Tliv rhetoric ema-nating from top managers in japan emphasizesquality to a greater extent than statements byAmerican executives. More concretely, Japa-nese manufacturing companies rely muchmore heavily on line managers for quality as-surance, rather than the staff specialists com-mon in American firms.

Despite this and other organizational differ-ences, most of the methods that Japanese man-ufacturers use in pursuit of quality and reliabili-ty have been borrowed from the United States,just as for product technologies. Japanese in-dustrialists have been noted for their study mis-sions to visit foreign companies and researchlaboratories. Engineers and managers from Ja-pan have become skilled at picking out usefulideas from such visitswhether related toproduct technologies or to aspects of manufac-turing such as quality controland improvingon them. The theory and, practice of quality as-surance may have diverged more in the UnitedStates than in Japan.

Origins of Qualify ConsciousnessStress on quality and reliability within Japa-

nese manufacturing firms goes back at least tothe period of postwar reconstruction." Man-agers realized that Japan's exports were wide-ly viewed as cheap and shoddy. Much of theearly effort toward improving Japanese prod-ucts was orchestrated by the Union of Japanese

14"Quality and Reliability of Semiconductors and CTVs: UnitedStates v. Japan," op. cit., pp. 38-40. The historical material thatfollows is drawn Largely from this report.

Scientists and Engineers (JUSE), which helpedto locate foreign expertise in quality and reli-ability, and diffused this knowiedge throughpublications, training courses, and confer-ences. As many as 10 million workers may nowhave passed through JUSE training courses."

During the 1950's, well-known Americanssuch as W. E. Deming and J. M. Juran traveledand lectured extensively in japan; Juran, in par-Acular, is credited with much of the visibilitythat quality control now enjoys at upper man-agement levels in Japanese companies. Inmany respects, the quality control movementin Japan began at the top and r-pread down-wardin considerable contrast to the situationin the United States, where the principal ad-vocates of quality assurance have often beenlower level technical specialists. The well-known Deming Prizesestablished in 1951,and given to both companies and individualsfor achievements in quality controlillustratethe prestige of such activities; they are amongthe most coveted industrial awards in Japan."

Japanese executives like Hajime Karatsu,Managing Director of Matsushita Communica-tion Industries, have been quality control ad-vocates for years; the Reliability Center forElectronic Components of Japan was formedin the early 1970's at the urging of industryleaders, including Karatsu. Financed private-ly by more than 200 electronics firms, the cen-ter conducts tests on components and systems,establishes procedures for determining relia-bility, drafts specifications, and diffuses infor-mation on quality improvement within the in-dustry."

The Japanese emphasis on line responsibili-ty has led to extensive training programs for--assembly workers and foremen. Efforts/toreach the latter have included radio and TV

is"American Manufacturers Strive for Quality JapaneseStyle," op. cit.

leQuality Control Handbook op. cit.. sec. 48, p. 48.9. On theprominence of the Deming Prizes, see U. C. Lehner, "JapaneseFirms' Stress on Quality Control Is Reflected in Dogged Vyingfor Award," Wall Street Journal, Sept. 24. 1980. p. 52. There iseven a widely publicized "Quality Month" in japan.

11"Guide to RE]," Reliability Center for Electronic Compo-nents.


broadcasts; about 100,000 of the accompany-ing textbooks were sold in the first year (1956)of the radio series alone. A monthly magazineGemba -to -QC (QC for the Foreman), was estab-lished about the same time and evidentlyserved as a breeding ground for quality cir-clesa technique that has recently received agreat deal of attention in the United States (seech. 8). The first quality circle was registeredwith JUSE in 1962; within 15 years, member-ships in registered quality circles had grownbeyond 800,000. JUSE reports that about100,000 circles are now in operation, withabout 80 percent of the nation's blue-collar

ork force involved.18

Standards

In the United States, product standards tendto be voluntary, but Japan's Industrial Stand-ardization Law, passed in 1949, places the re-sponsibility with government. The law dealsexplicitly with quality and provides that allJapanese exports must carry the approval of theJapanese Institute of Standards (NS)." Con-sumers in Japan are also said to look for theJIS mark. In 1957, the Japanese Governmenttook a further step aimed at- upgrading-thequality image of the country's products, pass-ing the Export Inspection Law. This regulationcreated set of standards aimedmostly at smaller companies, and also providedfor the establishment of testing laboratories.

Organizing for Quality

Despite the visibility of quality circles, theyare only one tool among the many that Japa-nese electronics firms have adopted. Becausetraining in quality is widespread, and respon-sibility for quality assurance diffused withinthe organization, quality control departmentsin Japanese firms tend to be small compared

la"Quality and Reliability of Semiconductors and CTVs: UnitedStates v. Japan," op. cit., p. 60. Circles also enroll clerical andmanagement personnel. It has been claimed that the averagequality circle in ;apan saves an employer about $100,000 peryear.

p. 66. A numher of other Asian nations have followedthe Japanese example in trying to improve the quality image oftheir exports. In Taiwan, a small tax is levied to cover the costof inspection: the tax drops as quality levels go up. See "Amer-ican Manufacturers Strive for QualityJapanese Style," op. cit.

to the United States. Companies in Japan haveoften dispensed with some fraction of in-proc-ess inspectors, making each worker responsi-ble for accepting or rejecting the parts passingthrough his or her station. This is but one ex-ample of the diffusion of responsibility throughthe organization. It is effective in part be-causeat least in the larger companiesem-ployees are carefully selected even for un-skilled, entry-level jobs. Transfers of blue-collaremployees within the firm are commonapractice facilitated by unions organized on acompanywide rather than craft basis, and new-ly hired workers, or those transferred to an un-familiar job, typically pass through trainingprograms that are lengthy compared to thosein the United States. At Matsushita, for in-stance, new assembly line workers are givena month of trainingwith a week devoted toquality controlbefore they begin to work onthe line." In the United States, new assemblyline workers would typically get a few minutesinformal instruction by a foreman, who wouldthen monitor their performance as they learnedby doing. Both approaches have their advan-tages.

An apparent paradox has developed in thewake of the 30-year history of quality controlactivities in Japan outlined above. Many of theoriginal techniques imported from the UnitedStates were concerned with statistical qualitycontrola subject in which Deming and Juranwere authorities. Yet there is little evidence thatthe application of statistical teglniqUes to quali-ty or reliability has advanced, any further inJapan than elsewhere. In fact, applications ofstatistics are seldom mentioned in descriptionsof the quality control procedures of Japaneseelectronics firms. Rather, the Japanese appearto have focused their efforts on making individ-ual employees aware ofand committed tothe achievement of quality. Statistical qualitycontrol is no more than a small part of the

""Quality and Reliability of Semiconductors and CTVs: UnitedStates v. Japan," op, clt., p. 52. While three-quarters of the work-ers in Japanese electronics firms were classed as unskilled atthe end of the 1970's, the proportion of skilled as compared tounskilled employees is expected to rise rapidly as automationproceeds. Presumably this is an important motivation for thetraining programs found in many companies.

2q


quality programs of typical Japanese electron-ics manufacturers, which the companies them-selves often refer to as "companywide qualitycontrol." Intangibles and consciousness-raisingare at least as important.

Quality and Reliability ofIntegrated Circuits

Manufacturing and Testing

Chapter 3 outlined the steps in making ICs.Most of the larger American merchant firmsperform some but not all of these in domesticplants, with labor-intensive operations carriedout offshore. A typical pattern might be asfollows:

Operations performed in the UnitedStates:1. Silicon crystals, generally purchased

from outside vendors, are sliced intowafers and prepared for lithographicprocessing.

2. Wafer fabrication processes such as lith-ographic patterning, oxidation, etching,diffusion of dopants, metallization, andannealing are carried out; some of thesemay be highly automated.

3. Each of the hundreds of ICs (chips) ona wafer is tested; those that fail aremarked, typically with an ink drop.

Operations often performed in offshore fa-cilities:4. Individual circuits are separated from

the wafer, and the defective chips de-tected in step 3 discarded.

5. Each good chip is mounted on a sub-strate (chip carrier).

6. Lead wires are bonded to pads on thechip (the lead wires connect to externalpins, which plug into sockets installedon circuit boards).

7. The chip is encapsulated in a metal,plastic, or ceramic package that pro-vides mechanical and environmentalprotection (metal and ceramic packagesare normally heTmetically sealed),21

"For a more complete description of packaging and assembly,see A. B. Glaser and G. E. Subak-Sharpe, Integrated Circuit En-gineerirg: Design, Fabrication, and Applications (Reading,Mars.: Addison-Wesley. 1977). ch. 10.

1111 -7-1'..J- gin

2.11._

Photo credit: GenRed. Inc

Test equipment fir integrated circuits

8. The packaged ICs are subjected to func-tional tests.

Often the circuits are shipped back to theUnited States for the final testing in step 8, par-ticularly if destined for American rather thanthird-country markets. (Economic aspects ofoffshore assembly are outlined in app. B.)

Outcomes at all these stages in processingpurity of the silicon crystal, wafer flatness, lith-ographic precision, integrity of wire bonds,hermetic sealing can affect quality and relia-bility. Some are more important than others;patterning flaws and poor wire bonds areamong the most common causes of failures.22During the manufacturing process, inspection

"For a discussion of failure modes in semiconductor devices,see E. A. Doyle, Jr., "How Pails Fail," IEEE Spectres; October1981, p. 36. An important technique, particularly for, 'Suringreliability, is the analysis of failures. ICs that fail during tingor in service can be examined by a variety of methodse.,rect observation in a scanning electron microscopeandcauses of failure diagnosed. Corrective action, which might ravefrom a modified circuit resign to simple adjustmenxs In processparameters such as temperature, can then be taken. A compre-hensive treatment of reliability, emphasizing t.si importaine ofthe design of the circuit, is C. G. Poattie, et al., "Elemens ofSemiconductor-Device Reliability," Proceedings of the IE??E, vol.

62. 1974. p. 149.


and testing are possible at some points but notothers; in the absence of good methods for di-rect testing following-a-particular processingstep, the engineers must rely on control, ofprocess parameters based on downstream testresults.

Customer Requirements

Differing customer demands lead to a rangeof standards for the quality and reliability ofsemiconductor devices. Purchase agreementsoften specify the testing procedures to be fol-lowed. Military circuits must meet especiallydemanding .'specifications for resistance toshock, vibration, and severe environments (in-cluding radiatirm); reliability is emphasized forsatellite applications. Limited-volume marketsfor partsintended for military or space applica-tions are often served by small firms specializ-ing in,ultrahighleliability components. Whilesemiconductors for commercial markets have,seldom faced specifications as demanding asfor military and space applications, the actualfunctional requirementsparticularly for Ion-sgevitymay be at least as severe. For instance,'some computers operate virtually continuouslyfor years, albeit in a service environment thatis well controlled and benign; semiconductordevices for automobiles must function relia-blyalso over many yearsin an environmentCharacterized by vibration and extreme tem-peratures, as well as eXposure,to gasoline, oiland grease, rain, road salt, and do-it-yourselfrepair efforts.

Reliability estimatione.g., by acceleratedlife testingis costly, thus life testing of devicesintended for consumer products is minimal.Considerably more reliability testing is carriedout on parts destined for computers or com-munications systems. Because the service rec-ord of their products is critical for future sales,and because the costs of locating and replac-ing faulty parts are high, particularly after thesystem has gone into service, manufacturersof complex electronic systems demand reliablecomponents. This is one of the reasons firmslike IBM or Western Electric chose to buildmany of their own ICs. Regardless of applica-tion, however, the chip Manufacturer seeks a

production process sufficiently well controlledthat testing becomes simply a verification ofthat control.

Because of these varying customer demands,the electronic components industry has, sincewell before the semiconductor era, suppliedproducts to a range of quality and reliabilityspecifications; as many as half a dozen levelsdeveloped from the initial distinction betweenmilitary and commercial pmts. The lowestlevel has been for inexpensive consumer prod-ucts such as toys and games, ,the'-,highest forapplications such as communications satellites.Failure rates for the most reliable devices can-be more than a factor of a hundred below thosefor the least reliable.

Failuresin SemiconductorsThe time history of failures for a large pop-

ulation of semiconductor devicesas for man-ufactured products of many kindswill nor-mally follow a pattern like that in figure 39.,Early in life, the failure rate tends to be high,with most of the failures caused by randommanufacturing defects. The distinctions be-tween quality and reliability become rather r:r-bitrary during these early stages. A strict quali-ty- standard, for example, might weed ots'4: partsthat would otherwise fail during the infant

Figure 39.Typical Failure Historyfor Semiconductor Devices

TimeSOURCE: Ottice of Technology Assessment.

233


mortality perioci. "Burn-in" tests help detectinfant mortality failures; during burn-in, ICsare cycled to high temperatures and exercisedby computerized tasting equipment.23

Afte: the high failure rates early in life, fail-ure frequency usually declines to a nominallyconstant valuethe middle portion of the curvein figure 39. For Semiconductor devices, thisperiod typically spans hundreds of thousands,even millions of hours, during which the prob-ability of failure is extremely low. Eventually,the curve may turn up again as devices "wearout or otherwise deteriorate with age.

While semiconductor products do-not wearmechanically, they are susceptible to degrada-tion from environmental exposure, thermal cy-cling, and a variety of physical processes. Com-mon causes of long-term failures in ICs in-clude: loss of hermetic seal, with consequentdamage from moisture or other environmen-tal agents; thermal fatigue of the bond betweenthe chip and its substrate or of the lead wirebonds; gradual thinning and cracking of me-tallized layer due to electromigration associ-ated with high current densities (even thoughthe currents in ICs are low, the small conduct-ing paths result in high values of current den-sity). Failure probabilities associated with par-ticular mechanisms can be reduced by conserv-ative design at both device and system levels,a common tactic in applications such as satel-lites or submarine cables.

TestingTesting costs for ICs increase with levels of

integration. Although 100 percent testing iscommon during the early steps in fabrication,manufacturers normally screen their final out-put by random sampling; that is, only a smallfraction of the outgoing product is subjectedto a full battery of tests. Many customers dotheir own screening of incoming parts. On theother hand, a toy manufacturer may not testincoming chips at all, cutting costs by relyingon returns and complaints from the field to lo-cate problems. Such an approach is favored

"Eleven percent of nearly 20,000 ICs tested for the 1977 Pio-neer mission to Venus were rejected, many of these tests involv-ing burn-in periods of 100 to 200 hours. The very high rejectrate reflects the severity of the application. See "Quality and Re-liability of Semiconductors and CTVs: United States v. Japan,"op. cit., p 14.

where other parts, are less likely to fail than theICs.

Semiconductor products are normallyscreened and purchased to an acceptable quali-ty level (AQL), a procedure much less expen-sive than 100 percent testing. From the stand- ,

point of the purchaser, the AQL is the permissi-ble fraction of delivered parts that can be de-fectivei.e., that escape detection during in-spection and screening. A 1 percent AQLmeans that no more than one defective circuitout of every 100 is allowed, on the average, inan acceptable lot; statistical sampling methodsare tailored to this requirement.

Figure 40 outlines the testing programadopted by a manufacturer of point-of-sale ter-minals for purchased ICs. Tests are conductedat many points prior to shipment becausedownstream failures cost much more to findand fix. Costs are even higher for field fail-uresboth the direct expenses of warrantyrepairs and the possible.costs in terms of dam-age to the firm's reputation. Table 48 illustratesthe growth in costs of locating and repairingfaulty components at successively later stages.The indirect and intangible costs can be muchgreater than the direct expenses.

Testing and Screening in Japan

When first qualifying a new vendor, Japanese"-purchasers normally test all incoming parts.With satisfactory experience, statistical sam-pling replaces 100 percent testing. If the defectfraction remains below 0.01 percent (100 de-fects per million parts) and downstream fail-ures are rare, the purchaser may stop screen-ing. Even when purchaser, and supplier areunrelated firms, customers prefer to depend ontheir suppliers to guarantee quality levels. Jap-anese manufacturers do tend to rely rather ex-tensively on in-process testing, aging, and burn-inin part to minimize infant mortality fail-ures.

Such practices differ from the arms-lengthrelationships common in the United States.Perhaps because the major Japanese manufac-turers of semiconductors are also the majorusers, they often appear to take the attitude thatthe objective of quality control is to deliver

2-3

On. 6Manufacturing: Quality, Reliability, and Automation 229.

Figure 40.Testing Sequence forPoint-ofSale Terminals

polit;iioldprivIsuiC(after Ootitttriiigd".-

'!-a890041Y).;!.......

equV- rent

RATEAutomatic Test Equipment.

100% orselective,dependingon the part

Selective

100% inspection

1000/..

100% - ATE orspecial test -depending oncost trade-off

100%

100%

130%

100%

100%

SOURCE: Adapted from R. Fleishman, R. J. Lever, and R. N. Parente. "TotalTesting," Circuits Manufacturing, November 1979, p. 32.

parts that meet their own in-house standards.A more common attitude in the United Stateshas been that parts need only meet the cus-tomer's requirements; customers that demandhigh quality may get it, others receive lessattention.

Table 48.Typical Costs of Detecting and RepairingFaulty Components in an Electronic System

Point of detection Direct cost Intangible costDevice level Cost of device, if Minimized If more

not refunded by devices than neededmanufacturer. are purchased

Initially.Circuit board level Manufacturing process

dislocated.System level $50 Shipment may be

delayed, disrupted.In the field $500 Customer upset.SOURCE: "Calls Volume Key to Testing Decision," Electronic News, Feb. 18,

1980, supplement p. 20.

Quality and Reliability ComparisonsAlthough respective quality levels of ICs

made in the United States and Japan have beendebated for several/years, there is little concretedata bearing directly on this matter. For a validcomparison, circuits from U.S. and Japanesefirms should be tested under the following con-ditions:

1. The devices should be the same type andof similar designe.g., 4K dynamic RAMs,8080 microprocessors.

2. Test procedures should be identical, thetests conducted at about the same time. (Itis not possible to compare quality or re-liability at the present moment. Qualitycomparisons always refer to some pointin the past. And, while the most recent re-sults are always desirable, quality and re-liability are dynamic characteristics; theyfluctuate with the vagaries of the manufac-turing process.)

3. The ICs should be produced to the samepurChase specifications in terms of AQLor other quality requirements, ideally fordelivery to the same customer.

While it is no surprise that little of this kindof data has been made public, the unfortunateconsequence was a series of public relationploys obscuring the technical questions: Werethere real differences in quality? If so; whatwere the reasons?

By any measure, semiconductor quality andreliability have improved immensely over theyears, regardless of whether the devices havebeen produced in the United States, Japan, orEurope. As an example; figure 41 shows de-

2 3 5


"..=, 1,000

100

10

Figure 41.Reliability Trend forAnalog Integrated Circuits

1974 19751976 1977

1978

1979

1980

Cumulative units produced

SOUNCL. G. Peattie. "Quality Control tor. ICs." IEEE Spectrum, October 1981,p 93

creases in failure rates for analog (linear) ICsas used in consumer electronic products. Othertypes of ICs show similar trends. Nonetheless,sources in the American electronics industryboth manufacturers and purchasers of semi-conductorsagree that, during the mid to late1970's, quality levels delivered by Japanesefirms were superior to those delivered by U.S.firms. There is also broad agreement that quali-ty levels delivered by American semiconduc-tor firms have greatly improved since the pub-licity given the Japanese "quality advantage"during 1980.24 The available data is summar-ized in appendix 6A. But at the same time thatU.S. semiconductor firms have made rapidstrides, Japanese manufacturers have alsoproved. While the gap has certainly narrowed,Japanese firms on the average may remainahead in quality.

It is also important to recall that discussionsand data on IC quality have centered on prod-ucts sold in the merchant market. No data have

24Much of this publicity stemmed from a seminar entitled"Quality Control: Japan's Key to High Productivity," organizedby the Electronic Industries Association of Japan and held inWashington, D.C., on Mar. 25. 1980. Data first released at thatseminar appear in appendix table 6A-1, pt. A. A perspective com-mon i,i much of the American merchant semiconductor induStry

..at that time can be found in T. D. Hinkelman, "The Economicsof Quality: U.S vs. Japan." An American Response to the Foreignindustrial Challenge 1' /I High Technology Industries, Proceedingsof the Semiconductor Industry Association Government PolicyConference, Monterey, Calif.. June 18-19, 1980, M. Hodgson (ed.)(Palo Alto. Calif.: Worden Fraser Publisher, 1980), p. 85.

been made public on quality levels attained bycaptive producers such as Western Electric orIBM. Captives account for about 40 percent ofall ICs made in the United States (ch. 4); thequality and reliability attained by captive pro-ducers would, if available, be a useful indicatorof the relative technological capability of theAmerican industry.

The ranges in quality level included in ap-pendix 6A, particularly table 6A-2, show a re-markcble lack of consistency on the part of allvendors. Even the top 16K RAM suppliers ex-hibited a factor of five difference between bestand worst lots. Much larger spreads wc. re therule, particularly for the American firms: Thisillustrates the danger in generalizing from lim-ited samples of IC quality data. It also indicatesthe importance of a consistent and well-con-trolled manufacturing process, and the diffi:culty of maintaining that control.

Spokesmen for the U.S. semiconductor in-dustry have sometimes claimed that Japanesefirms create a false image of higher quality bysorting ICs and sending only the best to impor-tant customers like Hewlett-Packarda prac-tice that has been termed "quality targeting"or "quality dumping." The claim is furthermade that this is a high-cost strategy, intendedto "buy" U.S. market shareand that after theirAmerican competitors have been forced out,the Japanese will raise their prices and shiptheir normal product, which will be found tobe poorer in quality.25 Indeed, manufacturersin many industries and in many countriessometimes attempt such strategies. Americansemiconductor firms will sort ICs and shiphigher quality lots to purchasers who demandthem. However, as a widespread and generalapproach to marketing in the United States,quality dumping by, the Japanese seems im-plausible. In order to ship higher quality lotsto the United States, they would have to shiplower quality products to other customersineither export or domestic marketsthus run-ning the risk of jeopardizing those markets. Itis difficult to believe that Japanese IC manufac-

2,T. D. Hinkelman, op. cit. See also "The Quality Goes OnBefore the (Japanese) Name Goes On," Rosen Electronics Let-ter Mar. 31, 1980, p. 1.

Ch. 6Manufacturing: Quality, Reliability, and Automation 231"

turers would do so in any concerted way, par-ticularly at home.

It is clear from the data in appendix 6A that,at least until the recent past, Japanese large-scale ICs have had, on the average, both bet-ter quality and better reliability than compar-able American parts. This does not mean thatsome products from some U.S. companies werenot as good as or better than products fromJapan. As the tables in appendix 6A indicate,the range in quality and reliability delivered byany firm is likely to be wide; this is intrinsicto the technology of semiconductor manufac-turing. But as a generalization, the UnitedStates had fallen behind in both quality and re-liability. It is also clear that the performanceof American firms on these dimensions hasgreatly improvedin part because of the com-petitive pressures generated by the publicitygiven this issue. According to recent reports,the quality levels of 16K RAMs supplied by anumber of American firms are now, on theaverage, about the same as those of Japanese'devices.28

While this is a positive sign for the future,it does appear that Japanese firms devote moreresources to analyzing field failures so as tofind and eliminate their causes. In Japan, elec-tronics firms have normally maintai ,Jd cap-tive service organizations which gather andanalyze field service results, and feed themback to design and manufacturing depart-ments. One American purchaser of Japanesesemiconductor. devices was reportedly quitesurprised to find a team of engineers dis-patched to explore the reasons for a batch ofcircuits with a defect rate of only 0.25 per-cent.27

In the future, if American managers devoteas much attentionand as many resourcesto the quality and reliability of their productsas do the managers of Japanese firms, there is

"E. R. Hnatek, "Semiconductor Memory Upda, ., DRAMs,"Computer Design, January 1982, p. 109; "Faults Show Up inJapanese RAMs," Electmnics, Jan. 13, 1982, p. 33; "In Semicon-ductors, Perfection Is the Goal," Business Week Nov. 1, 1982,p. 72.

""Quality and Reliability of Semiconductors and CTVs: U.S.v. Japan," op. cit., p. 57.

no reason why U.S. firms should not keep pacewith, or surpass, their overseas rivals on thesedimensions of IC technology.

Quality and Reliability of Color TVs

That Japanese TV manufacturers haveachieved excellent quality and reliability, andlargely succeeded in their marketing strategies,is self-evident. In order to bypass the fran-chised dealer networks that American manu-facturers relied on, they had to forgo extensiveservice organizations. Failure by Japanese im-porters to maintain both the image and the re-ality of a reliable, trouble-free product wouldrisk the largest market in the World. Most sur-veys continue to show the reliability of TVsproduced by Japanese firms to be better,though differences in quality appear small.

Many of the TVs sold in the United Statesby Japanese firmsare now assembled here.Quality levels achieved in the U.S. operationsof both Sony and Quasarthe firm that Mat-sushita bought from Motorola in 1974havereceived a good deal of publicity.28 Such plantstend to combine features typical of Japaneseand American manufacturing operations; seechapter 8 for a discussion of managementstyles and their effects. At least as important,TVs assembled in the United States by Japanese-owned firms contain large proportions ofimported components. Based on the findingsfor semiconductor devices outlined in the pre-vious section, imported components might beexpected to exhibit somewhat higher levels ofquality -and/or reliability than similar partsfrom American suppliers.

MoSt of the information bearing on qualityand reliability for TVs comes from sources likeConsumer Reports. Several years ago this pub-

280n Sony, see "Statement of Sadami 'Chris' Wada, AssistantVice President, Sony Corp. of America," Quality of Productionand Improvement in the Workplace, hearing, Subcommittee onTrade, Committee on Ways and Means, House of Represents=fives, Oct. 14, 1980, p. 62.

At Quasar, quality levels improved rapidly after the-Matsushita .

purchase; however, the baseline is deceptive idthat Motoroladevoted few resources to its TV operations for a nbrither of yearsprior to the sale. This case is discussed in more detail in the ap-pendix to ch. 8.

237,


lication surveyed nearly 200,000 owners of19-inch color TVs, the most popular size, soldduring the period 1975-79. Nine of the fifteenbrands for which the origins are knownallthe Japanese makes but no otherswftre givenreliability ratings of "better than average"based on the average cost of repairs during the1979 calendar year. The remaining six brandnames-/--for practical purposes, all the Amer-ican brandswere rated "average" (one brand)or "worse than average" (five brands).29 TheConsumer Reports survey reflects reliabilitiesof sets sold" during the period 1975-79 only.However, TV designs do not change rapidly;these trends should still be a reasonable indica-tion of comparative reliability levels.

Similar but not identical reliability rankingscome from a survey conducted by Trendex inthe same year, 1979, but again, covering TVsmanufactured over a period of years.30 Of the12 brands included in this survey, TVs madeby Japanese-owned firms filled four of the topfive places in terms of reliability. The reniain-ing Japanese brand ranked seventh, with thebottom five positions filled by American firmsplus Magnavox.

Table 49 presents data from a survey of TV. repair shops that ,:oint in a direction rather dif-ferent from the L ..)nsumer surveys discussedabove. ""7'"; ;able covers a smaller number ofbrands. .13e American (Zenith, RCA, and Syl-vaniathe latter at that time U.S.-owned,though since purchased by Philips); three Japa-nese (Sony, Quasar, and Panasonicthe lattertwo are Matsushita brand names); plus iviag--yriavox. The repair shops rated the Americanbrands generally superior on all three cri-

"19-Inch Color TVs," Consumer Reports, January 1981. p.34. The nine brands in the "better than average" reliatilitycategory included TVs sold by Sears, most of which are madeby Sanyosome imported, some assembled in the United States.Other private brand merchandiserse.g., Montgoniery Ward,J. C. Penney tend to purchase from both American and foreignsuppliers. Exchuding both the Wards and Penney TVs becauseof th,- Airlortain origins; 15'brands remain. Of the 16, 5 are

,; are Japanese, and the other Magnavox is ownedby Philips. As Magnavox is much more nearly independent ofits parent than the American subsidiaries of Japanese firrnS. ithas been considered a U.S. brand for purposes of.this compar-ison.

""Quality and Reliability of Semiconductors and CTVs: U.S.v. Japan," op. cit., p. 78.

Table 49.Rankings by Repair Shops of TV Receiversfor Quality and Reliability

Rankings in terms of picture quality andother performance features:

1. Zenith2. RCA3. Sony4. Sylvania5. Quasar6. Magnavox7. Panasonic

Ranking", in terms of reliability:1. Zenith2. S,)ny3. F CA4. Quasar5. Sylvania6. Panasonic7. Magnavox

Rankings in terms of increasing costs of repair:1. Zenith2. RCA3. Sylvania4. Quasar5. Magnavox6. Sony

SOURCE: "Quality and Reliability of Semiconductors and CIVs:11.S, v. Japan,"report No. C972, prepared for OTA by Consultant Services Institute,Inc., under contract No. 033.1170.0, p. 79. The survey, conducted dur.Mg 1980, covered 60 repalr,shops In Chicago, Boston, and NorthernNew Jersey.

teriaperformance, reliability, and costs ofrepair. In particular, the largest-selling U.S.TVsthose made by Zenith and RCA show upvery well, with Zenith top-ranked in, eachcategory. In contrast, Zenith and RCA are rated"worse than average" in reliability by Consum-er Reports. Because the Consumer Reports sur-vey covered such large sample sizesmorethan 40,000 owners of 19-inch Zenith sets, and35,000 made by RCAit must be given considerable weight. However, the data in table 49are not restricted to any particular screen size,and might be more representative of each man-ufacturer's overall product line.

As is true for ICs, American manufacturersof TVs have clearly made considerable stridesin improving quality and reliability-7-spurredby competition among themselves as well aswith the Japanese. Consumer electronics firmsnow screen and burn-in components morethoroughly; they also burn-in complete

outboards and assembled sets to weed out earlyfailures. Automation has helped quality. Final-ly, American TV makers are using larger num-

__

Ch. 6Manufacturing: Quality, Reliability, and Automatio.' 233

hers of imported componentsmostly fromJapan and other Asian countries. Importedcomponents often cost less, but in at least somecases they have been chosen because of su-perior quality and/or reliability. Even picturetubeswhich are bulkier and more difficult toship than other componentsare being im-ported in increasing numbers; one ,U.S. man-ufacturer stated that Japanese picture tubes hadone-third the in-process failure rate of Ameri-can-made tubes.31

"Ibid, p. 80. Japanese-owned firms that assemble and sell TVsin the United States still import many components, but are grad -uallyjncreasing value added here. Mitsubishiwhich producessets in the United States for sale under the MGA brand nameimports about 30 percent of their parts from a subsidiary inSingapore, and another 15 percent from Japan. Sony continuesto bring in from Japan about 35 percent of the parts, for theirAmerican-made sets. In general, the more critical componentsand subassemblies from a performahpe and quality standpointare importede.g., circuit boards. Cabinets and nonelectronicparts are the first to be purchased domestically. See Quality ofProduction and Improvement in the Workplace, op. cit., p. 85.

Consumer perceptions created by and re-flected in surveys like those discussed abovecan be extrapolated with some confidence intoat least the near future. Furthermore, becauseTVs have a design life of 7 to 10 years, thesurveys discussed above should do a good jobof predicting the reliability of sets presently inuse. The weight of the evidence points towardan advantage in reliability for Japanese TVmanufacturers during the 1970's. Even ifAmerican manufacturers today are producingTVs as reliable as their Japanese competitors,the image of reliability that the Japanese havegained will persist for a number of years tocome. On the other hand, differences in quali-ty among TVs appear to be sma11.32

"For example, "Small-Screen Color TV's," Consumer Reports,January 1982, p. 17, where the distribution of brand ratings byset performance and quality shows no systematic differencesamong U.S. and foreign brands.

AutomationManagers make decisions involving the auto-

mation of production processes largely on thebasis of costs. Automation typically involvestradeoffs between labor cost and capital costthat depend on production volumes; mecha-niZed production facilities also tend to lackflexibility,. which raises the costs of adaptingthem to new product designs. Factors less di-rectly related to costs include the impacts ofautomation on quality, and the possibility ofmechanizing unusually dangerous, dirty, oronerous jobs.

Modern automated production systems usu-ally rely on electronics, although electrome-chanical control systems were common untilrecently. Examples of automated processinginclude:33

automatic machine tools, ranging fromlathes and milling machines controlled by

"See, in general, M.P. Groover, Automation, Production Sys-tems, and Computer-Aided Manufacturing (Englewood Cliffs.N.J.: Prentice-Hall, 1980).

99-111 0 - 83'- 16

mechanical cams, to those that operateunder computer control, to machiningcenters;automated gaging, inspection, and testing;examples include inspecting circuit boardsby means of video image processing tocheck for solder runs or other sible de-fects, measuring the dimensions of ma-chined parts, and determining the chemi-cal composition of steel;mechanized systems for materials han-dling, ranging from computer-controlledconveyors to fully automated warelmses;process control systems incorporating sen-,sors and processors that implement con-trol algorithms based on feedback, feedfor-ward, or some combination (see ch. 3, app.3C, on industrial process control);use of computers in management or sup-port functions such as scheduling of jobflows, inventory control, or statistical qual-ity control; andcomputer-aided design methods to aid ingeometric modeling, in engineering anal-

233

234 Intern:Itional Competitiveness in Electronics

ysis, or in preparing design drawings orequivalent design information coded forautomated production processes.

The earliest numerically controlled (NC) ma-chine tools operated from instructions on apaper tape or similar storage medium, analo-gous to the cams and other electromechanicalcontrols used for many years to automate man-ufacturing. The NC tape, however, could beprepared with the aid of a computer, and easi-ly duplicated or modified. In the next stage,rather than following 'a sequence of instruc-tions held in ayead-only memory such as apaper tape, direct numerically controlled(DNC) and computer numerically controlled(CNC) machines were developed. These re-spond in real time to commands from the proc-essor of a computer. As a result, control algo-rithms based on gaging or sensing of machin-ing parameters can, at least in principle, be im-plemented. In a DNC system, one computercontrols several machines; CNC machinesoperate under the control of a dedicated proc-essor, typically a small minicomputer or a mi-crocomputer.

Sophisticated control systems use informa-tion from sensors for regulating the process,typically by adjustments that keep/measuredparameters within predetermined bounds. Fora machining operation, dimenions can bemeasured; for a. wafer fabrication line in asemiconductor plant, possible control param-eters include temperatures, pH of reagents, andcurrent flows during ion bombardment. In con-trast to such "closed loop" systems, in which,information flows from the proCess back to thecontroller, systems in which there is no sens-ing and transmission of information, but whichoperate purely on preprogramed instructions,are called "open loop." A skilled machinistcloses the loop just as does an automatic con-trol system on a CNC lathe equipped for auto-matic gaging. But in fact, most NC machinesstill run on an open loop basis.

Electronic control systems make possible theautomation of many processes that in earlieryears were too difficult or too expensive to

mechanize.34 In essence, the flexibility gainedthrough electronic controls makes automationcost effective in applications where productionvolumes are low. In the past, automation waspractical only in continuous process industriessuch as food preparation and packaging, or inhigh-volume batch production industries likeautomobile manufacture. In the automobile in-dustry, simple assembly operations, as well asmachining, have been carried out by transferlines linking a series of machines for manyyears; human operators have worked along the4line performing tasks that were difficult or cost-ly ly to mechanize.

Fixed and Flexible Automation

Automated production in either continuousprocess or batch industries can be thought ofas spanning a range from "fixed" or "hard".automation to "flexible" or "programmable ".automation. Fixed automation is exemplifiedby an automatic lathe in which the "instruc-tions" are encoded in the profiles of cams. Toset up the lathe for a different job, the camsmust be changed. Designing and machining anew set of cams .is a time-consuming job per-formed by skilled craftsmen. Conventionaltransfer lines are examples of fixed automationapplied to a sequence of tasks. When an auto-mobile manufacturer designs a new engine ortransmission, the entire transfer line mighthave to be scrapped and replaced. Much thesame is true if an electronics firm using suchequipment wishes to introduce a new designfora printed circuit board, TV chassis, or com-puter termi- J.

Until recenti , automated production equip-ment with the flexibility to accommodate sub-stantial variation in the design of the productwas the exception rather than the rule.35 Ma-chines seldom adapt very well to perturbations

34j. A. Alic. "Government Attitudes Toward ProgrammableAutomation," Proceedings of the Twenty-third International Ma.chine Tool Design and Research Conference, B. J. Davies (ed.)(London: Macmillan. 1983), p. 521.

3,Flexibility in the context of manufacturing systems carriesa number of possible meanings; see, for example, D. Gerwin."Do's and Don'ts of Computerized Manufacturing." HarvardBusiness Review. March-April 1982. p. 107.


in the processe.g., a part that comes downa conveyor sidewaysmuch less to new prod-uct designs. When flexibility has been needed,manufacturing operations have depended onpeople. Engine lathes, which are operated en-tirely under manual control,are the flexiblecounterpart of the automatic lathe. A skilledmachinist can make an amazing variety of dif-ferent parts on an engine lathe, but the cost perpart will be high.

One reason for the lack of flexibility charac-teristic of fixed automation is that new hard-warefixtures, toolingis needed to accom:modate a new design. A second reason is thatthe controls must be reprogramed. A hard-wired electronic control systemWhether analog or digitalrequires new circuitry everytime the control logic is altered. This is costlyand time-consuming, just as for an automaticlathe that requires a new set of cams. In recentyears, computer control has became cost effec-tive for replacing many mechanical or electro-mechanical control systems.

While the performance of a computer-basedprogrammable controller as a control sys-temwill generally bq superior to the alter-natives, this is not neceSsarily the.case for hard-ware. Often, flexibility l'n hardware trades offagainst performance, a d perhaps capital costas well. For example, a r bot can be programedto weld together section of pipe, but might notbe as fast as,ta specially designed automaticwelding machinewhich might also produce -:better quality welds. However, the robot couldbe programed to do other. tasks. In general,then, a fleXible facility may be less efficient formaking any one prods 1 than a dedicated,hard-automated manufacturing system.38

"Recent R&D work at Westinghouse illustrates a typical appli-cation of flexible manufacturing Ihere assembly, one of the mostchallenging tasks for automatic) . Westinghouse makes morethan 450 different models of sma !electric motors. with :..n aver-age lot size of 600: model changbs average 13 per day. In suchcases, labor-intensive manufacturing "Methods have generallybeen preferred. Fixed automation using transfer lines has beena real option only for long prodtiction runs of similar or identi-cal products. For a description f the flexible assembly systemunder development, see R. G. Atraham. "APAS: Adaptable Pro-grammable Assembly System."' Computer Vision and Sensor-Based Robots. G. G. Dodd and L. Rosso! (eds.) (New York: PlenumPress. 1979), p. 117.

As flexible automation technologies incorpo-rating computer-based control systems im-prove, an enormous pool of potential applica-tions will open; the consequences will include,not only cost reductions and productivity im-provements, but shifts in the composition ofthe factory work force. Skill mixes needed inmanufacturing industries will change, and thetotal number of employment opportunities inthe manufacturing sector of the U.S. economymay shrink even as total output increases. (Em-ployment levels and work force compositionare discussed in chapter 9.)

Automation in Electronics Manufacture.

Reasons for AutomatingMost applications of automation by U.S. elec -.

tronics firms have been driven by costs; non-cost factors have perhaps weighed more heavi-ly for Japanese manufacturers. The industry inJapan has at times faced labor shortages; in ad:dition, Japanese firms may sometimes havebeen motivated by potential quality improve-ments to automate earlier than their Americancounterparts.37

Table 50 presents the results of a 1979 surveyin which. Japanese electronics manufacturerswere asked to list reasons for their decisionsto automate. The most common response wasto reduce costs, with quality improvements sec-ond; in contrast, a 1975 survey found laborshortages ranked at the top. Comparison of1975 and 19'9 results shows a rapid increasein automation by Japanese electronics manu-facturers,38

Another example of flexible automation in assemblythis onealready in useis a machine developed in Japan by Nippondenv,that can put together 288 different versions of an automobiledashboard indicator. The average lot size is 40, with 200 change-overs per day. See "British Government Finances Robotics De-velopment," West Europe Report: Science and Technology, No.70, Joint Publications Research Service JPRS 78820, Aug. 25,1981, p. 14.

"K. Ito, E. Taira, S. Yagi, K. lwamoto.and K. Tsukamoto, "TheProgress of Automation and the Improvement of Reliability inProduction of Color TV Receivers," IEEE Transactions on Man-.ufacturing Technology, vol. MFT-3, December 1974, p. 55.

" "Quality, and Reliability of Semiconductors and CTVs: UnitedStates v. Japan," op. cit., p. 47. -Japanese consumer electronicsfirms reportedly spend about a third of their R&D dollars on man-

241


Table 50.Reast,is Given by JapaneseElectronics Fly ms for Automating

Percent of firmssurveyed' Reasons for automating

84% Reduction in manufacturing cost69 Improvement in quality43 Increase in production volume43 Improvement in workplace:.

conditions32 Labor shortage

°Multiple responses were common.SOURCE: Nikkan Kogyo ShIrribun, May 1, 1979, and July 11, 1979, cited In "Quality

and Reliability of Semiconductors and CTVs: United States v. Japan,'report No. C972. prepared for OTA by Consultant Services Institute,Inc.. under contract No. 033.1170.0, p. 50.

Consumer ElectronicsManual assembly was at one time the rule

for electronic products ranging from radiosand TVs to computers. Components were firstinserted and/or soldered into circuit boards-, theboards then installed in a chassis, the assemblyfinally tested and adjusted. Component inser-tion was one of the first tasks to be mechanized.This is relatively. easy for discrete componentswith axial leads, more difficult for ICs. Con-sumer electronics manufacturers first movedto automatic insertion of discrete parts; as ICswere designed into TVs, they were at first stillinserted manually. By the end of the 1970'smuch of this work had been automated as well,using' equipment roughly similar to that pic-tured in figure 42..

41'he spread of automation in the U.S. con-sumer electronics industry has been incremen-tal. Firms automated at different times and fordifferent reasons, depending in part on strate-gic responses to foreign competition. In mostcases, the initial reaction, based on Japaneseadvantages in labor costs, centered on movinglabor-intensive operations offshore rather thanautomating.

When American TV manufacturers did re-spond to competitive pressures by automating,cost was the driving force, quality and reliabili -.ty improvements secondary. Meanwhile, the

ufacturing developments: see Transfer of Technology in the Con-sumer Electronic Industry, Sectoral Study No. 2 (Paris: Organi-zation for Economic Cooperation and Development. Sept. 14,1979), p. 41. This percentage is probably a good deal higher thanin the United States or Europe.

Japanese continued to take the initiative inautomation, even though their labor costs re-mained lower. By 1976-77, 50 to 80 percent ofall component insertion in Japanese TV fac-tories had been mechanized." Computer-con-trolled testing and inspection of components,subsystems such as circuit boards, and com-plete TVs also spread rapidly. Concurrently,chassis were redesigned to take advantage ofthe characteristics of automated productionequipment. According to one study, labor pro-ductivity in the assembly of color TV receiverswas a little greater in the United States thanin Japan in 1970, but by 1977 productivity inJapan was more than twice that here--figure43.40

Semiconductors

In the earlier years of the semiconductor in-dustry, virtually all production operationsfabrication, assembly, inspertior, and testingwere labor-intensive. Among LI firms, thespread of automation may have Wen retardedby the widely publicized difficulties of PhilcoCorp., which invested ho''' in automatedmanufacturin, duri7 g th AP; 1950's only tosee rapid ulidages ill trabstor technologyrender its equipment obsolt;te.41 Philco laterdropped out of the semiconductor. business.

At present, semiconductor firms in all partsof the world -are automating rapidly, not onlyin manufacturing but in computer-aided circuitdesign. A few companiesboth American andforeignhave installed fully mechanized pro-duction lines, from wafer fabrication throughinspection, testing, wire bonding, and packag-ing. The benefits include increased yields as

"For a description of some of the technology used by Mitsu-bishi. see T. R. Crossley. "Study Tour of Industrial Robots inJapan," European Office of Aerospace Research and Develop-ment report No. EOARD-TR-80-3. London, August 1979, pp. 32-33. At the time of this visit. Mitsubishi was using robots of theirown design for assembling printed circuit boards for TVs. Theassembly line could be changed over for a different board con-figuration in 2 hours.

01. C. Magaziner and T. M. Hout, Japanese Industrial Policy`(London: Policy Studies Institute. 1980), p. 22. The data comesfrom work performed by the Boston Consulting Group.

41J. E. Tilton. International Diffusion of Technology: The Caseof Semiconductors (Washington. D.C.: The Brookings Institu--tion. 1971). p. 83.

24,ti-

Figure 42.

a result of better proexample, claims thatrication facility has15 percent, and prodscent; the control sy:mainframe compute]ers and 74 microcor

.

Rates of Automaticthe United States a

While much of theward automation in

42"Toshiba Completes FullJapan Report, Joint PublicatiiJan. 19, 1982. p. 85.


Figure 43.Average Labor Hours for Assembly of21-Inch Color TV Receivers (1977)

6

5

3

2

Japan United West UnitedStates Germany Kingdom

SOURCE: I. C. Magazlner and T. M. Hout, Japanese Industrial Policy (London:Policy Studies Institute, 1980), p. 23. Data are from a client study per.formed by the Boston Consulting Group.

electronics companies do seem to have adoptedrobots more quickly than their American coun-terparts. Two to three! times more robots areat work in Japanese factories than in the UnitedStates; about 40 percert of these have been in-stalled in the factories of Japanese electronicsand electrical equipment firms." But again, anumber of the larger American electronic com-panies are well known; both for their R&D inrobotics technology and for their applicationsof robots in manufacturing.

Some American electronics firms may havehad difficulty in finding the capital needed forautomation. The replacement of labor-intensiveproduction operations by capital-intensiveequipment aggravates the problems of financ-ing expansion (ch. 7); a transfer line for insert:ing components in a circuit board can easilycost halfa dollars. In contrast, to theirsmaller' American competitors, Japan's in-tegrated electronics manufacturers can take ad-vantage.of internally generated fundsas wellas somewhat lower costs for external capital=to invest in mechanized equipment. Further-more, the Japanese GOvernment has reported-ly given preferential tax treatment for in-vestments in production equipment that willimprove the quality and reliability of Japanese

44R. Ristelhueber, "RobotiicsThe Applications Gap," Elec-tronic News. Jan. 11, 1982, p. 60.

goods, particularly those for export." Such ac-tions have probably affected the timing of in-vestments more than the eventual extent ofautomation.

Although a few American semiconductorfirms make some of their own manufacturingequipmentnotably the larger, more highly in-tegrated companiesmost such equipment isdesigned and built by independent suppliers.In Japan, it is more common for electronicsfirms to design and fabricate their own. Mat-sushita, for example, meets 30 to 40 percentof its equipment needs internally.4e Internalcapability for equipment development can helpspeed automation.

As integration levels for ICs continue to in-crease. automation will become a necessity.Fine line widths and other requirements forvery large-scale integration (VLSI) demandlevels of cleanliness that are much easier toachieve if human intervention is minimized.More complex circuits can only be designedwith computer-aided techniques, together withcomputer-aided drafting and mask generation;once built, such circuits can only be tested withcomputerized equipment. Better process con-trol modelsnow limited by gaps in funda-mental understanding of the physics of elec-tron deviceswill be needed to ensure the qUal-ity and reliability of VLSI circuits, Continuedautomation may reduce pressures for offshore

'6"Quality and Reliability of Semiconductors'and CTVs: UnitedStates v. Japan," op. cit., p. 25.

461J. C. Lehner., "Japan Strives To Move From Fine Imitationsto Its Own Inventions," Wall Street Journal, Dec. 1. 1981. p. 1.Japanese firms continue to purchase a good deal of automatedproduction equipment from American suppliers; as pointed outin ch. 4, about half the equipment used by Japanese semiconduc-tor manufacturers comes from the United Stites. This percent-age has been declining, however, with some observers predictingthat Japan will produce 70 percent of its needs by 1985. Japanesefirms are reportedly already designing and building fifth-genera-tion automated wire bonders, while U.S. firms are still work-ing with first or second generation machines; see "Pushing forLeadership in the World Market," Business Week Dec. 14, 1981,p. 61. In some casese.g., automated testing equipmentjapa-nese 'products have the reputation of being somewhat more reli-able than those of American suppliers, largely because they aresimpler. However, industry opinion generally holds that U.S.equipment is as good as or better than that made in Japan orin Europe, as well as being less expensive. See "Can Semicon-ductors Survive Big Business?" Business Week, Dec. 3, 1979.p. 81.

24.1


manufacturing because labor costs will becomea smaller fraction of total manufacturing cost.

Robotics

Industrial robots comprise a subset of pro-grammable automation technologies mimick-ing some of the attributes of the human workforce. They are more flexiblein terms of abili-ty to perform new tasks, or to carry out com-plex motion sequencesthan other types ofprogrammable equipment. Because advancesin robotics depend to considerable extent onelectronics and computer science they are dis-cussed in some detail below. In the future,robots will be used to automate many of theoperations in making electronics products nowcarried out by hand. In Japan, robots arealready being ined to produce more robots bya subsidiary of one of the major electronicscompaniesFanuc, a part of the Fujitsu organi-zation.'"

The changing proportion of costs associatedwith electrical and mechanical componentssince the first industrial robots were intro-duced in the 1960's demonstrates the impor-tance of electronics for this technology. A fewyears ago, about half the direct cost of makinga robot was associated with the electronics, theother half with mechanical components. Nowonly about 15 percent of the costs go into elec-tronics, largely because of the increases in com -.puting power/dvailable with cheap ICs. Costsfor the mechanical components have notchanged greatly, but the total 'osts of robotshave decreased, the mechanical parts now ac-counting for 85 percent of the total.

TechnologyIndustrial robots, those used for factory

work, are machines that can move a manipu-lator, or end effector, at the end of a chain ofmechanical links. The end effector maybe agripping device similar to fingers; alternative-ly, the end of the robot arm may carry a tool,welding torch, or nozzle for spraying paint.

47N. Usui, "Untended Machines Build Machines," AmericanMachinist. June 1981, p. 142. (b) This robot moves rectilinearly

The simplest. robots have only two or threedegrees of freedom; an illustration of a twosdegree-of-freedom system would be ah "iifili"that could only extend and rotate, as for tight-ening a bolt. The most sophisticated robotshave seven or eight degrees of freedom, whichallows them to reach around obstacles. Figure44 shows a pair of typical robot designs.

While robots trace their desumt from moreprimitiv6 manipulators having little flexibil-itye.g., with position and sequencing con-

Figure 44.Two Approaches to,the Design'of Industrial Robots

I I I I 1.3is

A..1

Photo credit: Cincinnati Milacron

(a) This robot emulates the human arm

"V '

Photo credit: Westinghouse


trolled by limit switchesstate-of-the-artdesigns now operate under computer control,often a microcomputer. In routine productionapplications, the robot is commonly "taught"a sequence of motions by a human operator,who leads the arm through these motions whilethey are stored in memory. The machine canthen repeat them on command. Although sat-isfactory for simple applications like spraypainting or some forms of welding, off-line pro-gramingin which the instruction sequence isdeveloped independently and loaded into thecomputer when neededis a major R&D goal.

Virtually all robots still operate with relative-ly primitive open loop control systems.48 Thisis one of the factors limiting operating speeds,as well as the accuracy with which the end ef-fector can be positioned. Current-generationrobots are also burdened by complex and ex-pensive actuators that tend to restrict perform-ance. At some future time, one robot may beable to throw a part across the factory floor toanother, but this i3 far in the future. Makingrobots "smart"i.e., with the ability to gatherand process substantial amounts of informa-tion, then make decistons based on that infor-mationis a related problem. Few robots canyet make even simple decisions.

As such examples indicate, robotics technol-ogy depends on computer technology. Whilecomputer firms like IBM, Digital EquipmentCorp., and Texas Instruments are expected toenter the market for robotsand Fujitsu andHitachi are already two of the leading pro-ducers in Japanmany of the robots in currentproduction come from machine tool builders.In the United States, for instance, CincinnatiMilacron has a substantial share of the market.

From the perspective of the machine tool in-dustrywhether the portion that builds metal-cutting equipment or the manufacturers ofhard-automated assembly equipment such astransfer linesrobots trace their descent most-

. ly from NC machines. In fact, much of the tech-nology in the control systems of current-gen-

""Government Attitudes Toward Programmable Automation."op. cit. Much of the material that follows is drawn from thispaper.

eration robots is similar to that for DNC andCNC machine tools. Companies that intend tocompete in the design and manufacture of fu-ture generations of robots will need a broaderrange of technical capability; those moving intcprogrammable automation from other high-technology fields may have the advantage, par-ticularly firms with experience in automaticcontrols and the modeling of complex mechan-ical systems.

Robots in Manufacturing

Robots are usually installed in factorieswhere they can cut costs (compared to humanworkers) and increase labor productivitythesame motives that drive other capital, invest-ments. In many early installations of robots,human workers wore replaced on a one-for-onebasis, a substitution facilitated-by the robot'sability to emulate the human arm. In compar-ing the costs of robots to those for human work-ers, one-for-one replacement has often been as-sumed. This is potentially misleading becausea more thorough redesign of the production fa-cility means that some robots may each replaceseveral people, while in other cases severalrobots might be needed to do the work of oneperson. A cost analysis comparing robots andpeople must also take account of the numberof shifts planned for the facility, maintenanceand repair costs, depreciation. and energy con-sumption. A robot may use a hundred timesas much energy, as a human worker. General-ly speaking, when production volumes aresmall, human operators will still be the least .

cost alternative because of the expenses asso-ciated with setting up and programing therobotfigure 45. Moreover, at sufficiently highproduction volumes, fixed automation will becheaper because there is no need to trade offperformance for flexibility. In general, robotstend to be the low-cost alternative ,for annualproduction volumes of roughly 300,000 to 3million units."

The clot in figure 46 has been well publicizedby Unirnation, one of the largest robot manu-facturers in the United States. It compares the

"R. Allan. "Busy Robots Spur Productivity," IEEE Spectrum,Sepiember 1979, p. 3i.


Figure 45.Manufacturing Costs for Robots, Hard'Automation, and Human Operators as a Function

of Production Volume

RobotsRob

Humanoperators

Hardautomation

ge forwhichrobotsare thelow -costalternative

Production volumeSOURCE Of hce of Technology Assessment.

Figure 46.Ccst Comparison for Human Operatorand Robot Assuming One-for-One Replacement and

TwoShift Operation in the Automobile Industry

13

notot Price $40,000Useful life-8 years at two shiftsCost of money-11 percentInstallation cost $12,300Maintenance cost $1.05 per hourPower cost 35 cents per hourOverhaulDepreciationInstallationMoney cost

40 cents per hour$1.25 per hour80 cents per hour65 cents per hour

_ Human operator hourly cost

53

1960i

1963 1966 1969

- Year

Robot hourly cost

1972 1975 1978 1981

SOURCE n. Allan,"Busy Robots Spur Productivity," IEEE Spectrum. September1979, p. 31.

costs for one of their robots with the costs ofwages plus fringe benefits for an autoworker,assuming the robot to be a direct-replacement.-According to figure 46, hourly costs for indus-trial robots have gone up only slightly sincetheir introduction in the 1960's, a period over

r r

the robot is included, but not the engineeringcosts for the application. For a new installa-tion, development costs, including programingand general debugging, can easily total twicethe purchase price of the robot.

In electronics, robots can contribute to quali-ty and reliability by minimizing mechanicaldamage to delicate partswhich also cuts di-rect costsand by helping improve cleanliness.The latter is particularly important for large-scale ICs, where "clean rooms" with levels ofdust and other contaminants reduced to verylow levels are required. Because people bringcontaminants into the production area withthem, automation helps raise yields and reducefabrication costs.

Beyond the now routine applications likespray painting and welding lies a great deal ofscope for robots that extend or improve on hu-man capabilities. Some of these robots will belarger than those currently on the market,others smaller. While a few robots now avail-able can handle loads in the range of 500 to2,000 lb, most are designed with lifting capaci-ties in the range of 50 to 200 lb. Until recently,robot , intended for low loads (e.g., under 10lb) and precision work have been rare. Limita-tions on precision ai.d repeatability haveplaced severe constraints on potential applica-tions.,"

Robots and JobsDespite the fact that robots are simply one

type of flexible automation, with roots in anumber of familiar manufacturing methodsand that automation itself is as old as the in-dustrial revolutionit is as difficult for manypeople to be dispassionate and objective aboutrobots as it is for them to regard nuclear poweras just another means of generating electrici-

"Electronics firms need robots with high accuracy becausethe small and delicate parts used in electronic equipment areso easily damaged. Nippon Electric Co. (NEC) has recently de-scribed a machine with a load capability of about 5 lb and aclaimed positioning accuracy of 0.00016 inches, an order of mag-nitude better than previous high-precision robots. NEC plansto use it in circuit board assembly, as well as wire bonding for

242 international Competitiveness in Electronics

ty. The entire set of technologies for automat-ing manufacturing and services poses very realthreats to the employment opportunities andcurrent job skills of a large segment of the U.S.labor force, as discussed in chapter 9. But itis the whole family that is the proper focus ofattention. . "While it is too early to predict thefull range'ange of impacts of computerized manufac-turing, it is likely thatas with most instancesof major technological change these willcome relatively slowly and in piecemeal fash-ion. Just as these impacts are likely to be ran-dom and incremental, many will be unex-pectedand to the extent that they are, peo-ple and institutions will be unprepared forthem.

Market Growth

If the effects of programmable automationwill not become visible overnight, this is in partsimply a result of time scales for productionand installation; rates of increase will be high,but' total penetration will mount rather slow-ly. Figure 47 gives estimates for worldwiderobot sales through 1990. According to this pro-jection, the market will exceed $3 billion by theend of the decade, an increase of nearly 10times' over the 1980 level. Other estimates rangeboth higher and lower. To set these figures inperspective, note first that spending for capital

Figure 47.Worldwide Annual Sales of Robots,Past and Projected

400

1=3USA

C=1Japan West

GermanyOther

WesternEuropeancountries

200

100

0

equipment in U.S. manufacturing industries iscurrently $80 billion to $85 billion per year, andsecond that during the 1980's expenditures onrobots are expected to remain well under 10percent of total expenditures just on automatedequipment.51.Again, from a job displacementviewpoint, all types of automation must be con-sidered.

Figure 48 illustrates the growth in sales byapplication expected in Japan over the period1980.90. While many of the robots sold in 1980were for use in casting, metal forming, (i.e.,forging and stamping), and painting, the pro-jections in the figure indicate that these ap-plications will be far outstripped by assembly,welding, loading and unloading of machinetoola, and inspection. Some observers predicteven more rapid market growth in the UnitedStates than figure 48 shows for Japan.

When markets grow this rapidly, a good dealof technical and market volatility can be ex-pected, with many opportunities for entrepre-neurial firms pursuing innovaive technologies.While there are no guarantees that robot man-ufacturing will follow a path similar to semi-conduCtors, it would not be surprising to seestartups in the United Stated challenging es-tablished leaders like Unimation 'and Cincin-nati Milacron. The multidisciplinary demandsof advanced robotsboth hardware (microelec-tronics, kinematics and mechanical design, in-strumentation) and software (artificial intel-ligence and computer engineering, automaticcontrol theory, the production engineeringskills needed to integrate robots into theworkplace)create conditions favoring innova-tion and fresh thinking. New companies mayemerge to lead this industry into the thirdgeneration of robotics, just as Unimationoriginally a startup and now owned by a largercorporationled the first and second genera-tions.

"J. T. Woodward, E. P. Seskin, and J. S. Landefeld, "Plant andEquipment Expenditures, the Four Quarters of 1982;" Surveyof Current Business, September 1982, p. 35 (capital equipmentspending); "Industrial Robotics," Emerging Issues in Science

flMnr,L; .,.,tom., r AInfLsrl ol Cnitannr

800

700

600

500

400

300

200

100


Figure 48.Robot Sales in Japan by Application

1990 Projection

Casting Heat Metaltreatment forming

Welding(arc and

spot)

Painting Machinetool

loading

Assembly Inspection Other

SOURCE: Based on data in "Robot Industry to Grow Rapidly in 19131,- Japan Report, Joint Publications Research Service JPRS U9744, May 19, 1981, p. 33.

Japanese firms are applying robots in man-ufacturing more intensively than their com-petitors in other countriesfigure 49. Althoughmost observers feel that the United States stillleads in the relevant technologies, there aremore firms building robots in Japan-130 to150, perhaps five times the number hereand,indications that Japan may be ahead in learn-ing to apply robots in typical factory environ-ments. As figure 47 indicated, future robot in-stallations in Japan are e:Tected to at leastmatch those in the United States.

While the critical technical problems in thefurther development of robots center aroundmodeling and control. the critical imulemen-

process. More than half the costs of typicalbatch manufacturing operations are associatedwith scheduling and otherwise managing theflow of productioni.e., with software, broadlyspeaking. These management and productioncontrol costs involve: getting the right parts,materials, and supplies to the right place at theright time; seeing that shop floor personnelhave the information (now including computerprograms) they need; and ensuring that ma-chinery is available and in good repair whenscheduled for use. Tasks involving productionplanning and machine scheduling, job flows,inventory control, and the related routing andcoordinaung functions might-seem 'straightfor--ward, but in fact they are extraordinarily com-plex; experience shows them to be among themost critical factors in controlling mannfantur-

244 International Competitiveness in Electronics.

Figure 49.Estimated Numbers of Robots in Use, 1980

SOURCE

Japan United'States

Sweden WestGe-many

Italy

'CAM: An International Comparison," American Machinist, November 1981, p. 214.

least some of these costse.g., those associatedwith materials handling, control of the inven-tory of tools, jigs, and fixtures, routir ofpartsbut only when appropriately integratedinto the production environment. Integrationwill require a great deal of rethinking at boththe design and manufacturing stagesrethink-ing for which cheap computing power is nec-essary but hardly sufficient. The potentialbenefits in terms of productivity are huge, butno one anywhere in the world knows at pres-ent how _to realize them.

Computer-integrated manufacturing will af-fect the cost structures of marl;r industries; aslabor productivity improves. fixed costs will ir-

UnitedKingdom

France OtherWesterncountries

compared to blue collar labor costs. Flexibili-ty will make small-batch production more at-tractive; product differentiation and productcustomization will become relatively less, ex-pensive. The result will be far-reichingchanges in the product and marketing strat-egies of manufacturing companies throughoutthe world.-

International Trends

As has happened in so many other industries,Japanese firmswhich began to manufacturerobots by importing technology from theUnited States and Europeare now quite capa-ble of advancinc the state of the art on their


4-0*-Cc

Photo credit: Cincinnati Milacron

Ot welding frame assembly for computer

ns; General Electric, for example, haso build robots under license fromWhile fewer than 5 percent of the

roduced in Japan during 1980 were ex-mports were comparably small), Japanto export 15 to 20 percent of its robotion by 1985..62

as in other industries, the Japanesenentvia the Ministry of Internationalrid Industry (MITI)has designed pro-o encourage and support builders of\. number of Western European govern-re following suit. MITI sponsored sev-nufacturing-oriented R&D programsncompassed robotics during the 1970's.the agency's first steps to support theindustry itself was the establishment

of a leasing program. The Japan Robot Leas-ing Company, Ltd. (JAROL), incorporated in1980, is owned 70 percent by 24 robot manufac-turers and 30 percent by 10 insurance com-panies. About 60 percent of JAROL's capitalhas been borrowed from the Japan Develop-ment Bank and from commercial banks. Theconsortium uses this capital to purchase robots,which are then leased, primarily to smallerfirms.53 While the program is similar to thatoperated by the Japan Electronic ComputerCorp. for financing computers, JAROL doesmore than just lease equipment; its engineer-

, ing staff provides assistance in installing andprograming robots. Among its other initiatives,the Japanese Government has also granted anextra 13 percent depreciation allowance to pur-"chasers of advanced types of robots, whilesmaller firms that buy robots for moderniza-tion or to automate hazardous jobs can take ad-vantage of low-interest loans.54

Much more ambitious is MITI's plan for ajoint R&D program aimed at making robotssmarter.55 This effort, with a proposed annualfunding level of about 30 billion yen (roughly$135 million), will be organized much like othergovernment-sponsored R&D programs in Ja-pan. About 10 companies are expected to beinvolved, plus the Electrotechnical Laboratoryof the Agency for Industrial Science and Tech-nology. Major thrusts planned over the 7-yearschedule beginning in 1982 include:

improved 'sensory *capabilitiese.g pat-tern recognition, force/torque sensors;control algorithms incorporating true adapt-ive or "intelligent" behavior which wouldallow the robot to operate more-or-less au-

53Y. Machida, "Industrial Robot in Japan," LTCB Research,Long-Term Credit Bank of Japan. March/April 1981, p. 4.

"The special depreciation pro Jisions apply to robots pur-chased between 1980 and 1983 that are computer controlled,pave six or more degrees of freedom, and meet specified stand-ards for positioning accuracy. The added 13 percent deprecia-tion in the first year means that a total of 53 percent can be writ-ten off (assuming the 5-year, double declining balance deprecia-tion procedure that is normal in Japan). See "Robotics: They AreSmart and Never Need a Tea-Break," Far Eastern Economic Re-view. Dec. 4. 1981, p. 70.


tonomously, making decisions based onsensory data it receives from the operatingenvironment; andimproved mechanical design, includingmanipulators and mobility, the latter verymuch a controls problem as well.

The program is in part a sequel to previousMITI-sponsored work on remote control de-vices for maintaining and repairing the radio-active portions of nuclear powerplants. How-ever, the robot program will be much broader.

Summary and ConclusionsIn the past, Japanese electronics firms mak-

ing both TV receivers arid ICsnotably mem-ory chips for the merchant markethave em-phasized quality and reliability more heavilythan their counterparts in the United States.This does not mean that the best performingAmerican firms may not have had quality andreliability as good as the Japanese, or that cap-tive manufacturers in the United States maynot have been as good or better than IC makersanywhere. It does mean that specific types ofproductscolor TVs and dynamic RAMshave, in the past, been produced to higher aver-age levels of quality and reliability in Japan.The picture at present is less certainindeed,reliability cannot be estimated until productsare well along in their service lives. It is plainthat American firms have made major effortsto improve quality and reliabilityWith con-siderable success. But it is not obvious that theyhay. ghtmuch less surpassedthe Japa-ns::_:: who have been improving their own per-formance at the same time.

Japanese manufacturers may have succeededin creating perceptions of quality and reliabilityoutstripping any actual performance margins,particularly for color TVs; certainly the strat-egies of Japanese electronics firms have par=allels in other industriese.g., cameras or auto-mobileswhere the emphasis placed on boththe image and the reality of quality had an im-portant role in the penetration of U.S. markets.For American firms to match this image de-mands top management attention.

assurance has often been an orphan in theUnited States. Quality control personnel herehave been viewed as obstacles to production;they have had integral roles in neither designnor manufacturing. Japanese firms learnedmany of the basic techniques of quality con-trol from American engineer s, but they havegone a step beyond conventional practice inmuch of the rest of the world by, for instance,making individual workers responsible for thequality of their own efforts.

Electronics firms in Japan also invest moreheavily in employee training. At all levelsassembly workers, engineers and designers,foremen and supervisors, sales and manage-mentemployees of American electronicsfirms tend, on average, to be less knowledge-able concerning quality and reliability thantheir counterparts in Japan. Although many ofthe recognized authorities in these fields areAmericans, expertise is not spread as widelyhere as in Japan. Moreover, U.S. electronicsmanufacturers may still to some extent be 'pay-ing lip service to quality and reliability. Overtime, the depth of their commitments will be-come more apparent. In particular, it takestime to design quality and reliability into aproduct line.

In general, Japanese electronics manufactur-ers also appear to do a better job of managingthe interface between design and productionMoreover, the characteristically close w3rkingrelationships between vendors and purchasersin Japan's electronics industry evidently yieldsbenefits in quality a.ud reliability. Production

Ch. 6 Manufacturing: Quality, Reliability, and Automation 247

chase a good deal of manufacturing equipmentfrom U.S. suppliers.

Japanese companies automated the produc-tion of TV receivers and other consumer elec-tronic products earlier than most Americanfirms. Extensive applications of robotsinelectronics'and other industrieswill help theJapanese increase manufacturing productivi-ty still further, and will also improve qualityand reliability. At present, robots remain asmall subset of automated production equip-ment with limited impact, but they will be akey part of future manufacturing facilities.And, while robots will spread rather slowlythrough industry in both the United States and

Japanwith uni..redictable effects, as for anynew technologythe cumulative impacts ofthese and other types of factory automationwill be massive, affecting productivity andcompetitiveness, the skill mix in the workforce, and the total number of job opportunitiesin the economy. Computer-integrated manufac-turing will shift corporate strategies in manyindustries toward greater product differentia-tion. Japanese companies can be expected toapply computerized manufacturing technolo-gies at least as fast as American firms, wher-ever there are benefits in terms of cost, quali-ty, worker safety, or product design and mar-keting.

Appendix 6A Quality and Reliability Comparisonsfor Integrated Circuits

This appendix summarizes the data that havebeen made public concerning quality and reliabili-ty levels of .:hips supplied by American and Japa-nese firms to the merchant market. The most wide-ly noted comparisons have come from Hewlett-Packard Corp., a U.S. firm that purchases largenumbers of semiconductors on the merchant mar-ket, and also manufactures ICs for internal use.

Quality Levels

As indicated in table 6A-1, part A, at the end of .

1979 the quality levels of Hewlett-Packard's U.S.suppliers were poor relative to Japanese firms.While Hewlett-Packard had an obvious interest inpressing their suppliers to provide high quality, thisdata is not just another case of a purchaser play-ing its vendors off against one another; the semi-conductor industry has generally accepted Hewlett-Packard's test results as valid, although offering avariety of explanations for the relatively poor show-ings bydomestic manufacturers.

The data in table 6A-1 are all for 16K RAMs; in-deed, most of the public discussion of quality hasfocused on RAMs, because they are bought in largequantities by many customers and have become a

during 1980, but that they continued to trail Japa-nese firms. Improvements by Japanese suppliersover the time period covered in the table were negli-gible. Note that failure rate after burn-inparts Band C of the tableis essentially an indication ofinfant mortality failures, and thus more closely as-sociated with quality than with reliability. Needlessto say, conclusions based on such results should begeneralized only with care; the table refers solelyto circuits purchased by Hewlett-Packard, and dif-ferences from shipment to shipment even from asingle manufacturer can be large.

Table 6A-2 presents data gathered for OTA onquality levels for RAMs, 4K as well as 16K. As forthe first set of the Hewlett-Packard data, the Japa-nese 16K RAMs were superior by a large margin.The 4K RAM datathough limited to only one Jap-anese vendor are quite different, shbwing theAmerican-made devices to be superior.

Along with quantitative data on RAMS such asthat in tables 6A-1 and 6A-2, purchasers of ICs sur-veyed by OTA's contractor sometimes providedmore general comments. One purchaser, for in:stance, included the fcilowing comparison:

Percent of ICs rejected on


Table 6A-1.-Hewlett-Packard Data on 16K Random Access Memory (RAM) Circuits

A. Reported March 1980.Country of

manufacturePercent failing

incoming inspectionField failure rateper thousand hours)

H-P compositequality indexa

Japan- 1b 0 0.0100 /o 89.92 0 0.019 87.23 0 0.012 87.2

Unitrid States- 1 0.19% 0.0900/0 86.12 0.11 0.059 63.33 0.19 0.267 48.1

aThis index is based on 10 equally weighted factors, of which incoming inspection and field failure rates are two; the othersinclude such things as cost and delivery schedules.

°Evidently, the three Japanese suppliers (not necessarily in order.) were Fujitsu, Hitachi. and NEC. tge American suppliers(again not in order) Intel, Mostek, and Texas Instruments. See The Quality Goes On Before the (JapaneSe) Names Goes On,"Posen Electronics Letter, Mar. 31, 1980, p. 1.

SOURCE: R. W. Anderson, "The Japanese Success Formula: Quality Equals the Competitive Edge." Verbatim Record, Seminaron Quality Control: ,!span's Key to High oroductivity, Washington, D.C., Mar. 25, 1980, p. 40.

B. Reported November 1980.Country of

manufactureFailure rate after

burn-in

Japan- 1

20.05%0.10

3 0.12 Average = 0.17%a4 0.355 0.25

United States- 1 0.602 0.503 1.20 Average = 0.75%a4 0.70

aAverages are not weighted by numbers of circuits from each manufacturer.

SOURCE: B. Le Boss, "U.S. Reject Rate Still Trails Japanese," Electronics, Nov. 6, 1980.

C. Reported April 1981.

Country ofmanufacture

Failure rate after burn-inFirst half

1980Second half

1980

Japan- 1 0.06% 0.042 0.13 Average = 0.25%a 0.13 Average = 0.24%a3 0.40 0.404 0.40 0.40

United States-1 0.60 0.352 1.20 Average = 0.97%a 0.20 Average = 0.35%a3 1.10 0.50

aAverages are not weighted by numbers of circuits from each manufacturer.SOURCE: R. W. Anderson, Seminar on Management, Productivity and Reindustrialization: East Meets West, Washington, D.C.,

Apr. 2, 1981..

Consistent with such patterns, an independent test-ing firm noted that rejection rates following screen-ing and burn-in were twice as high (about 4 per-cent) for American-made ICs as for Japanese (1 to2 percent). End users of ICs generally reported sim-ilar results, several noting that they no longer'felt

Reliability Levels

While screening and other tests can locate defec-tive circuits and :neasure quality, determinationsof reliability must await field experience; long-term

equivalent of average /apnoea() products. Good Japanese lots run at a re-

Ch. 6-Manufacturing: Quality, Reliability, and Automation 249

Table 6A2.-Quality Levels of Japanese and U.S.Random Access Memory (RAM) Circuits

Table 6A3.-Reliability Levels of Japanese andU.S. Random Access Memory (RAM) Circuits

Percent rejected onincoming inspection

Country ofmanufacture

Field failure rate(°,2i, per thousand hours)

Country ofmanufacture Average Range

A. 16K RAMsJapan' 1

20.0062%0.0263 Average = 0.027%4

A. 16K RAMs 3 0.0507

Japan- 1 0.30% United S:ates- 1 0.01672 0.53 Average = 0.87%a 0-1.8 2 0.06873 1.77 0.1-5.0 3 0.088c; Average G.125%a

United States- 1 0.70% 0-2.7 4 0.1072 0.85 Average = 1.7% 0-4.7 5 0.12683 1.07 0-6.2 6 0.34214 4.11 0-12.4 8. 4K RAMs

8. 4K RAMs Japan- 1 0%Japan 1 07% 0-5.3 UnitedUnited States- 1 0 32% 0-2.0 States- 1 0.0b24

2 0.41 Average 0.53% 0-24.5 2 0.05263 0.87 0.1-1.1 3 0.1018

3,,,ufacies are not ,sec;ateo oy nuiroufs of cocuos from each manufacturer C. 1K RAMsr-souncE uasi, .a id hisi,3t.ehi,, of SE31',:.011CIUCtOrS and CTVs United States!

v Japan," report prepared for OTA by Consultant Services Institute..under contract No 0331170, p 72

Japan- 1 0.0756%

UnitedStates- 1 0.0667

tests are very expensive, and burn-in failure:; morproperly ascribed to quality problems. Field failuredata w;sembled by I fewlett-Packard for 16K RAM'swere included in table 6A-1, part A. Table GA-3 con-tains the remainder of the reliability data availableto OTA. Consistent with the Hewlett-Packard re-sults, this shows the reliability of Japanese 16KRAMs to have been markedly better than ArnericfmProducts.

Soft Errors

Failure modes for ICs can be divided into "ha d"and -soft" Failures. } lard failures are repeatable mdfinal; the device no longer functions properly. Incont rast, soft failures are random and nonrer eat-able. Alpha radiation can cause soft errors in I AMcircuits, a problem that was not appreciated nntildensities reached 1.:1:. The radiation-emint d atlow levels from ceramics and other materials isedin packaging It;s:--sortaitimes causes a bit ste7 ql ina memory cell to switch front -0" to "1" ori viceversa., The result is a soft error that appears 'vhenthe contents of that cell are next recalled.

'T I. MO. SOft Errors in VI.S1: Present and Future," /LT!:ems nn Components. Hybrids, and Afarni factoring 'Thcfmol

(AIM Ii noble I979, p. '177.

Trans-oy.

aAverages are not weighted by numbers of chips from each manufacturer.SOURCE "Quality and Reliability Of Semiconductors and CTVs: United States

v Japan." report prepared for OTA by Consultant Services Institute.Inc., under contract No 0311170.0, p. 72.

The frequency of soft errors caused by alpha radi-ation can be reduced by a number of techniques.which Japanese manufacturers evidently imple-mented more rapidly than American firms-per-haps because Japanese semiconductor firms were,.more willing to accept the extra costs. One pur-chaser of 64K 12AMs reported soft failure rates of10"a per hour for circuits from two Japanese manu-facturers; the rates for the products of a pair ofAmerican firms wire 10-3 and 1076 failUres perhour.3

,"Quality and Reliability of Semiconductors and CTVs: G.S. v. Japan,"op. cit., p. 73. Several years ago, the alpha-induced soft error rate forFujitsu's lid< RAM was reported to be three orders of magnitude betterthan that of Musters device. The differeoims were attributed to the de-signs of the circuits. See J. G. Pose, "Dynamic RAMS: ;Nita! to ExpectNext," Electronics, May 22, 1980, p. 119. Tile resistance of Mostek's IRKHAM to alpha radiation has since been greatly improved, and U.S. firmain general have adopted measures that substantially reduce sensitivityto alpha particles.

CHAPTER 7

Financing: Its Role inCompetitiveness in Electronics

ContentsPage

Overview 253

Sources of Funds and Financial Leverage in the United States and Japan 255:tnui Exiornal Firiancin "57

Pisk 2581)1..' lntlimiimn and Banking Practices 26(3Eitects mu! Financial I.VVOnigc on Tax Paynumts 262

k Absurption in Japan 262

Other Factors in Costs of Financing 263"17.1. anti Diversification 263

Nat.os 264Costs of Capital for Elei !rimics Firics in the United States and Japan: Summary 266

Financial Structure: An International Comparisor 268268280

Fraii.a 2891,1:!-.! 292

Summary and Conclusions 294

List of Tables

- 11.,si- (1a6ital 11.S. and lapatieso SemiconductorLii1C111,1101 1)V Chase Fitoincial Policy

acd F\ternal Sources of Corporate Financiiii.;7.56

To . JuimiR Ratios fur Selected din] Janese Electuinics Firms 258Savi:1_Is Pates in Several Industrial ()trims 264

iimistiliold Asset; in the l'inted States and Japan. 1978 265Adjme,tH Pates of RI! mire in the I nilmud States anti japan 266' \11,',-11.c;',,Ile ".S. Venture hive:anoint Activity 270

Internally Cieliorated Funds as a Percent of' 'Fotill.Capital From All Sour':-; 27:1.,11. Ay.,ets of and Japanese Semiconductor Nlanufacturersml). Sources of External Financing for U.S. and Japanese Corporations 28261. Short- and Long-Term Debt as a Percent of r.till('',apitalization tom

:.S. 111(1 fapiinese Electronics Firms 2113

6.1 Asset (:ategories as Listed on Balance Sheets al Electronics Firmsin the l'Itited States and japan 284

6.t. Aver:we Assets and Liihilitimus of laparnise Commercial Banks 28564. ilitortA Rates and Price Inflation in Japan 21665. Household Borrowing in Japan and the United States 288oti.lfolders of Liiihilities and With the Nonfinancial Sector in the

Slates and France 9 9(1

List of FiguresPage

incicaso in Capital Costs fur Iligh-Volinne Integrated Circuit Production, lane 2711,1ales of Capital Spending hy 1.1.5. and Japanese Semiconductor Firms 27,1lila Filnity Patios for Japitnesti ElectricalfElectrimi, 282

CHAPTER 7

Financing: Its Role inCompetitiveness in EJe6tronics

OverviewDeclines, real or imagined, in U.S. competi-

tiveness in electronics have been ascribed atvarious +.. imes and by various people to suchcauses as: unfait competitive tactics by foreignfirms, trade barriers that keep American prod-ucts out of overseas markets, government sub-sidies in other countries, and costs of capitalthat are lower than in the United States. Low-cost investment funds are said tobe availablein countries like Japan for-reasons rangingfrom higher rates of consumer savings to allo-cations of capital by gollernments or direct sub-sidies.

This chapter deals with only this last set ofpossible causesthose related to corporate fi-nancing. Although limited in scope, the discus-sion has clear implications for other facets ofcompetitiveness. For example, financing costscould be lower where a protected home marketreduces risk and provides a stable foundationfor international operations. Government sub-sidies might he indirectly channeled throughfinancial markets as implicit or explicit loanguarantees, as well as in more obvious formssuch as grants for research and developmentor tax havens encouraging regional develop-ment.

In mature industrial economies, a vast andvaried network of channels links companiesseeking funds with individuals and organiza-tions that have moneys to lend or otheiwise in-vest. The capital markets where transactionsbetween those seeking and those providingfunds take place accommodate both direct andindirect investments, for short time periodsand for long. Among the direct and long-term__AL _ L

ownership relation with the issuing company,but receive a fixed rate of return, as well as pos-sible capital gains (or losses).* Shareholders ac-cept a variable rate of return in the form of div-idends, as well as changes in the value of thestock depending on the success of the com-pany. Both stocks and bonds are traded in ac-tive secondary markets in the United Statesand many other industrialized nations. In gen-eral, holders of debtof which bonds are onlyone typehave first claim on the residual as-sets of a corporation in the event of liquidation;the claims of stockholders are subordinate.

Highly developed capital markets such asthose in the United States also provide indirectfinancing mechanismsi.e., one or more finan-cial intermediaries are interposed between theinvestor and the final recipient of funds. Banksare the most common intermediaries. Investorsdeposit moneysfor instance, in ordinary sav-ings accountswhich the banks then lend tobusinesses. Other financial institutions func-tion in generally similar fashione.g., the post-al savings system in Japan, an important chan-nel for capital that ultimately helps financeJapanese industry. Investment banks, insur-ance companies, and pension and retirementfunds are other examples of financial interme-diaries.

The fundamental questions in this chapterdeal with costs of capital faced by electronicsfirms in various parts of the world. Spukesraenfor American companies have often compared

'Several types of corporate bonds exist. Straight bonds carrya fixed interest rate, but their market price varies with economic

International Competitiveness in Electronics

their costs for investment fundswhether debt(bank loans, bonds) or equity (primarily stockissues)unfavorably with costs in other coun-tries. in particular, costs of capital in Japan areoften said to be as little as half those in theUnited States. Some observers also claim thatthe pool of funds potentially available for in-vestment in the U.S. industry is too small.

Such concerns are particularly relevant forthe rapidly growing, high-technology portionsof the American electronics industry. Firmswhose business centers on semiconductors,computers (including software), and even themore rapidly expanding portions of consumerelectronics (e.g., electronic games) can findthemselves with markets outstripping theirability to finance expansion.

Problems in securing funds for rapid expan-sion riot only of production, but of R&D andproduct developmentare compounded by therapidly increasing capital intensity of someportions of the electronics industry. Semicon-ductor manufacture is a prime example; capitalcosts are going up rapidly, not only becauseof escalating design cost as circuits becomemore complex, but alsc 'because new genera-tions of production equipment are much moreexpensive (ch. 3). Given the predilection of U.S.firms, in electronics as in other parts of Amer-ican industry, for relying on internally gener-ated fundsi.e., retained earnings and depre-ciationwhenever possible, financial mana-gers have often been hard pressed to securefunds for growth.

Because the chapter centers on costs of cap-ital, interest rates and the mechanisms bywhich they are determined become one of thefundamental bases of comparison. The cost tothe borrower of acquiring funds is the interestrate' on the loan or bond. Costs of equity, fol-lowing conventional practice, can be relatedto costs of debt. In countries with well-devel-oped capital Markets and -modest levels ofgovernment interventionas in the UnitedSteinsrnerket-rietermineri interest rates are

The interest rate thus serves a critical func-tion in the economythat of the price for bor-rowed funds. This price serves to allocatefunds so that the pool of available capital goesfirst to the most productive investments. Themechanism is as follows. Managers of profit-seeking enterprises make investment decisionsby comparing their costs in acquiring fundswith the expected profits from the uses of thesefundsi.e., with the returns on alternative in-vestments. These projects might be new man-ufacturing facilities, R&D programs, or the ac-quisition of other corporations. If the antici-pated returns are greater than the costs of ob-taining funds, then the investment might bemade using money generated within the enter-prisee.g., from retained earningsor fromoutside capital markets. In either case, the in-terest rate is the primary factor in determin-ing the cost of financing the project. For ex-ample, if market interest rates are high, a cor-poration might choose to invest in securitiesrather than in its own business. In general, lessattractive investment projects will be post-poned when interest rates rise, the market serv-ing to allocate funds to other uses both withinthe firm and among various companies seek-ing financing in the capital market.

The market-driven process described doesnot always function ideally, but as a rule in-terest rates allocate funds quite efficiently. Still,governments can act in various ways to influ-ence investment decisionseither on a case-by-case basis or by favoring some sectors of theeconomy over others. Outright subsidies andloan guarantees are two of the more obviousand common tools. Less visible and less directpolicies are also possible; some of these are ex-plored in the discussion of financing practicesin Japan later in the chapter. Where govern-ments intervene in capital markets, one con-sequence can be higher interest rates for allborrowers except those favored by the govern-ment.

To explore international differences in

Ch. 7 Financing: Its Role in Competitiveness in Eiectronfcs 255

More limited discussions of France and WestGermany follow. The objective is to understandthe effects of financing patterns on competi-tiveness in the international marketplace,

where governments may try to complementcorporate strategies or to implement nationalstrategies.

Sour es of Funds and Financial Leverage in theUnited States and Japan

Many executives in the U.S. electronics in-dustry believe the firms they manage to be con-strained in efforts to defend or expand inter-national markets by relative costs of financing,and in some cases also by shortages of capital.The electronics industry is not alone in thisconcern. Other American industries, especiallythose facing intensified international competi-tion, voice the same complaint, more especiallyif they feel threatened by the Japanese. The ar-gument has been articulated bestand empir-icai* supported in mcst detailby the U.S.semiconductor industry, largely because itsrate of expansion and changing technical char-acter place extraordinary demands on the fi-nancing capabilities of independent merchantfirms. The semiconductor industry's positionwith regard to financing is summarized helovt:to the extent possible the argument will be gen-eralized to other sectors of electronicsi.e.,computers and consumer products.

The basic contention of the semiconductorindustry is straightforward, and for the mostpart directed toward the industry's primary for-eign competitor, Japan: the ability of Japaneseelectronics firms to gain market positionagainst American companies over the past fewyears, both in the United States and abroad, hasbeen eased by cheap capital. (The meaningsthat attach to cost of capital will become clearerbelow.)

For one reason or another, in this view, Japa-nese corporations in many industries enjoycosts of capital markedly lower than their

them at lower prices, At times, U.S. firms havealso associated low-priced products with "un-fair" practicesin international trade (see ch.11).Certainly, q broad range of business tacticswhether or not fair within the accepted frame-work of international tradeare easier to im-plement if capital is inexpensive.

The U.S. semiconductor industry has also as-serted that favorable access 'to funds has en-abled Japanese manufactures to add capacityin advance market demandindeed, to cre-ate excess capacity even in times of recessiona "luxury" decidedly unavailable to Ameri-can firms. As a consequence, when the econ-omy improves, the Japanese are better placedto quickly move into expanding markets, whiletheir competitors here struggle to build capac-ity and catch up. Finally, it is alleged, amplesupplies of cheap capital allow. Japanese corpo-rations to spend lavishly on the advanced R&Dso necessary in this rapidly changing field.Lower costs of capital_ together with full con-trol over their domestic market, are viewed asprimary underpinnings of Japan's global strat-egy.

What are the perceived reasons for these low-er capital costs? Two main causes are frequent-ly cited, along with related structural featuresof the financial system in Japan: 1) the distinct-ly different capital structures of Japanese elec-tronics companies; and, 2) the very high rateof savings within the Japanese economy. Forstructural reasons; Japanese firms can tap relatively large amounts of nonequity funds, pri-


revenues to finance growth. NonequitY funds,it is claimed, tend to be less costly.

The second source of Japanese advantage- -high savingsby increasing the pool of fundsavailable to be lent, should depress interestrates. This would have the effect of making alltypes of investment capital less expensive com-pared to countries where savings are a small-er proportion of gross-national-produck-Savings rates are discussed in more detail in a latersection; household savings in Japan run atabout 20 percent of income--nearly four timesthe rate in the United States. There is littleagreement on why the savings rates in differentcountries vary so'much, and in particular whythat in Japan is so high and that in the UnitedStates so low. Variations in the average pro-clivity of individuals in different countries tosave under otherwise similar circumstancesappear to be a factor; so do the extent of socialwelfare programs and differing tax structures.

Ccimbined, these two sources of financial ad-vantage are said to give Japanese electronicsfirms capital costs barely half those of theirAmerican competitorsin 1980, about 9 per-cent compared to 15 to 18 percent for U.S.semiconductor firms.1 Such a result, if true, hasimplications for competitiveness in many otherindustries.

The conclusions of the Chase Financial Pol-icy study cited above are summarized in table5.1. According to Chase's calculations, the typi-cal Japanese manufacturer of semiconductorsenjoys substantially lower costs of capital thanmerchant firms in this country. Only Matsu-shita (table 51) incurs financing costs larger

"U.S. and. Japanese Semiconductor Industries: A FinancialComparison," Chase Financial Policy for the Semiconductor In-dustry Association, June 9, 1980, p. 2.5. The most thorough dis-cussion so far of the impact of corporate financial structure onrelative costs of capital, this report seeks to quantify Japaneseand U.S. financing costs with considerable care to identifyingthe sources of the differences,

A more receot anal; ::s comparing industry as a whole in theUnited States and Japan finds average costs of capital for 1981to be 16.6 percent here versus 9.2 percent in Japan. See, "AHtsterical Comparison of the Cost of Financial Capital in France,the Federal Republic of Germany, Japan, and the Uiiited States,"Department of Commerce. April 1983. In this report, no attempt

Table 51.Costs of Capital for U.S. and JapaneseSemiconductor Manufacturers as Calculated

by Chase Financial Policy

Weighted averagesof debt and equity

° costs as ofJune 4, 19808

U.S. companies (calculations in dollars)Advanced'Micro Devices 17.7%Fairchild Camera and Instrumentb 15.5

16.8!Mel . .............. ..Irrtersilc 21.1Mos7,ekd 16.7Motorola 13.8National Semiconductor 17.4Texas- Instruments 16.5

Japanese companies (calculations in yen)Fujitsu 8.8%Hitachi 12.1Matsushita Electric 17,1Mitsubishi Electric 7.7Nippon Electric CO. (NEC) 7.7Toshiba 7.7

aln terms of required overall rate of return on Invested capital.b Subsequently acquired by schlumberger.cSubsequen9y acquired by General Electric.dSubsecitieritly acquired by United 7ectihologies.SOURCE:"Q.S. and Japanese Semiconductor Industries: A Financial Com

garlson,- Chase Financial Policy for the Semiconductor IndustryAssociation, June 9, 1980, tables 4 and 9, pp. 5.3 and 7.6.

than any of the U.S. companies, at least for thetime period examined. Nonetheless, the rangein capital costs faced by firms in either coun-try is relatively large.

There are two major reasons for the wide di-vergences in capital costs in table 51. First, bor-rowing costs used in the calculations for Japa-nese firms were lower than rates for Americancompanies, The second primary solircQ:of dif-ference lies in the dissimilar capital structuresof corporations in the two countries--thegreater use of debt by Japanese firms, principal-ly in the form of bank loans (most Americanfirms with substantial debt carry this in theform .,f bonds).

If Japanese firms use.substantially more debtthan U.S. companiesas they doand if debtfinancing is less costly than equityaS the com-putational methrd used by Chase assumes-4101-1 n frith] rnet riarivarl from a wpichtpd aviar.

Ch. 7Financing: Its Role in Competitiveness in Electronics 257

would imply cost advantages for Japanese com-panies, not only in electronics but in any indus-try making similar use of leverage.

Internal and External Financing

Table 52 illustrates something of the rangein international differences in corporate fi-nance. Japanese capital structures are heavilyweighted toward external financing. Japanesecorporations, on the average, received less thanhalf their capital from internal sourcesi.e.,from depreciation and reinvested profits. And,while Japan is at the high end in use of exter-nal capital, the United States is atthe low 'end,relying much more heavily on internally gen-erated funds.

The category of external finance includesboth loanswhich in all,five countries are ex-tended primarily by banksand securities. Thetwo- major categories of securities are bonds(like loans, debt) and stocks, reptesenting equi-ty holdings.2 Note that Japanese firms rely

,much more heavily on loans than securities(either loans or equity) for their external fund-ing; in general, companies in Japan employmuch higher financial leverage than do Ameri-can corporations (leverage can be defined inseveral ways, perhaps the most common beingthe ratio of debt to equity in a firm's capitaliza-tion). Table 53 compares debt/equity ratios forU.S. and Japanese electronics companies. Thereasons that corporations in Japan make

For a standard introduction to...corporate finance, see J. C.Van Horne. Fundamentals of Financial Management, 4th ed.(Englewood Cliffs. N.J.: Prentice -Hall, 1980).

greater use of financial leverageand theconsequencesare take'i up later.

The conclusions of the Chase study concern-ing the impact of debt financing on capitalcosts in Japan are grounded in well-acceptedmethods of calculation. The cost of capital fora particular investment can be estimated usingthe relative proportions of the company'ssources of capital as weights in the computa-tion for the investment. For example, if a com-pany pays 15 percent interest on debt instru-ments, and its risk-adjusted cost Of equity (ex-plained below) is 20 percentand if the debt-equity ratio is 1.0then the firm's overall costof capital would be 171/2 percent. The returnsexpected from a given investment can then becompared to this estimated cost of capital. Thecomputational method is deceptively' simpleandexcept for various subtleties involved indetermining the appropriate interest rate fordebt and the risk measures for equitycan beapplied in straightforward fashion.

All other things equal, then, Japanese firmswould enjoy clear financial advantages, fromtheir greater relative amounts of debt (higherleverage) so long as the interest rate on debtis less than the risk-adjusted cost of equitythe normal case. Several questions follow: Iffinancial leverage lowers costs of capital, whydon't U.S. firms emulate the Japanese by usihgmore debt in their capital structures? Wouldn'tstockholders benefit from this choice by earn-ing higher returns? There are also potential taxbenefits: since corporations can deduct interestpaid on debt as an expense, but not dividendspaid to stockholders, would not greater use ofdebt decrease Federal tax obligations and in-

Table 52.Internal and External Sources of Corporate Financings

Internal finance(reinvested profits, External finance

depreciation)Ratio of internal

Loans Securitiesb Total to external financeUnited States 69.4% 12.4% 18.2% 30.6% 2.27Japan 40.0 49.0 11.0 60.0 0.67United Kingdom 51.4 10.3 38.3 48.6 1.06West Germany 63.1 29.6'. 7.3 36.9 1.71France 65.0 27.4 7.6. 35.0 1.861966.70. These Patterns have Probably not chanaed °really.


Table 53.Total Debt-toEquity Ratios forSelected U.S. and Japanese Electronics Firms

1975 1979

United StatesAdvanced Micro Devices 81% 8%Control Data Corp. (CDC) 38 20Digital Equipment Corp. (DEC) 30 32General Electric 41 25Honeywell 65 32IBM 4 17Intel 0 0Motorola 28 30National Semiconductor 25 37RCA 106 125Texas Instruments 14 21

Japan°Fujitsu 200% 190%Hitachi 160 96Matsushita Electric 14 16,Mitsubishi Electric 370 270Nippon Electric Co. (NEC) 350 400

aThe financial data for Hitachi, Matsushita, and Toshibaas used by ChaseFinancial PolicyIncludes affiliated trading companies among the consolidatedsubsidiaries, while that for Fujitsu, MitsubiShi Electric, and NEC does not. Seethe Chase Financial Policy report cited in the source note below, p. 6.1.

SOURCES: United StaLS;Derived from annual reports; also "Financial Issuesin the Competitiveness of the U.S. Electronics Industry," report pre-pared for OTA by L. W. Borgman & Co. under contract No. 033.1550,0,pp. 52, 56. JapanDerived from data in "U.S. and Japanese Semicon-ductor Industries: A Financial Comparison," Chase Financial Policyfor the Semiconductor Industry Association, June 9, 1980, Appendix:Japanese Semiconductor Companies, Financial Statements and Sup.porting Schedules.

crease aftertax profits? If so, isn't this anotherreason to encourage U.S.- electronics firms toincrease their leverage? (Japanese tax treatmentof interest payments is similar to. U.S. law inthis respect.) At this point, the layperson mightthink that Japanese firms have simply taken ad-vantage of financing choices also open toAmerican companies.

Risk

The answers to the questions above, and thekey to understanding the U.S. electronics in-dustry's unhappiness with Japanese financingpractices", relate to a second aspect of finan-cial decisionmaking risk. Investment deci-sions inevitably involve risks for those whosupply fundswhether external funds or inter-nalbecause there can; e no certainty that fu-ture cash flows will be sufficient to compen-sate investors. In essence, the risks borne byinvestors are of two types. First, cash flows are

than others. In one year, the funds remainingafter expenseshence available for distributionto shareholders or for retention in the enter-prisemay be plentiful; in another, such mon-eys may be scarce or nonexistent. Stockholdersare generally.believed to desire stable earningsfrom year to 'year, accepting greater variabili-ty in rate of return only if compensated by ahigher average return.

In contrast to stockholders, who share in theownership of the firm, creditors merely lendit funds; they generally have first claim on cashflow, as well as on the assets of a firm, andreceive a "guaranteed" rate of returni.e., theinterest rate on bonds or other debt instru-ments. While creditors seldom share in the firsttype of riskvariability in returnsthey maysometimes choose to subordinate their claimsrather than force a firm into bankruptcy. In therecent example of Braniff International, theairline's creditors several times allowed pay-ments of both principal and interest to be de-ferred before Braniff finally entered bankrupt-cy.

The Braniff case illustrates the second typeof riskloss of all or part of the investment it-self, as well as loss of revenues from interestpayments or distributions of profits. This is arisk borne by both owners and creditors. Butbecause creditors have first claim, they aremore likely to recover at least part of their in-vestment in the event of business failure. Thisis the reason interest rates on debt are generallylower than the risk-adjusted cost of stockhold-ers' equity: holders of debt face lower risksbecause they have first claim on assets. At thesame time they must accept a nominally fixedrate of returngenerally lovirenthan that accru-ing to shareholders. (In fact, the effective rateof return on bonds is not necessarily fixed, aspointed out earlier, but this is not importanthere.)

The discussion above is necessarily schemat-ic, and corporations- can avail themselves ofother methods of financing, which fit into thesubordination ordering in various ways. Butas a general rule, common stockholders come

Ch. 7Financing: Its Role in Competitiveness in Electronics ' 259 .

other creditors and investors have been paid.This subordinated status makes shareholderssensitive to the degree of leverage employedby the firm; their exposure to risk increaseswith higher leverage. Not only does more debtin the firm's capital structure tend to increasethe variability of returns to shareholders, butadded debt worsens their position in the eventof a forced_ liquidation.TypicallyT- commonstockholders must be compensated throughhigher returnswhich can include capital ap-preciationbefore they will accept the risk in-herent in greater leverage.

As a consequence, adding more debt will notnecessarily lower a firm's cost of capita1,3 In-deed, neglecting tax effects, the choice of debt-equity ratio, over rather wide ranges, shouldhave little, if any, impact on capital costs. Evenassuming no increase in interest rate as a firmborrows morewhich is not very realisticthelower costs of debt are generally offset by thehigher required returns to common sharehold-ers as leverage increases. Several cautions mustbe adde`d. While this conclusion is commonlyaccepted as applying for U.S. capital markets,it is not clear that it always .holds in the sameway in other countries. Furthermore, taxes domatter, and the fact that interest paymentslower a company's tax bill usually would arguefor adding to the proportion of debt in a firm'scapital structure. But Et some point more debtwill be accompanied by higher interest rates,since the debt itself becomes increasingly riskyfor potential holders.

With all of this said, how is it possible thatJapanese companies can, on the average, em-ploy debt-equity ratios markedly higher thanAmerican firms, without seeming-to bear high-er costs of both debt and equity? The usual re-sponse holds that the Japanese financial systemdiffers from that in the United. States, andforces that tend to raise the cost of capital asleverage increases are absent in Japan (or func-tion differently than they do here). This impliesone or more of the following:

1. Japanese investors exhibit risk aversioncc--

2. Some Japanese investors are acceptingrisks for which they receive less compen-sationfor whatever reasonthan theydesire.

3. Some classes of borrowers in Japan paypremiums for funds, these premiumscounterbalancing the low rates available

" to other borrowersor, alternatively, somepotential borrowers cannot get funds at allbecause of capital rationing.

4. Some risks which private investors in theUnited States must bear are, in one wayor another, reduced for private investorsin Japan.

Each of these four possibilities will be brieflyexamined.

Risk Aversion Behaviors

Financial and busine3s risks must be ab-sorbed within any system that operates totransfer funds from savers to commercial bor-rowers. The presumption is that people havean aversion to risk and, if they'are to acceptsuch risks, they must be compensated by in-terest payments, capital,appreciation, or divi-dends on shares. Still, it is not necessarily truethat all peopleor all economieswill exhibitidentical patterns of risk aversion. Japanese in-vestors, for example, might demand less re-muneration for a given level of risk thanAmericans. (It is also possible that the Japaneseare less reluctant to postpone consumptiona possible explanation for the higher savings ,rate mentioned earlier, with its tendency toforce down interest levels.)

Compensation for RiskThe second possibility suggested above was

that some individual or-institutional investorsin Japan might be compelled to accept ,lesscompensation than they desirei.e., less thanthey would receive in a capital market thatfunctioned differently. It does appear that Jap-anese bankswhich provide much of the cap-italization for electronics firms throughloansare accepting what are essentially quasi-

esc


crated with equity. While banks can diversifythese risks by maintaining a portfolio of cor-porate loans, to the extent that such risks aresystematic, diversification will be ineffective("systematic risk" is simply that which cannoibe reduced by diversification). The question re-mains: Why do banks in Japan accept finan-cial risks that would not be acceptable else-where in the world? This question is taken uplater in the chapter.

Preferential Treatment ofSelected Borrowers

Some observers assert that "target" indus-tries in Japan area selected to receive bank loansat interest rates well below market levels.4 Thiswould imply, in the normal circumstance, thatother borrowers wil! pay higher than marketrates; still Other potential borrowers might notbe able to secure funds at all. The bias is usu-ally alleged to favor large firms at the expenseof the far greater number of small establish-ments in Japan, and to favor growth industrieseven though such industries may, in theshorter term, offer rates of return both lowerand more variable than alternative invest-ments. In general, both semiconductors andcomputers would be classed as rapid growth,long payout industriesin Japan as well as inthe United States.

If some borrowers are, in fact, favored withlower than market interest rates in Japan, thisquestion follows: Why doesn't competitionamong lenders force Japanese banks to allocateresources to the firms and industries promis-ing the greatest returns, as in the United States?

4L. C. Thurow, for example, has claimed that Japanese inroadsinto U.S. and world semiconductor markets are financed withfunds provided by,the government-owned Bank of Japan: '.'Butthe Japanese are entering this industry with massive amountsof debt capital ultimately lent by the Bank of Japan. Their aimis to jump directly into large-scale, capital-intensive techniquesof pioduction; proceed rabidly down the learning curve; sell atprices lower than those of the rest of the world; and capture most,if not all, of the market. If American industry limits its invest-ments to those that can be financed by retained earnings, theywill simply be driven out of the semiconductor industry." See,"Prepared Statement of Dr. Lester C. Thurow," Productivity in

Subsidization of RiskThe foregoing question is often answered, at

least in part, by appeal to the fourth pointnamely, that on some loans Japanese banks canshift part of the risk to other parties. In partic-ular, for loans to companies whose activitiesare deemed to further national interests, theJapanese Government may effectively guaran-tee the loan, at least to the extent of providingprotection against bankruptcy. Some observ-ers, indeed, suggest that many loans by Japa-nese banks are simply government loanspassed throtigh the banking system. In thisview, some of the "normal" risks of debt fi-nancing are absorbed by the government ratherthan the banks; interest charges below marketrates reflect government subsidization.

As with the other points raised above, thequestion of whether and to what extent the Jap-anese Government subsidizes risk can be an-swered, at least in part, by empirical evidence.While the actual functioning of the country'sfinancial system is taken up in a later section,the Japanese economy is no different fromothers in-that capital is a scarce resource allo-cated by various mechanisms among an enor-mous variety of investments. If the governmentor the banking community chooses to step inby selecting target sectors to receive capital atrates that were directly or indirectly subsi-dized, the consequences are quite predictable.The target sectors would gain at the expenseof the rest of the economyfor which creditwould normally have to be rationed. In otherwords, no country can subsidize all industrialsectors simultaneously although manufactur-ing might be favored over agriculture, or theprivate sector over the public. In fact, there isno question that capital was allocated by theJapanese Government during the earlier post-war period; what is not so clear is whethermore than remnants of these policies remain.

Price Inflation and Banking Practices)

nne_nnestinn that even the more sonhisti-


is the impact of inflation oc, international finan-cial comparisons. The effects are too complexto fully review here, but differing inflation ratesamong the world's economies are a major fac-tor in apparent differences in costs of capital.The reason is that observed market interestrates depend in part on expectations by invest-ors with respect to price inflation. If expecta-tions cliff& between a pair of countries, thenmarket interest rates will diverge from thiscause alone. But to the extent that the diver-gence in interest rates simply reflects the un-derlying inflation rates in the two economies,differences in costs of capital based on theseinterest rates are not "real." Only a differencein costs of capital after adjustments for infla-tion would confer advantages in internation-al competition.

The difficulty is that future inflation can onlybe projected; the presumption is that the mar-ket mechanisms which set interest rates takesuch projections into account. Interest levelsenter the calculation of capital costs for U.S.and Japanese electronics firms in table. 51 inat least two ways: 1) through the cost of equi-ty computation, which is based on a "riskless"interest rate; and, 2) in the choice of interestrate for the cost of debt. The riskless interestrate applies to investments fOr which the riskborne by the lender can be considered negligi-ble, at least in comparison to the risks of equi-ty. Government notes, bonds, or bills are typi-cal examples of "riskless" investments.

The riskless rates applied by Chase FinancialPolicy were: 10.2 percent for the United States,derived from the June 1980 Treasury bond rate;and 9.0 percent for Japan, the yield on the mostwidely traded debenture (a type of bond) on theJapanese market-10-year issues of Nippon Tel-egraph and Telephone Public Corp. (NTT). Theanalysis takes the degree of risk for these in-struments to be, if not zero, at least small andcomparable in the two countfies.*

These two interest rates do not differ bymuch; indeed, their closeness accounts for part

of the cost advantages calculated for the muchmore heavily leveraged Japanese companies,But are they close in real, rather than nominal,terms? The answer depends entirely on long-term inflationary expectations ,expectationswhich were probably considerably higher inthe United States than in Japan during 1980.*The nominal interest rate differential favoring _

the Japanese might well reverse, and favor theUnited States, could the real rates- be com-pared.

Differences in banking practices between thetwo countries also affect the true costs of capi-tal. For instance, banks in japan typically de-mand that greater compensating balances bekept on deposit against corporate loans.5 Be-cause the firm pays more for the funds it hasborrowed than it receives on its deposits, thispractice raises the effective interest cost of theloan. Large compensating balances mean thatthe usual measure of financial leveragedebt-equity ratiooverstates the true degree of lever-age.

These observations on the effects of inflationand compensating -balances emphasize thecomplexity of cost of capital comparisons asapplied to funds from external sources. Theydo not confirm or deny the general trends intable 51, though perhaps throwing doubt on themagnitudf; of the differences resulting from theChase analysis.

'The Chase study from which table 51 comes did attempt tocompensate for varying inflationary expectations. The Japanesecost of capital in yen was converted to a dollar cost using thedifference between the two interest rates cited above, said torepresent "the premium investors require for receiving interestand principal payments in U.S. dollars rather than yen" (p. 7.7),But by implying that the 1.2 percentage point differencerepresents dissimilar inflationary expectations, this prOcedureamounts to assuming that the riskless rates in the two countries,expressed in dollars. are equal. This seems unlikely: it wouldimply that the capital market in Japan is both efficient and per-.fectly linked to that in the United States, neither of which is true.It is more reasonable to assume that differences in inflationaryexpectations were considerably larger than 1.2 percent inmid-1980.

5y. Suzuki, Money and Banking in Contemporary Japan (NewHaven. Conn.: Yale University Press, 1980), p. 50. Japanese firmsoften keep 25 percent or more of borrowed funds deposited withlending banks. Furthermore, banks are more likely to lend tofirms that are already large depositors. In the United States. com-


Effects of Financial t_everageon Tax Paymehts

While the Japanese Government might sup-port electronics firms through a variety of cap-ital and other subsidies, 014 study by ChaseFinancial Policy summarizied in table 51 isbased solely on a leverage argumenti.e., onthe advantages of debt as a source of corporatefinancing. In the absence of other sources offinancial advantage, leverage provides lowercapital costs primarily through its effects oncorporate tax payments. Although these are nottrivial, the tax advantages that accrue to Japa-nese firms as a result of high debt-equity ratiosreduce their costs of capital by only a fewpercentage points--probably less than 2com-pared to American firmS.. The reasons are out-

.1i ned below.

In order to isolate the effects on cost of capi-tal stemming from tax shields on debt, assumethat interest rates. in the United States andJapan are the samesay 10 percentbut thatcorporate tax rates differ. For purposes of il-lustration, use the nominal tax rates in the twocountries-48 percent for the United States, 40percent for Japan.° For leverage, assume a ratioOf total debt to .total capital equal to 0.67 forJapan and 0.17 for the the United States.* Asa result of the tax shield created by leveraging,costs of capital would he lowered by:

Japan 0 67 (0.4) (0.1) = 0.0268United States .0.17 (0.48) (0.1) =. 6.00816Subtracting gives 0 01864 or 1.864%

'That is. the tax shield created by greaterleverage would give Japanese firms a cost of

"This is the nominal rate for retained income in Japan; distrib-uted profits are taxed at 39 percent. While nominal rates suf-fice for illustration, they bear little resemblance to the taxes thatcorporation:, actually pay after deductions, credit, depreciationallowances, etc. The -effective" tax rates in the two countriesin the late 1970'stotal corporate taxes paid divided by total cor-porate-profitswere about 37 percent in the United States, 29percent in Japan. See H. Gourevitch, A. Wilson. and D. Culp,fax Rates in Atajor Industrial Countries: A Brief Comparison,Congressional Research Service report No. 80-224 E. Dec 15,nitn, p. 8.

'Tile 9.97 fieure is used at several mints in the summary of

capital advantage of about 1.9 percentagepoints compared to firms in the United States.

In fact, this example, overstates the advantagebecause the median figure (0.67) for Japanesefirms ignores the impact of absolute size.Hitachi and Matsushitawhich have less thanmedian leverageare much larger than fheothers; Nippon Electric Co. (NEC), which hasthe highest debt-to-capital ratio-0.80is con-siderably smaller. When the debt-to-capitalvalues are weighted by total assets, the debt-to-(total) capital ratio for the Japanese compa-nies is 0.52, a result that is considerably af-fected by Matsushita, which had negligible levrerage. Using this figure for Japan. along withthe leverage value that Chase suggests in theirstudy as a desirable target value for Americanfirms desiring to reduce their costs of capital-0.33the comparison becomes:

Japan 0 52 (0.4) (0.1) = 0.208United States .0.33 (0.48) (0.1) = 0.01584

-Substracting gives 0 00496 or 0.496 %.

These two changes reduce the cost of capitaladvantage of the Japanese firms from 1.9 per-centage points to only !-alf a point. This secondcomparison is not necessarily either more orless meaningful than the first; the lesson is thattax advantages are quite sensitive to smallchanges in leverage. The computation is farless sensitive to the tax rate itself.

These examples show that, while the use ofleverage does lower a firm's cost of capital, theeffects are relatively small. A cost of capitallower by 2 percentage points might transla'eto manufacturing costs lower by 1 percentagepointnot very significant. The difference thatdoes result can be regarded as an implicit fi-nancial subsidy from the Japanese Governmentvia the tax system. (The question of special taxprovisions for certain industries is independ-ent.)

Risk Absorption in Japan

As mentioned earlier, Japan's banking sys-tPrn nhcnrhc ricicc nnrmnIlliacciimpri 1-11r chirp_


fact the risks of bankruptcy, reorganization,and business failure are increased by thegreater use of financial leverage in Japan. If so,the frequency of failuresespecially duringeconomic downturnsshould be higher thanin countries like the United States where, onthe average, leverage is much lower. Indeed,bankruptcies in Japan have risen to rather highlevels in times of general economic downturn.In 1977, for example, Japanese enterVisesfailed in record numbers.' The rate of bank-ruptcy that year was four times greater thanthe comparable U.S. figure, and these failures,.involved corporate liabilities of over $16 bil-lion, inure than Five times the 1977 level in theUnited States. While 1977 remains the peakyear in terms of both number of business fail-ures and total liabilities, in every year since1976 Japan has experienced more than 15,000.bankruptcies (excluding small businesses) with-total liabilities exceeding 2 trillion yen (roughly$10 billion)." Although bankruptcies in elec-

'G. R. Saxonhouse, "Industrial Restructuring in Japan," Journalof Japan:?se Studies. vol. 5, 1978, p. 273. Elsewhere Saxonhousestates: "Debt- equity ratios which are four or five times the Amer-ican level result in bankruptcy rates which are also four or finetimes the American level. Large Japanese firms cio go bankrupt.In recent years /alma has had two close to $1 billion in liabil-ities bankruptcies." See "Statembnt of Gary R. SaxonhouseBefore the f:ouse Foreign Affairs Committee, Subcommittee onAsian and Pacific Affairs and Subcommittee on International.Economic Policy and Trade," Oct. 1, 1980.

6Only firms with liabilities of more than 10 million yen areincluded. Figures for 1968 through 1980 can be found in "Japan1981An International Comparison,' Japan Report, JointPublications Research Service JPRS L/10760. Aug. 24, 1982, p:7. those for 1981 in "Corporate Failures in Japan Last Year Fell1.5% to 17,610," 1311 Street Journal. Jan. 15, 1932, p. 28.

tronics have been infrequentin part becausegrowth rates and cash flows have remainedhigh, allowing firms to service their debt-,---riskis clearly present.

In the Japanese financial system, these riskstend to be shifted to, the banking community.Banks assume quasi- equity positions by ilLcept-ing debt in hig,y leveraged firms. If a com-pany finds itself in financial difficulty, thebanks are literally forced to take action aimedat reorganization. The alternative is to proceedwith bankruptcy. When large corporationshave faced trouble, the choice, not surprising-ly, has often been restructuringsometimes ac-companied by infusions of even more funds.Typically the banks have forced a completereassessment of corporate strategy; not infre-quently the ailing firm's executives have beenreplaced by a bank-appointed managerial team.

Sometimes observers in the United Statesconclude that these interventions by Japanesebanks serve to reduce risks to businesses, orrisks to the banks, or both: To believe this im-plies believing that bankers are on the averagewiser than managers of industrial corpora-tions. In fact, these interventions do not lessenfinancial risks, but are caused by the muchgreater exposure of the banks. Such interven-tions are less common in the United States be-cause American banks do not provide as greata' fraction of corporate financing.

Other Factors in Costs of FinancingSize and Diversification

Many of the leading international clmpeti-tors in electroi, are larp,,..1, diversified c--,m-panies (ch. 4). ins is the case for Ainericancorporations like GE or IBM, European man-ufacturers like Philips or Siemens, and manyIaDanese electronics firms. But other U.S. en-

These firms have generally depended moreheavily on one or a few product lines than theircompetitors in japan. Size and diversification,affect capital costs quite directly, with the ad-vantages going to big companies with broadproduct lines. Such firms can absorb andspread risks more effectively.

264 IntQrnational Competitiveness in Electronics

own repayment. Diversified companies exhibitmore stability, hence lenders are willing tosupply funds at lower rates of interest. Largediversified firms are also, on average, evaluatedas better risks by bond rating companies likeMoody's 'or Standard & Poor's. Texas Instru-ments taps a lower cost bond market than, say,Advanced Micro Devices; IBM lower than Dig-ital Equipment Corp.; GE lower than Zenith.The conclusion is: costs of capital will behigher for U.S. as opposed to foreign electron-ics companies when the American firms aresignificantly smaller and less diversified.

There is a great deal of variation in the sizesof firms within the electronics industries of theUnited States, Japan, and the European coun-tries. Nevertheless, in Japan and Europe it isprimarily larger companies that are active ona world basis, more the exception than the rulefor companies the size of Nixdorf (West Ger-many) or Oki Electric (Japan) to be strong inter-national competitors. In the United States; thesituations of companies like Mostek, Fairchild,or Intersil have changed dramatically as a re-sult of their acquisitions by much larger con-cerns. Still, the United States remains uniquein the number of relatively small electronicsfirms that seek to compete worldwide, includ-.ing many of the new startups making semicon-ductor and computer products.

In Japan, the evidence suggests that the gov-ernment and banking system as in a numberof other countriesovertly discriminate amongborrowers on tilt: basis of size.° But even if for-i..;;n capital markets were identical in everyrespect to American marketb, and operatedwith no more government intervention than inthis country, the larger average size of the ma=jor foreign competitorsparticularly in semi-conductorswould give them at least a smallrelative advantage.

On the other hand, larger semiconductor andcomputer firms in the United States clearlyhave not reaped great benefits from their ownability to tap somewhat lower cost sources ofextern al canital. Thus. one can Question how

important such differences might be interna-tionally. Over the years, fast-growing and prof-itable small- and medium-sized firms have co-existed with the giants of the U.S. industryindeed have often Outstripped them; size anddiversity did not appear to give RCA (ir GEmuch help in computers or semiconduciurs.In dynamic, technologically advancing indus-tries, other competitive forces far outweighsmall differences in interest rates on bonds orbank loans.

Savings Rates10

As mentioned earlier, international differ-ences in savings rates could affect relative costsof capital. Within a closed economy, a high rateof savings creating an ample supply of invest-ment funds will tend to depress interest rates.Given the international linkages among capitalmarkets, this simple /argument is not by itselfsufficient to relate savings levels to interestrates, but may still have weight. For reasonsthat are poorly understood, the savings rate inJapan has been extraordinarily high for manyyears. Table 54 gives data on household sav-ings for the 1976-79 period; the figures for allfive countries have remained fairly constantover the past two decades. High savings rateshave characterized both the corporate andhousehold sectors in Japan, but it is personalsavings that have been most surprising in viewof interest rates that have been below prevail-

"For a general introduction to savings rates. see C. Elwell.Investment and Saving: The Requisites for E6onomic Growth.Congressional Research Service report No. 81-207 E. Nov. 15.1981.

Table 54.Household Savings Rates inSeveral Industrial Countries'

Average savings rate,1976-79a

United States 5.8%Japan 21.1West Germany 13.4France 13.5United Kingdom 8.3

Ch. 7 Financing: Its Role in Competitiveness in Electronics 265

ing rates of inflationas shown later (see table64) and have also been significantly lower forsavings accounts than for alternative invest-ments such as bonds." Despite this, Japanesehouseholds carry the largest portion Of theirsavings as cash or deposits; the contrast withbehavior in the United Statesillustrated intable 55is striking. Householders in Japankeep ai substantially lower portion of their as-sets in corporate stocks.

This extraordinarily high savings ratehalfagain as much as in France or West Germany,and nearly four times that in the United States(table 54)--when combined with closely con-trolled savings institutions, is often said to pro-duce artificially low interest rates on loans toJapanese businesses. The argument is essen-tially on the supply side: low interest yet abun-dant sources of capital have allowed Japanese

' corporations to expand rapidly while maintain-ing low prices, especially in export markets.

/ As one consequence of rapid growth, the com-/ panies would enjoy economies of scale, along

/ with modern, highly productiVe manufacturingfacilities. Past this point, low capital costswould be less of an advantage but Japan'sfirms could by then compete Comfortably onother grounds.

Numerous variations on this theme havebeen propounded many stemming from theJapanese Government's well-known low inter-

"See, for example. H. C. Wallich and M. I. Wallich. "Bank-ing and Finance," Asia's New Giant, H. Patrick and H. Rosov-sky (eds.) (Washington. D.C.: Brookings Institution. 19761. pp.260-26 I .

Table 55.Distribution of Household Assetsin the United States and Japan, 1978

United States Japan

Cash, plus demand and savingsdeposits 39.2% 68.7 %

Bonds 9.6 8.1Stocks, including mutual funds 23.5 10.0Life insurance 5.6° 12.6Othera 22.1 0.6

100.0% 100.0%_

est, high growth strategy in the earlier postwaryears. Some observers go so far as to imply thatno resource allocation problems exist in Japanbecause of a virtual glut in investment funds.12Often such assertions are linked with the targetindustry argument mentioned earlier. If true,this would mean that Japan's chosen industriesenjoy low financial costs for reasons entirelyapart from their high debt-equity ratios.

But capital markets should not be viewedfrom only one side. In tit, case, there are po-tential Impacts on the demand side as well asthe supply side. Growth affects both the de-mand for funds and the supply. Businessmenforesee abundant sale opportunities in ex-panding economies ar. ..,est to meet the newdemand. This places heavy pressures on capitalmarkets. On the supply side of these markets,individual consumers may experience rapidgrowth in real income but adjust their con-sumption habits more slowlymeanwhile sav-ing their unspent income. Thus, a case can bemade that Japan's high savings rate is a con-sequence rather than a cause of rapid econom-ic growthi-i.e., that income growth has out-stripped cnsumption.

In fact, neither demand-side nor supply-sideperspectives tell the whole story; both areneeded. din/ Japan,In/ Japan inflation-adjusted interestrates on /savings have recently been comparableto rate in the United Statestable 56. Thistable n rates of return available onlong-teirn government bonds in both countries(a rather/arbitrary choice) to the respective in-

ation rates, the difference being "real" rateof return. As the table shows, since 1S78 inves-tors inijapan have received higher real returnsthan,Viose in the United Sates. This suggeststhat lqrtificially depressed interest rates havenot recently been a source of abnormally low

'=Response of W. Rapp, Technology Trade, hearings, Commit-tee on Science and Technology, Committee ot, Interstate an'dForeign Commerce. and Subcommittee on Inteinational Trade,Investment and Monetary Policy of the Committee on Banking,Finance and Urban Affairs, House of Representatives, and HouseTask Force on Industrial Innovation. June 24. 25. 26. 1980. p.

266 International Competitiveness in Electronics--,--

Table 56.-Inflation-Adjusted Rates of Return in the United States and Japan

United States Japan

Longtermgovernment rate

Inflationrates

Real rateof return

Governmentbond rate

Inflationrate°

Real rateof return

1975 8.2% 9.2% -1.0% 9.2% 11.9% -2.7°A1976 7.9 5.8 2.1 8.7 9.3 -0.61977 7.7 6.5 1.2 7.3 8.1 --0.81978 8.5 7.5 1.0 6.1 3.8 2.31979 9.3 11.3 --2.0 7.7 3.6 4.11980 11.4 13.5 -2.1 9.2 8.0 1.21981 13.7 10.4 3.3 8.7 4.9 3.88Elased on consumer price Indexes.

SOURCE. Based on data from International Financial Statistics, International Monetary Fund, various issues.

costs of capital for Japanese electronics firms.In other words, there is little evidence of anyacross-the-board supply-side stimulus thatmight stem froth an ability by Japan's Govern-ment to persuade people to save even at rela-tively unattractive interest rates. As the tabledemonstrates, corporate (and recently govern-ment) demands for funds have driven up in-terest rates in Japan much as in other devel-oped economies.

This does not dispense with the possibilitythat the Japanese Government intervenes incapital markets to subsidize target industries.Certain industries-or firms--could be favoredby government repayment guarantees to lend-ing banks, effectively transferring the risk ofdefault from the commercial banking systemto the public at large. Alternatively, the Bankof Japan could be encouraged to rediscountbank loans at favorable rates. Finally, throughthe postal savings system the government itselftakes in about a third of all savings deposits:"These funds could be channeled to favored in-dustries at interest rates largely determined bygovernment fiat.

It is quite true that in early post-World WarII Japan, allocations of investment funds weremore a function of administrative control thanrelative interest rates; favored industries hadaccess to funds at subsidized rates of interest,while personal savers and small businessesbore the brunt of the costs. This point is takensin infor r-vc,ormii ofrIlrflIrcs r1f tb. hna-

nese financial system is described in more de-tail. Still, this practice seems largely to havedisappeared; government capital allocations donot now seem to provide borrowers in Japanwith a notable edge over U.S. competitors. Gov-ernment financial institutions accounted forabout 30 percent of all corporate loans in 1 -1502but as of the end.of 1980 their share. had fallento 13 percent; during the 1970's, loans fromgovernment institutions accounted for onlyabout 5 percent of total capital flowing intoJapanese industry." The percentage is evenlower in the electrical machinery sector, whichincludes electronics; here, government institu-tions accounted for only 8.2 percent of out-standing loans as of December 1.980. Nonethe-less, some observers continue to hold that theJapanese Government effectively socializes the,risk of corporate borrowing.15

Costs of Capital for Electronics Firmsin the United States and

Japan: Summary

It does seem clear that Japanese electronicsmanufacturers can get external capital at some-what lower rates than their counterparts in the-United States. But at present, this capital costadvantage, in inflation-adjusted terms, appears

"E. Lincoln, "The Japanese Government's Role in BusinessFinancing,"JEI Report, Japan Economic Institute. Washington.D.C.. Jan. 8. 1982, p. 12.

157 Catratrihara P Pc.lrirnan and V Pararla ThP la na mPcp


/4410

Photo credit: GCA Corp.

Wafer processing equipment for making.integrated circuits

to be small- -certainly less than 5 percentagepoints. The sources of this advantage are multi-ple: government policies in Japan that have theeffect of subsidizing interest rates for favoredinvestments no doubt continue to account fora good deal of the margin. Except for .the tax-shielding efiects of their higher leverage, Japa-nese electronics companies do not have accessto cheaper capital because of the preferencefor bank loans in their capital structures. Whilegreater leverage transfers business risks to thebanking system, it does not allow the Japaneseto avoid risks.

A differenc:, in financing costs of 2 or 3percentage points is nontrivial but will notdrastically alter the market postures of com-peting firms. For purposes of comparison, as-ellTeStl n ,ninn -.r

age point margin applies for the total invest-ment in a production facilitywhich is unlike-ly. Even then, the result-would be a pdtentialmanufacturing cost difference of about 11/2 per-cent, and might translate into a price differenceof the same magnitude. Smaller capital cost dif-ferences would reduce this advantage com-mensurately.

Such a 2 or 3 percentage point advantage incapital costs represents an average over manyfirms in both Japan and the United States. Thedifference in average costs of capital in the twocountries is smaller than the differences incapital costs among competing electronicsfirms within either country. Although table 51does not accurately portray cost of capital dif-ferences between the two countriesty will il-lustrate this point if taken as representative offirm-to-firm differentials. The range in costs ofcapital for the eight U.S. semiconductor firmslisted in the table is 7 percentage points, thatfor the six Japanese manufacturers nearly .10percentage points. The interfirm differences--and the resulting competitive advantages or-handicapsare much larger than OTA's esti-mate of the average differential between thetwo countries.

Relative availability of capital for electronicsfirms in the United States and Japan has great-er potential impact. Favored Japanese electron-ics firms seem to have little difficulty in acquir-ing funds for expansion and modernization. Incontrast, some U.S. companiesparticularlythe smaller onesbelieve themselves subjectto capital constraints. That is, they may findthemselves unable to raise as much capital asthey would like at any reasonable cost.

Capital availability is a subject left to a latersection, but note one major difference betweenelectronics and other American industries thatmake this same complaint. Some domeitic steelcompanies, for instance, have had difficultytracting external funding because of their Lability to convince prospective investors of theindustry's potential for growth and future prof-_._ Tit rl -1


capitalparticularly compared to existing networthrequired to maintain their share in arapidly expanding worldwide market. In semi-conductor manufacturing, rising capital inten-sity compounds the difficulty. These matters,concerned more with the dynamics of growth

than with absolute costs of capital, are takenup below. The next section extends the com-parative treatment of financing to several othercountries, while carrying forward the U.S. -Japan comparison.

Financial Structure: An International ComparisonMany countries are attempting 'to build com-

petitive electronics industries because they be-lieve them essential for a growing and healthyeconomy. Government assistance, often finan-cial, has flowed to electfonics companies:Great Britain provides explicit subsidies;France has combined subsidies with trade pro-tection. A number of the rapidly developingcountries are following suit, as outlined inchapter 10. Still, Japan remains the primarycompetitor in electronics, and its financial sys-tem is treated in much more detail than ihatof France or West- Germany, the two' othercountries examined below. The focus on semi-conductors continues because U.S. firms inthis oart of the industry have fae6d the mostpronounced financing problems.

Funding rapid expansion is a challenge thatsemiconductor companies share with manu-facturers of small computers and peripherals,software firms, and new entrants in other por-tions of the industry; Atari, the manufacturerof video games, has reputedly been the fastestgrowing technology-based company in history,while one firm making game cartridges saw itssales grow 1.000 percent (to $50 million) dur-ing 1981.16 In order to remain competitive,firms in such markets must be able to financegrowth at _rates that are literally explosive.

United States

Sources of financing for American electron-ics companies vary depending largely on their

New corporate startups have been frequentduring the industry's postwar histo-ynot onlyin semiconductors and computers, but in many,other product lines. Hewlett-Packarda div9r,,sified manufacturer of test and laboratorYequipment, calculators and computers, and aleader in integrated circuit. (IC) technologythrough its captive operationsgot its start justbefore the war in a garage in Palo Alto, Calif.17In many respects typical of the firms that latergave this region the name Silicon Valley, thecompany's founders began by designing itsfirst product themselvesan audio oscillatorsupplied to Walt Disney Studios for the pro-duction Fantaia.

Venture CapitalBusinesses typically draw on far different

sources of funds in their early stages of devel-opment than later, progressing through a fairlypredictable sequence as they grow and mature.This progression, which illuminates some ofthe unique characteristics of U.S. capital mar-

, kets, is rather different from that in other coun-tries. For purposes of illustration, consider astartup firm with origins like those of Hewlett-Packard or the many semiconductor manufac-turers that sprang up during the 1960's. Oftenthese enterprises were formed by small groupsof engineers and managers spinning off froma somewhat older company with the aid offunds from the venture capital market. Theprocess is not unique te.., the semiconductor in-dustry: Control Data Corp. was founded in1958 by a grutip of ex-Univac. employees. While

Ch. 7Financing: ifs Role in Competitiveness in Electronics 269

many new ventures in microelectronics, com-puter peripherals, and software have been es-tablished since 1980.18

Where rapid. growth creates expectations ofhigh returns, capitalization for new companiesoften comes from specialized financiers whoprovide equity funds to the venture capital mar-ket. Along with electronics, biotechnology isa current example. In such cases, ownershipis typically shared among venture capital orga-nizations and the founders of the firm. Stockoptions have been a common means of attract-ing talented individuals to startups, and havebeen used to build handsome 'compensationpackages for key executives or technical spe-cialists without cutting too decply into cashflow.

Annual returns of 25 to 50 percent over aperiod of perhaps 5 years are typical goals ofinstitutional venture capital organizations. Inthe past, wealthy individuals of families some-times founded private corporations for seek-ing new and riskybut potentially highly prof-itableinvestments. Today, corporate venturecapital organizations are also prominentsubsidiaries of larger companies seeking to di-versify. Corporate venture funding is more like-ly to go toward second- cr third-round financ-ing of young companies than to new startups,and investments tend to be larger than thoSeof independent venture capital organizations.Sometimes eventual ownership is an objective;in other cases corporate venture capitalists aresimply seeking capital appreciation. Occasion-ally the funding organization provides capitallargely to gain proprietary technology. This hasbeen the apparent goal of investments in U.S.electronics firms by a number of foreign com-panies. Both Siemens (West Germany) and Fu-jitsu (Japan) have invested in this way. Siemensowns 20 percent of Advanced Micro Devizies,Fujitsu 26 percent of Amdahl, a leader in tech-nology for large computers.

An alternative source of venture funding, theSmall P.:sinacc Invactmant rn (SRIC.1 wag rra-

ated by the Small Business Investment Act of1958. Although. most SBICs concentrate onneighborhood businesses, a few have nationaloutlooks. Venture capital partnerships andfunds have also blossomed in recent years; theU.S. venture capital industry now includesabout 600 firms and should continue to expandas a consequence of new ERISA (Employee Re-tirement Income Security Act of 1974) rulesallowing "prudent" participation of pensionfunds in venture activities." In rare circum-stances funds, be raised through publicstock offerings, but this avenue is more likelyto be ailable later in development.

Verdure capital markets are highly cyclical.One influence has been taxation of capitalgains. In general, high taxes on capital gainsdiscourage potential investors. Table 57 sum-marizes the results of a study prepared for theNational Venture Capital Association, togetherwith more recent data, that bears on this point.The maximum tax on capital gains in theUnited States was reduced from 49 percent toless than 30 percent in 1978. Although totalventure investments rose dramatically, suchtrends do not prove a cause-and-effect relation-ship. They do suggest that the tax revision wasa powerful contributing fal:Aor in the upswing.

At the same time, a variety of other forcesaffect the ups and downs of venture funding.The cyclicality reflects the confidence of poten-tial investors on the supply side and of poten-tial entrepreneurs on the demand side concern-ing prospects for the economy and the propi-tiousness of risky new undertakings. The tim-ing of startups depends on more than economicconditions. Venture organizations look care-fully at the abilities. of a new firm's leaders;both capitalists and managers look for "tech-nological -windows" that offer unusual oppor-

190ne estimate of the Ind ictnt hrnalcrinwn is ac fnlIntvc. nriva to


Table 57.Aggregate U.S. Venture-investment Activity (millions of dollars)

New commitmentsHigher round investments

(prior commitments) Totals1977Amount $56 $28 $84Number of investments 126 126 .252

1978'Amount $92 $31 $123Number of investments 196 150 346

1979'Amount -145 $65 $210Number of investments 29 248 538

1980Amount NA NA

1981Amount $500 $900 $1,400NA - not available.1379 ::.ota annualized from first 6 months. Data taken from 55 respondents. To OTA's knowieige, no fulty comparable data

venture Investments for later years we available.SOURCES 1977-79"Financial Issues in the Competitiveness of the U.S. Electronics Industry," report prepared for OTA by

L. W. Bergman & co. under contract No. 033-1660.0, p. 9, quoting from "Survey of Ventu:e Capital Investment,1977.1979," prepared for the National Venture Capital Association by D. J. Brophy and P. L. Schaefer o1 the Univesity of MIchigin. 1980, -Government.industty Cooperation Can Enhance the Venture Capital Process,GAO/AFMD-82-35 (Washington, D.C.: Genert,1 Accounting Office, Aug. 12, 1982). p. 3. 1981J. W. Dizard. "Do WeHave Too Many Venture Capitalists? Fortune, Oct. 4, 1982, p. 106.

tunities. Some of the upsurge in investmentsin table 57 is related to booming interest in ap-plications of microprocessors; of the new ven-ture capital deals nationally since 1979, perbaps half have been in electronics or closelyrelated faelds.20 At the peak of the most recentcyclei.e., rnid-1981---many observbrs of ven-ture capital. markets concluded that-entrepre-neurs wue having an easy time finding start,:up fonds; some claimed that the supply of ven-ture capital considerably exceeded demandduring 1981., and that poor risks were beingfinanced. 21 In more normal times, potentialstartups may face a long and arduous searchfor Capital.

Table 57 will serve to illustrate another point:venture capital makes only a small contribu-tion to the over-all funding needs of Americanindustry. Annual venture financing at some-thing over $1 billion, and a total pool of yen-.

IFsues in the Competitiveness of the U.S. Elec.indost,T," report prepared for OTA by L. W. Borgmrin.

& Co. wirier contract No. 033-1550.0, p. "Startup Fever isSpreading," op. cit.

"Ser:i, for example, A. Follzick, "Few Places for Venture Capital:Funds Outpace Investment Opportunities," New York Times,June 17, 1081. p. Dl; J. Levine. "Once Again, It's A Buyer's Mar-ket," Venture, June 1'482, p. 80.

ture capital of pbrhaps $5 billion to $6 billion,pales alongside other sources of capital for U.S.business and industry: bank loans, $230 billion;other short-term debt, $250 billion; corporatebonds, $490 billion; commercial mortgages,$280 billion; equities, $1.3 trillion.22

Costs of EntryAlthough a substanti it fraction of new ven-

ture investments continue to flow into electron-ics, in some segments of the industry entrycosts are becoming prohibitive. Among the ex-ceptions are firms able to generate cash flowsin other lines of business. NCR is an example:an established manufacturer of computers andother business equipment, the firm had madeICs exclt:ilrtly for internal consumption. In1981 NCR announced plans to produce semi-custom logic circuits and certain kinds ofmemory chips to be sold on the outside, becom-ing one of the few new entrants in the mer-chant market contemplating a fairly broadproduct line.23 The reason is straightforward.Costs for establishing a new seinicbnductor

22"The Perilous Hunt for Financing," op. cit. The estimatesare totals outstanding at the end of 1981.

2"A Surprise NCR Leap Into the Chip Market," BusinessWeek July 13, 1031. p. 22. ..

2

Ch. 7Finarr:ing: Its Role in Competitiveness in Electronics 271

firm have risen by a factor of 10 over the pastdecade; as figure 50 indicates, the end is notin sight. The rapid increases in capital inten-sity shown in the figure stem largely from themore expensive prodtiction equipment re-quired for current-generation ICs (ch. 3, espe-cially table 2).

Rising entry costs are one reason why the1980-83 group of semiconductor startups havepicked narrow market niches rather than at-tempting to compete in a broad range of prod-ucts. Examples include: custom chips, or insome cases just design services; specializeddevice families .such as linear ICs or program-mable logic arrays; and, in one case, galliumarsenide circuits. While entry via niche mar-kets is the usual pattern in this ancilAher indus-tries, none of the semiconductor startups ap-pear to be aiming at the mass produced mem-ory or microp"ocessor markets: the most recentnew entrant to attempt this was Inroos, estab-lished in 1978 with the aid of $90 million indirect investment by the British Government.24

=' UK 13oard to Cut Stdko in inmos," Electronic News, Mar.29, 1982. p. 54. The National Enterprise Board, which has had-considerable autonomy to fund British industry (ch. 10), provided50 million pounds. with Limos receiving an equal amount inloan. guarantees and grants for specific projects.

50.Increase in Capital Costs forHigh-Vailmelc.tegrated Circuit Production LineSi r---bdi,

$1002 million

S10million

1970 1975 , 1980

Year

1985 1990

SOURCE R W. Broderson. "Signal Processing Using MOS.VLS itichr.olc.gy,"VLSI Electronics: Microstructure Science. vol. 2. N. Einspruch(ed.) (Ne?. York: Academic Press. 1981), p. 203.

Entry barriers can be higher still in main-frame computers, where the new firms in re-cent years have entered with plug-compatiblemachinesAmdahl, Magnuson, in 1981 Tril-ogy. The one exception during the 1970's wasCray, established with venture funding to buildspecialized supercomputers. An added hurdlestems from the preference by many customersfor leasing rather than purchasing large com-puters. Financing leases is a severe strain onsmaller companies; lease cancellations wereone of the immediate reasons that Itel, a con-glomerate that had entered the plug-compatiblemainframe business, declared bankruptcy in1981.

Leasing has been a major part of IBM's strat-egy in mainframes; competing firmsnone ofwhich has assets near IBM'sface a major con-straint in financing leases.25 Not only are theylimited by their smaller size in securing fundsat rates comparable to IBM's cost of capita,but the risks of competing with IBM are largeand v " L-nownitel's failure presumably add-ing to the concerns of potential lenders. With-out a source of particular advantage such asAmdahl or Cray get from their reputations as'technological leaders, cost and-availability ofcapital will remain formidable barriers for en-try into the mainframe computer market.

Entry costs are also daunting at the high per-formance end of the minicomputer industry,though firms such as Prime (1972) and Tandem(1974) did :begin operations during the pastdecade. The microcomputer segment has stillseemed attractive; the entry of IBM into thepersonal computer-market at the end of 1981has not seemed to dampen the enthusiasm ofprospective entrepreneurs and venture capi,talists.

Early Growth

In the startup stage, equity capital from ven-ture or other sources goes to purchase or lease'plant and equipment and cover the initial ex-penses of developing, manufacturing, and mar-keting the first products. Later, more familiar

J. T. Soina, The.Compnter Industry (Lexington, Mass.: Lex-ington Books; 1976), p. 4L


financial markets can be tapped. Exterrel fi-naming is ci cal at this stage: the Companymay be growing rapidly, with production out-stripping sales as inventories build and distri-bution channels are filled. New firms oftenoperate below their break-even points for anumber of years, and cash flow problems arecommon.

Once sales have begun, local banks Ivill nor-mally provide short-term loans up to about 80percent of net receivables, this amount beingrolled overi.e.. the loans _rewritten at prevail-ing interest ratesevery 3 months or so. Long-er term financing may come from incrementalventure capital commitments; many ventuyeorganizations prefer to participate in second-or third-round financing because they can bet-ter evaluate a company's prospects once it hasproducts to show. While at this stage litnitedpublic offerings may also be possible, stocksales to the general public have been less com-mon in electronics than in other industries.Many electronics manufacturers have beenable to finance quite rapid growth through -re-tained eavnings and employee stock option andpurchase plans: Indeed, the managers of elec-tronics companies started by individual entre-preneurs or small groups have often shunnedexternal equity markets to avoid stock dilutionand the loss of close control.

When sale reach annual levels in the :vicinityof $10 mi .. credit lines typically ecomemore regi.... ..zed. Revolving credit a d termloans provide short-term financing. In addition,banks will generally extend lease credit forcapital equipmentparticularly helpful in elec-tronics because it reduce: the pressure to raisefunds for purchasing equ:,:nent ate time whenlong-term investment requirements are ex-panding rapidly. Because capital equipmentCan quickly become obsolete, staying at theforefront of the technology can strain re-sources. On the other hand, for firms with ade-quate cash flows, rapid obsolescence meansshort writeoff cycles and tax savings from:depreciation.

In any case, firms with growth patterns thattake them beyond the $10 million level findcapital-more abundant-and-less-costly. At-this

point, electronics companies exhibit financingpatterns that depend on the preferences ofowners and managers, as well as opportunitiesin relevant capital markets. Some firms offernew equity shares to the public; others issueshares but limit purchases to their own execu-tives or employees. Some sell bonds, often inaddition to equity, to add leverage to the capitalstructure. But while financing patterns differ,they share a characteristic common to most ofU.S. industry: American electronics firmstypically attempt to finance expansion inter-nally, even when this delays dividends inde-finitely. Only if self-generated sources proveinadequate do companies enter externalmarkets.

Internal and External Sources of Funds

Table 58 illustrates the extent to which Amer-ican semiconductor firms rely on internalfundsi.e., depreciation and retained earnings.When the companies listed in table 58 haveresorted to external financing, this has rangedfrom virtually all capital stock (Intel) to substan-tial amounts of debt (National).

Although both new and established firms inthe U.S. electronics industry will no doubt wishto continue relying on internal funds, a numberof factors converging during the early 1980'sforetell financial dilemmas for some compa-nies. Among those common at least to manu-facturers of semiconductors are:

1. Growth in unit sales over the coming yearsray exceed even the .apid rates of the pre-:;ous decade.

2. Revenue growth will continue to trailgrowth in unit sales as manufacturingcosts and sales pricesIdecline. Historically,semiconductor prices are driven rapidlydownward as cosis fall because of learn-ing curve phenomena. The sharp drops inmemory chip prices during 1981when16K RAM (randorit .:ccess memory) pricesfell from $4 each to about $17-a1' e sympto-matic.

3. Capital intel -ity will continue to rise be-cause new pro: ..iction equipmente.g., forfine-line lithographyis much more ex-pensive-than-in-the-past-

Ch. 7Financing: ItS Role in Competitiveness in Electrorics 273

Table 58.Internally Generated Funds as aPercent of Total Capital From All Sources

Yea'xas

IriirunientsNational

Semiconductora IntelAdvanced

Micro Devices1974 L % 76% 89% 93%1976 79 82 79 961978 87 97 87 671980 65 60 46 . 71aFiscal year

SOURCES 1974-78 "Financiat Issues in the CompetitivAness of the U.S Electronics Industry," report prepared for OTA byL Vi. Borgman & Co. under contract No. 033-0550.0. p. 31. 1980 Annu al reports.

4. Engineering and design costs are also es-calating, due to the increasing complex-ity of ICs.

5. Global competition, particularly from theJapanese, is becoming more intense, andwill be based on low prices to an evengreater extent than in the past. Althoughforward pricing in anticipation of learn-ing economies has been characteristiceven of competition among domesticfirms, pressures from the rapidly expand-ing Japanese industry may cut still fur-thur into sales revenues.

6. Competition is also forcing companies topay more attention to quality and reliabil-ity, requiring costly test equipment. As ICdesigns increase in complexity, testingprocedures become more time-consumingand expensive.

Figure 51, comparing capital spending ratesin the United States and japan over the pastfew years, illustrates the rise-in capital inten-sity. Capital spending in both countries fellsharply in 1975 when the market for semicon-ductors slumped; since that time, the U.S. treadhas been steadily upward. Japanese spendingrates have been higher because they have beenadding capacity faster.

While capital spending in the United Statesaveraged around 10 percent of sales during the1970's; rates in the last 3 years have beensignificantly great: (fig. 51). Continued in-crease will be difficult for many merchantfirms without substantial outside funding.Capital needs of U.S. semiconductor manu-facturers during the current decade will prob-ably be in the range of $30 billion, with some

industry sources expecting considerably higherfigures. Such estimates compare with capitalexpenditures totaling $4.5 billion during the1970's.26

The changing character of semiconductorproduction and marketing is not unique.Smaller American manufacturers of comput-ers, as well as peripherals, are confrontingmore intense foreign competition at a timewhen developing technologies are placingsevere demands on their financing capabilities:New firms are entering markets for peripheralssuch as terminals and disk drives designed tobe compatible with the products of establishedcompanies. Microcomputer applications of allsorts are on the rise. Computer software is oneof the most rapidly growing portions of the in-dustry. New entrants have often depended onventure capital for their original financing,andagain like semiconductor manufactur-ersfollowed growth patterns relying on inter-_nally generated funds supplemented with lim-ited amounts of debt. ,

Parallels exist even in consumer electronics.A good deal of the production of establishedproductsradios, televisions, audio equip-menthas moved abroad, taken by Far Eastern

2°''Hungry for Capital-to Sustain the. Boom," Business Week.June 1. 1981. p. 74. J. F. r y of Texas Instruments has estimatedspending at $25 billion.to .-Z,15 billion for the decade of the 1980's,while G. Moore of Intel v..,,ices the figure much higher, in therange of $65 billton. An estimate of $30 bfllion results if salesare projected through 1990 based on the trend since 1975 (ch.4), with capital spending assumed to remain at 13.7 percent ofsales, the average over the past 3 years. Spending rates mayrisesome predi6tions point to 15 percent of sales in comingyears (see E. Williams. "Electronic Compo:::ents," FinancialTimes. Mar. 5, 1982, sec. III, p..1)but the total is much moresensitive to the sales projection titan to the capital sperciinc rate._

276

Cornce:Itpieness in Electronics

Figure 51.Rates of Capital Spending by U.S. and Japanese Semiconductor Firms

t Japanese averages

......... ....... .... ,./ s

-,r..

d''

.197 1975 197fJ

1977

Year

1978

ndnut.. 1`+73-137'4 wniqritml ,"144eraye e.! 12 manutac!ur,r5 19731979 11 Manutanturnrs 1980 981

t4480 '7444,3t44,1 ter 11 U S mirdtant serreCnndUCtor manutantitrers. t,:j11 estimated.

%or

1979 1980 198

'-i United States 1y 1377 E3uau 01 Census, 1978-1981, Departrri4)411 of Corrirnerce4Ind Semiconductor Industry A' lion Japan 1973 1979..L.IPJ, Fall)1..,8').). 0)111.:11 PUt)811)))1,081i, 1980) 1; 203. 1980, 1981. Japan ECOI70:771C Journal

competitors or transferred offshore by U.S.co-rtirtanies. But the broader market for coo-sumer electronics remains dynamic, as the ex-ample of video games showed. Along with vid-eo cassette and disk players, .as well as -homecomputers. these are precursors of an array ofelectronics-based innovations for personal andhome use to be introduced over the next twodecades. Many of these products will be hightechnology items made possible by advdnceselsewhere in the industry. Some may comefrom small companies or-ganized by entrepre-neurs with strong technical backgrounds, ashas happened with semiconductors, micro-.computers, and software. Unquestionably,

______fierce_competit ion. from abroad will continue.

in consumer products; financing could becomea problem here too if small firms need to :;up-port rapid expansion.

Financing Startups and Growthin 0.:3 Coming Decade

But why should future financing be trouble-se73e for an industry whichby all indications

can expect expanding markets overseas aswell as at home? Not all. observers believe theproblems will be that-great; those who do gen-erally point to a pair of related concerns:.

1. it may not.be possible to finance growthfrom internal sources in the proportionscommon in-eariiern.years, thus requiring

Ch. 7 Financing: Its Role in Competitiveness :7 Eitctronics 275

greater dependence on external sources ofcapital.Costs" of external funds will he high, andmay place U.S. firms at a disadvantagecompared to competitors in Japan, Theprofit levels required may be difficult forAmerican firms to reach.

The first point deals with the continuing abil-ity of electronics companies to fund growth in-ternally. To expand sales, manufacturers musteither purchase assets that are more produc-.tive or use existing assets more effet :tively. Tosupplementassets, funds must come from oneor more of the following sources: moneys accu-in LI 1;1! ed through co.porate operations (retainedea ii;gs plus depreciation); capital generatedby lee sale of additional stock; 07 borroWingsof one type or mother.

As discussed prie4iously, borrowing w'ehoutparallel increases in equity alters a fiee-i'sleverage. This, in turn, tends to increase thevariability of rettfrns to equity, increasing therisks to owners/. In countries like the UnitedStates, where capital markets are relativelycompetitive, managements choosing higher le-verage eventually confront two problems: 1) com-mon stockholders become increasingly sensitiveto their risk positions, and make adjustmentsthat tend to depress stuck prices; and 2) lendersalso may:'object, ultimately refusing to supplyadditional debt. As capital intensity increases,financial managers face an intricate set of de-cisions;

The important relationships can be ex-1-:,,Atssi!fl its ..oilows:

I,J1(. . .

0;

Value of Tot-1 valuesides t-f Net profit

Total vt:Itte Value ui ecitta,.. Va!ott of(-1 asset,:

17his expression is simply an identity. Thefirstferin on the right-hand side, asset turnover, isa broad measure of Asset productivity. Definedas I. A; ratio of sales to total assets, it in,',icates

/ the efficiency with whic:la company utilizesits assets to .i/enerate revenues, and deper-de-on-such -fa,evee-eas_thelirmls.capitai intensity.and the degree of competition cha.r:Y*......,ng

its markets. The second termthe ratio of as-sets to equityis one way of measuring ti firm'sleverage, an alternative to debt-equity ratio.Profitability, the third, depends on many fac-tors: product mix, competition, and labor pro-ductivity, to mention just a few.

Industry analysts tend to focus on the assetturnover ratio, which may fall as a conse-quence of expensive capital equipment (thisshould, however, improve productivity). Thusto preserve existing returns to equity, firms willeither have to adopt higher leverage ratios orsomehow improve profits on sales. Given theconstraints likely to be exercised by the U.S.r'nancial community, there are clear bounda-, les to the first optioni.e., to increasing thereletive proportions of debt in a firm's 'capital-ization. The second possibilitygreater profit-abilitywill be equally problematic if interna-tional and domestic competition continues tobe stiff.

There is a second difficulty. In the identityabove, note that if each term were to be heldconstant, growth would have to involve a bal-anced expansion of .lebt and equity. But the!eft-hand side of the equation dictates that, ifa given increase in equity is to be financed in-ternallyi.e., through retained earningstheremust be a proportionate increase in aftertaxreturns to equity (profits). (This assumes thatelectronics companies use aftertax earnings tofund growth instead of paying dividends, a pol-icy followed by most of the rapidly growingcompanies during the 1970's.) Thus, growthmust be accompanied by an increase in profits.

There is no requirement that new capital berestricted to funds generated internally. Firmscould tap markets for both equity and addition.al debt. /NOnagements, however, often objectto stock offerings on the grounds that new is-sues are expensive." Aside from the costs offloating the offering, before the current boomexecutives commonly cited what they per-ceived as low price-earnings (P--E) ratios in se-curitY markets. From the perspective of man-agenient, selling new stock-under such condi-tionS would have provided too little capitalcompared to their expectations of futuregrowth, earnings, and presumably dividend


payments. P-E ratios of 2 or ;:t for electronicsstocks were cited as indications :if "unrealis-tic:" rate of return requirementsas implyingthat the market vas demanding returns in therange of 33 to 50 percent. New equity issues.in this view, should be delayed until the marketreturned to more normal co7ditionsi.e., un-til stock prices and P-E ratios rise.

Whether or not this makes sense deperAs inpart on how stockholders are perceive. Tothose who believe that equity holders are, 1ifact, owners of the corporationand if existingstockholders have first rights in the purchaseof new issuesthen stock prices may appeartoo low to management but the firm's "loss"is exactly offset by the "gains" of these stock-holders. That is. existing owners would be ableto buy new stoc... "cheaply." The existing own-ers would be unaffected by variations in theissue price based on market conditions. On theother hand, equity may be viewed as effectivelyanother form of subordinated debtin realityseparated from any control. In this case, man-.agers would perceive equity as "borrowing"and its costs wo,ld be evaluated like any otherdebt in cost of capital calculations. Statementson the high cost of equity by managements ofelectronics firms oaen seem to point towardthis second belie

Then to the seconc: pond above: Will exter-nal funding be available at a cost manufactur-ers are willing and ablein the face of compet-itive pressty-s--tu : the moment. theavailability u( fur.3 does not appear to bea limiting factor. c, could change as firmsincrease their leve Lending institutionstypically establish standards on levels of debtconsidered prudentguidelines that depend onliquidity and the variability of cash flows to theborrowing firm. Effective limits .on debt. there-fore, differ front co.npany to company. Ever.if some companies can borrow what ,hey named,the total volume of funds required by the U.S..electronics industry seems so higil that otherwill almost certainly reed new equity.

For similar argumanis applied to the Ja.,.)aneira case iceWallich and Wallich., op. , it., pp. 268-269.

While equity from venture capital sourceshas been more freely available since the 1978revision of capital gains tax rates, most of thefirms needing capital will be well beyond thisstage. Nor is it likely that such sources couldprovide enough money; the venture capitalmarket is far too small. Still, there remain por-tions of the electronics industry where initialcapital requirements are less than in semicon-ductor manufacturinge.g., computer 'soft-

s!-..e.Cialized industrial equipmentandwhere startups should find capital relativelyabundant. Considering the effective reduc!i Insin corporate income taxes resulting from theEconomu_ Recovery Tax Act of 1981, the totalventure carital pool should continue to ex-pand.

Even so, many observers predict that the costof funds will be too high. From this perspec-tive, investments may simply not promise ade-quate rewards; American companies in partsof the industry that face mounting competition

Jrn abr--2,ad may have difficulty in attractingfunds from wary lenders with numerous alter-natives, some offering better prospects for safe-ty and high returns. Finally, some commenta-tors believe that the total supply of investmentcapital in tne United States is smaller than itcould or should be because of a variety of dis-incentives affecting savings and investmentbuilt into the American tax systein.29 While thecapital market will certainly supply funds inan amount equal to total demand at some setof interest rates, such observers believe thatsupply constraints artificially force these ratesto levels too high compared to other countries.

Several of these matters are at the heart ofeconomic policy debates still before Congress;the capital formation question, in particular,has been widely discussed for years, and hasnot been resolved by the tax policy changes in-stituted in 1981.29 The issues are often complexand teclsical. As a consequence, the discus-sion below concentrates on matters that areparticularly relevent to :)ctronics:

"For a typical commentary, see A. Sloan and C. Miles, Slow-down at Capital Gap," Forbes, Jan.;. 1980, p. 38,

2i lee, for wimple, Capital Formation, hearings, Joint Econotr.is Committee, Congress of the States, lone 9, 1976.


Capital Supplies and Financing Costsfor the U.S. Electronic.:, Industry

As mentioned previously. American elec-tronics firmswith some prominent excep-tions tend to be smaller than their major in-ternational competitors. And, in part beca:isetheir lack of diversification results in sharper-swings in cash flow, small companies confronthigher financing costs than large integratedmanufacturers. beret ire; on the basis of com-pany size alone, electronics producers arelikely to face higher costs of capital than manyof their competi.ers in Western Europe or Ja-pan.

The semiconductor industry again providesa ready example. Table 59 lists total assets fora sample of U.S. and Japanese companies.While Japan's semiconductor manufacturers

Table 59.Assets of U.S. and JapaneseSemiconductor Manufacturers

U.S. merchant firms

Total(millions

1979

assetsof dollars)

1980

Advanced Micro Devices $109 $165American Microsystems 80 81aIntel 500 767ntersil . 83 98'

161Motorr)r) .......... 1,904 2,112National Sernici..,nduct,)r 385 561Texas Instrumems 1.908 2.414U.S. captive producersIBM . ... S24.530 $26,703Western Electric' 7,126 3.048Japn.-wse Companies'Fujitsu $2,030 $2,380Hitachi 6,790 7,450Matsushita Elec:tric 5,190 5,640Mitsubishi Electric 4,490 4,910Nippon Electric Co. (NEC) 3.110 3,560Toshiba 6,180 6,450AC7un r1 in early 15'12 t'y GcAlir ,.rich nail 1980 assets of Al 6 talhon

LA:. of Sebteatser 1980. in early t Inter sa purcha4,ed 5r General Elect,", 1:090 a: nets acre. T.18 5 billion

"-%-r-tn,ea a fy; s7 7 LOH:,l-,:',.'n=tern Elector: only ,0 ,, not Include assets of `3.T. cperating

,asortns oche; A I T stAbsi,.;,tts"Asset ''Gore; do not include iatillata5 i..anyaraions forn yen to cIC.asas based

rotes trot Eronomio .Reoorf of t,lo Pf:sidort (Wasilinoon, t) C,0Q!, r, 145 1979, 218 yen per del,ar, 980. 226 pen per dollar

,;0151,; . IS, firms - it r.n1.1a1 -e 5 li.pank.,e firmsJana Handf-0 cpit yo Toyo Kel:a Ft.;-://Ii),.;t. 1579,,

are substantially larger than the typicalAmerican merchant suppliers, several of theU.S. firms already have been acquired by muchbigger companies. Intersil may look puny com-pared to Hitachi ce NEC, but this is hardly trueof its new owner, GE. It is quite possible thatfurther rationalizetions of this type will con-th e, perhaps in part to assure continued fund-ing for expansicu. FAthough many of the cur-rently attractive candidates for purchase havenow been acquired. Finally, as table 59 alsonotes, the two largest captive manufacturersin the United States, IBM and Western Elec-trc, have assets much larger than the Japaneseproducers.

For U.S. companies that are riot affiliated-vith larger firms, the question of differentialfunding costs remains. Based on the usualassumptions concerni i-tident amounts ofleverage as discussed ea.,::.er, the difference indebt financing costs bie.ween a firm the sizeof Intel (table 59) and companies like Motorolaor Texas Instruments probably averages about1.2 percentf,,c points (ir' fact, as table 51lustrateS, the firm-to-firm differences at anypoint in time clearly depend on many factorsbeyond size). _Ci)=11 else equal, Japanese firmsmaking semiconductors would probably haveabout the same advantage as a result of theirsize and diversity. But, because debt accountsfor only a fraction of a company's capitaliza-tion, the impact of this difference in interestrates is smaller, and other forces are likely toweigh heavier in the competiCve balance. Afterall, the :Xmerican firms included in ta'''e: 59were f.,rmed, grew, and flourished even thoughtheir capital costs were greater than such -.em-petitors as RCA, GE, and GTE, all of 1.vhor. en-titre(' the semico..ductor market in its earlyyears.

The related questionwhether interest ratesin the Un"d States have been driven to ex-cessive lei by supply constraintsi.i too invalved to cover in any detail. It is true, for ex-ample, that in ;erest rates to carp-07a o bor-rowers are lower in Japan for equivalent pro-portions of debt (to the extent that equivalencecan be ascertaine(.l). But such comparisons,

on diff,:rent currences, con easily !ME-

2 8 2

278 International Competitiveness in. Electronics

lead. In ;)art. they reflect differing inflationaryexpec:at ions as In.ircored in interest rates. Ad-justments for inflationarN,, expectations. areproilematical; to some extent, it might be possi-ble to make such adjushnents based on patternsof variation in the exchange rate. But a varie-ty of factors other than inflationary expecta-tions affect foreign exchange markets. especial-ly in the short run, and the gyrations of the yenagainst the dollar in recent years have gener-ated a wholly independent source of contro-versy."''

The inflationary trends reflected in high U.S.interest rates can themselves deter new invest-ment, in electronics and in other sectors of theeconoiny. Inflation adds to uncertainty in thecash flows that can be anticipated from in-vestments; these mount as investment horizonsrecede. High rates of price inflation tend todiscourage longer term commitments; instead.they/ favor investments with relatively quickpayoffs. American managersin electronics asin other indo trieshave been charged with ig-noring long-term growth opportunities whileconcentrating on the short run. Part of thishesitancy to commit resources is tied to uncer-tainties created by price inflation and the at-t impact on interest rates. That is, dif-

hces in short-term compared to long-termmanagerial behavior between American andJapanese corporations may not be caused .by&fierences in capital availability, or byferences ill real interest rates, so much as byuncertainties associated with high and variablerates of interescand inflation.

"The Unite Jtates has not"had 0 great dealof succw;.-; ill controlling inflation over the pastfew :years; while current economic policiesMay help to 1, et'!) down the inflation rate, ad-r stinerits jr xpectations often lag well behind,(:an the eh; conics industry expect to benefitmoire directly fruit) the changes in U.S. tax lawadopted in 1981? A Per all, Irrniy of these' ".tioo, for I'. Ilarilandl'Ilimborg "V,111! of vt Ftivls

;;;,1, Exrlintige }tats, Nt't ,inports aBetio' 1.os Anger/es 111, .iimkestnen forAim.rican (41..0 :6nm competitive problems im an....ndervalo,,d von mai, ng 1;,i)a..u.st, imports cheap in the 1,;.5.mirket.

changescuts in personal income taxes as wellas effective reductions in corporate taxeswere directed at enlarging the supply of fundsfor investment. Unfortunately, increasingcapital supplieswhich all else equal willdecrease the costs of investmenttends to bemore easily said than done.3' For instance, taxchanges that affect both supply and demandfor funds may leave interest rates unchanged.Under this circumstance, neither the availabili-ty nor the cost of funds for Amer' in elec-tronics firms would change.

In part because of these complicating factors,it is not yet clear what the net effects of theEconomic Recovery Tax Act of 1981 on aggre-gate capital formation will he. much less thedifferential nnpacts on various sectors of theeconomy. Thus far, there is scant evidence ofbroad positive effects on capital investment inindustry; many observers are skepti,-tal that therevisions to U.S. tax law will have niuch effecton levels of personal savings, which they feelare central to freeing up new investment forindustry.32 Internationally, even before themore rapid depreciation schedules and otherreductions in corporate taxes enacted in 1981,the United States had in place tax policies that,according to at least some analyses. favoredcapital investment more strongly than taxationin Japan and most of the European nations.33

"C .-er the postwar period, investment levels as a fraction ofthe ....mintry's gross national product have C.,,ctuated markedly

year to but with no in:idol-ice of any long-term ',lendup or down. SII! j. W. E. Conrad, and I.. I., mime).

Policy and Capital Formation," ederal Reserve Bulletin,October 1981, p. 749. For an excellent summary or internationaldifferer,co-. in capital formation. s(!I! V. C. Price. "capital For.!nation and Investment Policy." l' dern 1.;:conninies in Trans,-,i(w. (Jul:we and Adjte:treeel )'elides in Industrial( :otintr,e.,, I. Levestm and I. W. 11/:coler !eds.) Ilsoulder, Colo.:West vim.v Pro's, 1979). p. 135.

"K. W. Aronson. -The Low. Ll.S. Hate of Savings," NeerTimm,, Dec. 22, aPe p

C. F. Kopits,- rax Provisions lo Itmst Cape. it Formation VaryWidely in LIdosfrial Nations," 17, 131110, p. 055.The effects inter;.:Itional :1117ffelict!s in taxation on coin-petitiveness are, (1.^.1211001iIrirdy complicated. The .Si11114:ci(:111HCS thai ,oply to otbor international comparisons are it

work, compmepled by the co:nolexitii.,. of the tax' codes iu eachcountry :dies rf tax;',:on be compared in a rminherof d;f1ri.:0 wm..3-. with results depend on factors suchprojected in:.; ,1.1) s. Even taxIhems micro:, ap, caoot a. prohlm,a n)f re:atine the

Ch. 7Financing: Its Role hi Competitiveness in Electronics 279

At thu number of nations usesubsidies other than tax incentives

inure actively than the United States to supportcertain of their industries.34

Regardless of effects on overall supplies ofci:pital, the accelerated depreciation schedulesini7)lemented by the 1931 Tax Act seem likelyto place electronics firms at a disadvantagerelative to other industries with which elec-tronic. competes for funds. More rapid depre-ciation lowers tax obligations and increasescash flows from new investments. Faster de-preciation raises profits for projects that wereformerly marginal or unattractive; the resultshould be to increase the overall demand forinvestment. funds and the overall rate of invest-ment in industry. For most industries deprecia-tion is inure rapid under the new lawbut not`necessarily for electronics.

The 1981 Tax Act permits equipment such asthat used in production or R&D to he depreci-ated over either a 3- or a 5-year period. Former-ly, pro:Mc:Hon equipment was depreciated atrates at ieast nominally related to actual oh-

effects policy to economic performance remains:-See.k.ftv.c1;; Industrial Countries: A Ilriel Cori,

pareion. op. cit.Sobsolies and Corapetitiveness International-

' Iv NPA Cr.!. ii'' eon changing International1,:no.try

Prot(' Universal Insiruments

Aut.prnated machinery for ele..:tronics assembly

solese.N.:e. All equiHnent used for R&D cannow be wr:itt. , off over 3 years; so can anyequipment that could, under the old law, bedepreciated in 4 years or less. All other equip-ment is now eligible for a 5-year writeoff.35

Although electronics firms probably pur-chase little equipment with useful lives lessthan 3 years. this is nonetheless now theminimum depreciation period. To the extentthat a company was formerly able to write offsome of its equipment more quickly, it may suf-fer a reduction in cash flow. While this shouldbe no more than a minor factor, the new de-preciation procedures will place some elec-tronics manufacturersindeed, firms in any in-dustry where technological change results inrapid obsolescenceat a disadvantage withrespect to industries that reap greater benefitsfrom the new depreciation schedules. The lat-ter tend to be industries where technologicalchange is slower, and equipment has a long/useful life. Steel and other heavy industries areexamples.36

Focusing on levels of domestic savings canalso lead to underestimates of impacts flowingfrom international financial markets. Nationsneed no longer rely on domestic savings for in-vestment; international capital movements =Irelarge and continuing to grow. For the UniedStates, these long-term capital flows, both port-folio and direct investment, have been negativefor many yearsi.e., the flow of funds out ofthe United States has exceeded the inwardflow.37 This implies that rates of return are

',5Econornic Recovery Tax Act of 1981. title II. subtitle A, SSC.201. Also see P. Oosterhuis, "High Te;;:mology Industry ;IN: TikXpolicy in the 11Jfills,' National journal, Jan. 1. 19113, p. 40.

,"The only analyses of differential impacts across industriesthat have been published are on a highly aggreg4ted basis; thusit is not clear how electronicsmuch less particular portionsof the industrywill fare. The "machinery and instrumen0i-category. within which the electronics industry falls. is one ofthe least favored sectors under the new depreciation method:.See aurunilic Report of the President (Washington, D.C..lzibroary 19921, pp. 122-125. particularly table 5-7; also J. C.( ;rave.:'-'. -Effects of Ore Accelerated Cost 1-4;cove; T. Sy:item byAsset Type," Congressional Research Service,-A'.- ;. 31_ 1981.

"Althowh direct investment of U.S. funds kwerscas is still run-ning o to three times the level of forebiln investment in the!sited States, foreign holdin;ts of U.S. securities roughly ialance

American holdings of foreipn securities. Sit Statistical A, hstract11 the iThited Slates, 1982 40 103rd u.d. (Washington, D.C.:partment of Commerce. 1.1e,:einhet 1982)i p. I

28 ,1

Corn & :: fl eness in Eiectronics

11: :2. erseas .tt hoe te. 1.l ereFore gen-crating more savings in tile tjnited States v. iii

trily :':crease the rate of domesticeitnital lot aiat:i.e. funds :nay flow abroad in-stead. At the same time, were the cause of highinterest rates in the United State.zi simply ashortage of fends for capital investment, 111011-

flow here from abroad. In-ternal:tend capita: markets Operat:- effi-...1cr:t iv '111(if'1' '1.1111 t.irCITInStiinCeS.

Japan.-1!, at extent have .ii1F.:11eSe electronics

ft ;ice, beim aided by the unique structure of thecapital markets. a I .ictor that has

heel, (feattiu ei( to give tap ;''s corporations ad-vantages in niter-me:1(min cor)petition?

Postwar Trends in the JapaneseFinancial System

financial system lies changed moreihan ind.-4 over the past :33 years, and the cowl-

markets and financial institutions ere nowfar reinoved\from their grossly underdevelopedstate ill the early postwar period, 'the transfor-mation of the japanesc financial system haspaolleled, first, the reconstruction of Japan's,.!conoiny, and, following that. the count r( 'sdramatic industrial expansion. 'rile yen hasbecome a major internotional currency, whilel'eko only l).1 :iv by governmentalemerging its a (vorld banking, center.

(live.; the ,.peed with which the Japanesefinaricil system has evolved. it is easy to hemisled Ir., images frotil the past. Yet the future,e% en more than the present, should he the real.,oncet n: effects on.international competitive-IleSS ail" Functions ()I' the dynamics of changein markets ill the t Jnited States, Jima..and elsewhere. Insight into competitive trendsdepends oil these dynami s.

Because systemic changes fi; Japan havebeen evolutionary rat'-''r than revoiutionary,(tlelnelltS of validity on remain ill descrip-time; of financial institutions that are Otherwiseoutdated. For example; some discussieu:: im-ply that government, the Rank of the

cum:nen:A banking system, and varioussectors are all hierarchically

with control over resource allocation vested inthe Ministry of Fin -nice. AlC7ough hardly the:ase today, this is probably a fairif simplis-

erepresentation of the situation twodeca(Ifis ago. And even Low. at the= level of in-dividuni inw lment decisions, the Japanesesystem resi). much more directly to the(vishes of government than does that in theUnited StiLes. Hut if government guidance stillexists. it is a far weaker force now than 20 years.

(_), and the investment climate in Japan muchmore fluid,

Now to more concrete Questions: WI:A doJapanese corporations utilize much gieater pr)-portions of bank debt in their financial struc-tures than firms in the United States or West-ern Europe? (beverage in European electronicsfirms tends to be grclt.'r than for Americancool: ias out less that of jApai:ese.) Theanswer lies par)y in the historical developmentof industrial groupings ill Japan, roost of whichcontain one or more banks.28 Japanese Mall ti-fiAuring companies for Many years have de-pended much more heavit.: on close workingrelationships with bankseven to the extent ofparticipation by banks in management deci-sionsthan have firms in the United States orEurope (West (Thrtnany is a partial exception,as ';,.lcussed below). In other pa of tin world,

generally enter the picture only ifreorganization follows bankruptcy, whereas in

"Theso 2roups, s oli d idihat.,/; 1,1,;.,(:nietiling hire holditiL, ror general background, Svf!

It. F. Cave:. and !,.1. Cekilsa, luchrslridl lir,14ani....iaion in Japan.op cit., As(), (;ronpin1,.., tit r

pt(wil Crolyyo: 1)(.:t1 vd1 '.d.irkuting Cunsultituis,11:1an illustration, consider the Hitachi group II of

1)1',J v 500 firms; ;is (,I I..td. held maiorityand minority sharestypic:dly in the 20 to 5(1 per(amt

1,1: the remainder (I. Gres.,r. !lig!) Technology andine.q! InthistriA A Strategy for U.S. Po/it:pm/kers,

Siii,c(?;:nnittee on Trade, Committee on Ways and Means, Houseof l<epresentative;,0)ct. 1,1980, p. While many of the affiliatesmake electrical and electronic.,: productsand Hitachi, r: Ad. isthe 1,irgest electrical and electronics firm in Japan other. pm-duce ilimehN.trical mailtinery. tr:,risport it n equipment,chemicals, and primary ineta'...s. of the gro,,p are linkedwith the Kangyo flank, tire Fuji flank, and the SanwaHaut, at. so it as tile Indust' ial Hank r,f lapa71 (industrial (;runeinns Iii Japan. pp. 121)121).

Ch. 7Financing: Its Role in competitiveness in Electronics 2.11

lapin -e; en with the weakeniog of the indieetrio i groups following the postwar di,banding

:1-le zuibatsu hanks have continued to in-fluence 7nanagerial direction for firms that arefinancially healthy as well as those in trouble.This close relationship within the industrialgroup is one reason Japanese banks are will-ing to absorb risks more like those of share-holders. Lead banks. it is said, frequently subor-dinate their credit voltmtarily. That is, thebanks act inn eh like holders of equity, :Ind de-fer to others lower in the hierarchy of claimson assets when ecopeenic conditions dictate.kliliat appears to Ameticans an oppressive debtload, may not be so from the perspective of a'horrower in japan.

The (Anse relationship between banks andelectronics :companies in Japan is only part ofthe story. Following Work .War II. Japan'scapital markets were undeveloped, with viablepublic markets for neither corporate debt norequity. Financial intermediation was carriedout almost entirely through the banking system.Of necessity, industrial expansion had to befinanced by channeling funds through com-mercial bank ,.39 Moreover. it was not an acci-(lent that alter.e;-dive financing methods did notdevelop more ,)idly as Japan progressed eco-nomically. The government could convenientlyguide the economy through the commercialbanking system. Government decisions to fos-ter economic growth by extending credit to in-dustry at the expense of consumer credit andinfrastructural development were implementedin this way. Today, the extent of governmentinfluence over the lending decisions of majorcommercial banks in Japan remains consider-ably greater than that exercised by Wes'erngovernments, France excepted (while WestGerman banks have a good deal of leverageover corporate managements, the governmentinfluence on the banking community is muchless than in France, as discussed later in thechapter). The lack of alternative sources offinancing for Japanese companies enhancedthe government's ability to direct economicgrowth; "window guidance" and a variety of

Ackely and I I. Ishi, "Fiscal, Monetary. and Related Pol-icies,- Asia's Oiant, op. cit., p. 153.

- 1 1 1 - 83 - '

other industrial policy tools would ha.e been'weaker instruments if corporations had beenable to look elsewhere for capital.

Reliance on External FundingAlthough debt-equity ratios have decreased

considerably in recent years, a:s shown infigure 52, Japanese electronics companies re-mai:, heavily leveraged. And, because banksare so deeply involved, even the definition ofleverage must be adapted to the Japanese case.In the United States, the usual indicators ofleverage 'ire based on the long-term debt in acompany. 's capital structurei.e., obligationsthat mature after 10 or more years, most ofwhich are bonds. Leverage can then be definedas the ratio of the value of this long-term debtto the value of the firm's equity, or to its totalcapitalization. In japan, as the figure :TIclicates,such a ratio would be misleading becausemuch of a typical firm's capital comes fromshoi Am"? b.ink loans. Table 60, which listssources o external funds for Japanese andAmerican corporations, shows that companiesin Japan have depended much more heavily onloans from financial intermediaries, mostlybanks, than on bonds. Bonds are issued byAmerican firms at nearly 10 times the rate inJapan, although in recent years these patternshave been changing somewhat. U.S. industryhas been relying to an increasing extent onshort-term bank loansprobably because un-certainty concerning future rates of inflationmakes corporations wary of floating bond is-sues at recent interest levels.40 In neither coun-try have stock iss;ies been a major source, ofcapital; still, American firms have relied moreheavily on new equity than their counterpartsin Japan.

Financing patterns exhibited by electronicscompanies in Japan differ, but not greatly, fromthose for Japanese industry as a whole. Elec-tronics firms have financed growth with inter-nal funds to a greater than average extent.Table 61 compares leveraging for electronicscompanies in Japan and the United States. Nei-

..""Thu Perilous Hunt for Financing," op. cit.

2


0

120

100

40

20

0

Figure 52.Debt-Equity Ratios for Japanese Electrical /Electronics FirmSa

Tclal Long-termdebt/equity debt/eq,

1972 1974 1976

Year

1978 1980

(Long-termnot

availablefor 1980'

4tf S t:ro-,i In 1972. 54. tit-^,4 to 197478, 14 thlT.5 1980.

rES 1972.78" .+oalcat ,o the L;ornpet;tiveheSs of the U.S Electronics Industry," report prepa-ec for CIA by L Ui Bc,gman Co under.,-,ntract No 033-1.5. p 471.

trf)-1 Japan Company Handbook (Tokyo: Toyc, Ke.:w S:prposhalThe Orien'ai Econornps- 1981).

Table 60.Sources of Exte-nal Financing forU.S. and Japanese Corporations

Prcportions by source,

Loans from banks and financial

1965.72

JapanUnited States

intermediaries 51% 89%Bond issues ........... 37 4Stock issues . 12 7

100°4 100%SOURCE T raaru.yama. -Frnancing Japanese Business." The Conference Ronrd

May 4978. p 47

ther the six Japanese electronics Firms, nor, :-<-cept for RCA, the seven in the United Statesstray too far from the patterns typical of eachcountry. As ti 3 table shows, the Japanese I rrtrponies have depended much more heavily onexternal capitala point that has been em-phasized previously (table 52). The heavy reli-ance on both shor.- and long-t -rm debt inJapan's electronics industry contrasts sharplywith U.S. semiconductor firms. Electronicscompanies in Japan utilize very large absolute

Ch. 7Finar,cing: Its Role in Competitiveness in Elect-hnics 233

Table 61.Short- and Long-Term Debt as a Percent Total Capitaltzelion for U S. and Japanese Electronics Firms

Ai; Japanese1975

Si x Japanese electronicscompanies. 1979a

38 0' ii

Wei cited average offour 1!.S. sRmicor.n.:clor

rnanufactures. 197c;"8.1'.',

')trtc.rdilual U.S. companies. 1 79

OS';C...=.3".:,..

IBM FicA

5.3'..:, 164'.4-.37.5 29.0 12.7 23.1 9.0 37.2

71 7 67.0 20.8 24.0 14.4 55.5

Stlarer-,oklet 5 ec:utti 28 3 33.0 79.2 75 o as.,; 44.4

To!aJ ca;)ita'dzton li.'./s 100P,S 100,..-',, 100'..". 100%

Lcr"c; 7:,-r. .:.?..);'e:',:..:,ty 1 23 01'6 0.15 '1.30 0.11 3847,, .ii ::t.,:".::-.;.,' . . 253 2C; G:5 0.32 017 0.566

;,,S,, ,.-.:C-, ani: 7,-;sr,ipaA.i.;-,_,J V., i'.,, :).-....i, .., , ,,Dia Na7-,,a, F.,,,,,Tti,:.0,,Cl:Clc,f TP.a::1,,..rurrt !!;

SC.,.,',Ci-..S Japanese companies u S anis Ja;a-ese Semics-Iclor7 _:1 tr,:iiiiiiles A +,.:,lancti Go,,,;-,a,scn.- Cr'..se F-lan.c,at Pc..i:cy tc, ine Serncc,r1CL:clo,1,111..:stryAs.;,-:,.: 37 ,-,-_, ..,e g 1:,,eC D 62 American Iimt Anni;a1 ii-.iiirts

amounts of short-term debt which is automat-ically turned over as it falls duei.e.. the iaansare rew-naten. normally et an interest rate con-sistent with prevailing market conditionsapointed cut earlier, these practices seem toplace Japanese banks in positions of consider-ably greater risk than would be acceptable here.

Tables GO and 61 emphasize the extent towhich capital structures in Japan are weiglqedtoward externa! financing. These practicesmust be taken into account when calculatingcosts of capital. The American financial sys-ternand therefore the methods of establishingcapital costs commonly used herepresumesa mix of alterna:ive sources of financing. Notall of these are widely available in Japan; sonicdo not even exist.

In the United States for example, individualportfolio holders adjust their :security positionsin response to changing market conditions,riding off risk against potential returns.

Americanseven those with modest assetshave become sophisticated investors, switchingin and .iut of certificates of d -.)sit, moneymarke, funds. corporate stocks, and other in-vestments in response to small swings in rela-tive rates of return. The escalation in real estateprices during the 1970's. reflecting high ratesof re!, investments in land and housingas w auvantares, is another, example.In Jap.in, individuals do not have this range ofinvestment opportunities; for example, marketsfor corporate bonds barely exist. The well-developed capital markets in the United Stvtes

maintain a state of dynamic equilibrium withrespect to one another which is not always at-tained in countries like Japan. Japanese capitalrnat-ketsand those in many other countries,-are narrower; nor are they as closely linked.Neither individual markets nor individual in-vestment decisions respond as quickly or asconcertedly to changing conditions. Borrowershave few-- potential sources of funds.japanesecorporations must use bank financing; saversmust rely on bank deposits or government sav-ings institutions. Under such circumstances,J. panese banks have little option but to accept

that would be unacceptable in the UnitedState!;, _heir other lending opportunities aretoo limited.

Ledgers of Japanese companies differ fromthose of American firm 3 on the asset side aswell as the liability side. Table 62 illustratessome of these asset side differences. Japaneseelectronics companies carry much rnere cashand other liquid. assets, reflecting in part re-tp.:irements for compensating balances im-posed by banks; they have less tied up in ac-counts receivable and inventories, The fixedassets of !apanese firms plans, and equip-mentare proport;anately smaller, in partbecause of grossly understated land values; insome cases plant and equipment ,R:iluationsmay be reduced further by depreciation rateshighei than in the United States.4

"On f.he undervaluation of assets in japan, 1. Keroda andY. Oritani, "A Reexamination of the Unique Features of japan'sCorporate Financial Sp-tic:two," lepe.nose Eq.:ono:lb.(' Studies,


Table U.Asset Categories as Listed on Balance Sheets of Electronics Firms inthe United States and Japan (percentage of total assets)

U.S. companies (1979)

AdvancedMicro

Devices MotorolaNational

SemiconductorTexas

InstrumentsCash and marketable securities 7% 5% 2% 6%Accounts receivable, net 28 26 32 29!nventories, net 13 29 27 18Other current assets 7 5 2 4

Total current assets 55 65 63 57

Net plant and equipment 45 34 35 42Other noncurrent assets 1 2 1

Total noncurrent assets 45 35 37 43

Total assets 100% 100% 100% 100%Japanese companies (1978.1979) Hitachi Matsushita Mitsubishi ToshibaCash and marketable securities 22% 18% 150/0 22%Accounts and trade related

notes received 21 15 31 22Inventories, net 19 17 26 19Other current assets 5 6 9 6

Total currettt assets 67 56 81 69

Net plant and equipment 16 12 13 15Other noncurrent assets 17 32 6 16

Total noncurrent assets 33 44 19 31

Total assets 100% 100% 100% 100%SOURCES: U.S. firmsDerived from data in Moody's Industrials, 1980. Japanese firmsDerived from data in U.S. and Japanese

Semiconductor Ir4rustries: A Financial Comparison. Chase Financial Policy for the Semiconductor Industry Assoctation June 9. 1980. appendixes.

But the important point of table 62 is thelarge fraction of short-term, liquid assets heldby Japanese electronics firms. These assets donot earn high returns. Thus, on the one hand,lower financing costs for Japanese firms arepartially offset by lower returns on their largeholdings of short-term assets. On the, otherhand, the greater risks that Japanese banks ap-parently accept are partially ameliorated by thehigh levels of these same current assets. Theresult is to reduce the capital cost advantagesof Japanese electronics firms while helping ex-plain how they can carry such high levels ofdebt.

(footnote continued from p. 283)summer 1980, p. 82. In general, depreciation rates in Japan arecomparable to those in the United States except for selected in-vestment categories that are favored by accelerated schedules.See, "Corporation Income Tax Treatment of Investment and In-novation Activities in Six Countries," PRA research report 81-1,prepared for the National Science Foundation, Aug. 12, 1981,pp. 90-95; also J. A. Pechman and K. Kaizuka, "Taxation," Asia'sNew Giant, op. cit,, p. 317, and Tax Rate's In Major IndustrialCountries: A Brief.Comparison, op. cit.

What Role Does Japan's Government Play?

The relatively underdeveloped state of Jap-anese capital markets gives the government lev-erage over allocations of funds for investment.In the absence of a wide range of alternatives,both financial institutions and industrial cor-porations are more susceptible to governmentinfluence. Does, then, the Japanese Govern-ment indirectly subsidize target industriesusing the banking system as a conduit? A varie-ty of mechanisms would be possiblefor in-stance, government funds could flow to thebanking community in the form of low-costloans earmarked for certain uses. The fundsmight come from tax revenues or from privatesavings deposited in government-controlled in-stitutions-such as the postal savings system.

Unfortunately, just because the possibleroutes are indirect, evidence bearing on thisquestion is scarce. The most useful comes fromthe collective financial statements of Japanesebankstable 63. The liability side of the ledger

.28


Table 63.Assets and Liabilitiesof Japanese Commercial Banksa

Value(billions of yen) Percentage

AssetsCash 7,559 6.3%Securities 14,335 11.8Short-term assets 86,634 71.9Other 11,947 10.0

120,475 100%LiabilitiesDeposits 86,302 71.6%Borrowings from the Bank

of Japan 1,570 1.3All other borrowings 315 0.3Short-term liabilities 18,605 15.4Reserves 2,210 1.9Other 11,473 9.5

120,475 100%aAs of the end of 1975.

SOURCE The Japanese Financial System (Tokyo: The Bank of Japan. 1978).

is of particular interestspecifically, borrow-ing from the central b :nk, the Bank of Japan.Although quasi-independent, the governmentholds majority ownership in the Bank, the oper-ations of which are closely monitored bythe Ministry of Finance.42 As the table shows,lending by the Bank of Japan to commer-cial banksthe -"overloan" phenomenonamounted to only 1.3 percent of total liabilitiesin 1975 (the percentage is no doubt less today).43This is too little, by itself, to give the Bank orthe government much influence over the restof the banking system. Nor could funds fromthe Bank of Japan significantly reduce costs ofmoney to commercial banks. Such weight asthe government might exercise would, there-fore, have to flow from other sources; someAmerican observers hold that informal chan-nels are quite sufficient.

The situation was different in earlier years,when overloans were much larger; their de-crease as a proportion of the total liabilities ofJapan's commercial banks is another indicationof the changing character of the country's fi-

42K. Haitani, The Japanese Economic System: An InstitutionalOverview (Lexington, Mass.: Lexington Books, 1976), pp.164-165.

"An overloan simply means that commercial banksindi-vidually or in the aggregateare in debt to the Bank of Japan.See Suzuki. op. cit., pp. 1-'.3.

nancial markets. Then does Japan's Govern-ment still influence lending decisions? In thepast, target industries were certainly favoredwith low-cost capital, although the extent andforce of this on industrial development is muchless at present than 20 years ago. The govern-ment appears to act largely through informaland indirect mechanisms, rather than explicitlyallocating low-cost funds. Economic develop-ment goals set by government after extensiveconsultation with financial institutions and in-dustry have traditionally been supported by thebanks. Because of the close working relation-ships among government, the banks, and pri-vate industry, suggestions made by governmentofficials tend to be consistent with the predis-positions of bank managers.

In nracticr,, ins flow preferentially to largercompanies, r,:fr.fitt of which are associated withone or eror:. of -lie major banks through an industrial groupink, Interest rates on these loansare typically {hose for small borrowers(such discrEtthrotirm is common in all indus-trial f:. :ntries). ill, t.:;41),:r.pite the higher financ-ing c;..)sts faced by rKe.,.e and small companies,firms such as Sony Honda have beComehighly successful (luting the postwar neriod.Some have even managea to establish theirown industrial groupings; Matsushita', whichhad fewer than 2,000 employees before thewar, is the most prominent case in electronics.

Costs of capital in Japan dependfar more onbroad controls exercised over interest ratespaid by the banking sector on /deposits andcharged on short-term loans than on govern-ment guidance of investment flOws. High ratesof capital formation have been rooted in sav-ingsthe mirror image of investmentas illus-trated previously in table 54! The savings ratein Japan is especially surprising because formany years the government followed a delib,erate policy of keeping interest on savings andother investments low. The effect was to pre-vent interest rates from becoming the primarymechanism for allocating capital, as wouldhave occurred with market-determined rates.But if in earlier years capital rationing gave ad-vantages to some sectors of the economy,

2 9 u

286 ticeirVional Competitiveness in

others the cost. In general,.L.5:2:h.-.6ted at the exuenSe of

savers.

Table personal nr..ar$ in jahave zere, or negative -rates Ofturn (p:1.1:!r ;.Y.ii,-,IstineAts for infle:..-1.);1-(.,3r rraur-11of the rfa.-4 20 ?lependingi.oFfnrse extenton the (',.';'51e,!-Ltcs;-,-.P-5 chosen. neriod1961-69, major categw..:ts vf person-al 'in the tai'..';:,--remainetl.fixedLonger ttarned ,.st at a'ooutthe rate:IA Pa ti .1.141.1Ce the real ...-.1)turr(s wereessentially -;;tf.ri:). Sli):rter termnegative ',7eturns, Mvich the same --airs Vrrae dur-ing the 1.-!.-3-70';-;, WiCa price consider-ably more Large part 1-,..4::,11,1193 sud--den inc-,:Yec-r,iv:-::<; e...nergy pres, nr:,-tatAs. in1974-4 !ri were perio;71.ca517!13,444-justeri, returns remair, f.7/.1-:

'r;./I'ViPtgS been ttaal..4,hthe ba;.(1jr.:---.,; ,Ifyslem, appearing t

dusiry bow rates? The que5.7.'.:.ncan be asko...I another way.. W,-.sat tre.ncicwould interest rale.,..fi breve beeilfreer to adjust, and had saver e'n,:z--A-1 atoir-alternatives in placing their funds? 7. 1.7;the Japanese Government has ration::ci

Table 64.-Interest Rates and Price Inflation in Japan

,'.nnualchange in One-yearConsumer Demand time Postal ca:-Ir ws

Price Index deposits deposits Ordinary 2-3 year

Annual average interest rates

1961-69a 5.5% 2.19% 5.5% 3.6% 5.5%1970 7.7 2.25 5.75 3.6 5.751972 4.5 2.0 5.25 3.36 5.51974 24.5 3.0 8.0b 4.32 8.01976 9.3 2.5 6.75 3.84 6.751978 3.8 1.0 4.5 2.4 4.6

alnterest rates were held fixed over this period: the values given are the ceilingset by the government. The change in Consumer Price Index Is the average forthe period.

bTwo-year or more.SOURCES: Price Index-Karst/ye( Rodo Tokel (Useful Labor Statistics) (Tokyo:

Nihon Seisansei Honbu (Japan Productivity Center. Labor ProductivityDocuments Cents :., 1981), p. 136: data based on Shohisha BukkaShisu (Consumer Price Indax) (Tokyo: scrItu Tokei Kyoku (Prime Min-isters Office, Statistical Bureau).Interest Rates-1967-1974, H. C. Wallich and M. I. Wallich, "Bankingand Finance," Asia's New Giant, H. Patrick and H. Rosovsky (eds.)(Washington, D.C.: Brookings Institution, 1976), p.261. 1976 and 1978,Bank of Japan, Research Division, New York, N.Y.

the answer must be as follows: left free to ad-t;Lls`, interest rates would have been higher. Oni'tte other hand, if the government's actionss,..:1--ved primarily to allocate funds among sec---iws of the economy, then the answer is less ob-vi...)us. To the extent that commercial banks bor7rFivred from the Bank of Japan-and overloans

lame- during the 1960's-then interestr±Tres were probably depressed relative to levels211,:tt better developed financial markets would

set, particularly if overloans had been proli1 fed. current impacts are uncertain, if

i-rtily because overloans haVe declined in recentyears.

The Japanese Government has other meansle:; help selected industries get investment cap-MI. There is, for example, the Japan Develop-Treitt Bank-a public corporation throughwhich moneys from postal savings and trust

7:ixounts can be funneled. In the early postwaryears, the Bank was a major instrument of gov-ernment policy, but its influence rapidly de-dkacd; the-Development Bank provided 22 per-

-of all capital invested in industrial plant?.s.,..1 equipment in 1953, but only 5 percent in

Between 1965 and 1974, loans fromnment financial institutions-of which

'1",cs:.t,elopment Bank is but one-averaged4. :; percent of new industrial funds.45 Still,'.mall percentage came to about $3 billion

arninally, more than sufficient for major im-;pacts on international competitiveness if con-centrated on well- chosen targets.

Continuing Change in Japan'sFinancial System

In terms of competitiveness, and from theperspective of the American electronics in-dustry. the critical questions deal with thefuture. Change in Japan has been rapid, andthe pace may even accelerate. There are at leasttwo reasons:

1. Shocks to the Japanese economy stemmingfrom high energy prices beginning in1973-74.

4C. Johnson, Japan's Public Policy Companies (Washington, .

D.C.: AEI-Hoover Policy Studies, 1978), p. 98.sHaitani, op. cit., p. 169.

Ch. 7Financirg: Its Role in Competitivenesi in Electronics 287

2. Changing aspirations and expectationsamong savers, consumers, and the generApublic, along with the increasing complex-ity of Japan's maturing economy.

The explosion of energy priceswith deepimpacts on an economy that depends almosttotally on imported fuelshas stimulated areevaluation of economic goals. The gov-ernment is paying more attention to efficien-cy in allocating resources, backing away fromearlier commitments to high rates of economicgrowth regardless of costs elsewhere. As peo-ple's expectations rise, Japan is devoting moreresources to the public sector, aiming to im-prove the quality of life, broadly conceived.Although public sector expenditures remainwell below levels common in Western Europeor the United States, more money is goingtoward environmental protection and a varie-ty of social welfare programs. Defense spend-ing is slowly increasing, par6y in response toU.S. pressures. Finally, the Japanese are dis-covering that the days are over when a few sec-torsgrowing very rapidlycould lead the

=country's economic expansion. Future growthwill be slower and more balanced.

These trends in the Japanese economy ,willbe mirrored in the financial setting for in-dustry, where change is already well under-way. For example, the government has lostsome of its influence over interest rates; asJapan takes a more active role in internationalfinancial markets, with funds flowing in andout in greater volumes, interest levels will moreclosely follow those in other industrial nations.The dynamics of the Japanese financial systemare carrying it steadily toward the Americanmodel of open capital markets. This does netmean that the Japanese Government will aban-don its past efforts to guide the economy. Japanwill remain a nation where industrial policyis a powerful force. But, as large Japanese firmscontinue to expand internationally they willhave more lattitude for independent action,and the government will necessarily play alesser role in the allocation of resources.

Four clear trends can be discerned in theevolution of the Japanese financial systera:48

1. Interest rates are becoming more respon-sive to underlying conditions in capitalmarkets, and as a result are less subject tothe dictates of government.

2. Corporations, especially larger ones, aredeveloping alternative sources of fundingand relying less on banks.

3. Banks, partly as a consequence, are look-ing to individuals and small businesses asborrowers.

4. The Japanese Government, in response totrends already visible, is moving towardcloser links with the international finan-cial community.

While pressure in Japan for market-deter-mined interest rates is not new, only recentlyhave events combined to make this outcomea virtual certainty.47 One precipitating factor 7-has been the government's own need, follow-ing several years of deficit spending, to enterthe capital market. The national debt rose from11.7 trillion yen in 1972 to 62.3 trillion yen in1978.48 Iii earlier years, the banking systemwould have absorbed bond issues floated byJapan's Government to finance this debt, withinterest rates set at low levels to minimize coststo the treasury. But as such bonds have becomea larger portion of bank portfolios, and as anactive secondary market for government bondshas developed, bank managers have becomeless willing to accept new issues at below-market rates. The government has had to raiseyields to levels consistent with secondary mar-kets. At least for government issues, a long-term bond market more typical of industrial-ized economies is developing.

Banks have also sought more freedom to setinterest rates on certificates of deposit (CDs);

41. E. W. Kirby, "The Japanese and Their Changing EconomicEnvironment," Business in Japan, revised ed., P. Norburry andG. Bownas (eds.) (London: Macmillan, 1980), p. 85.

4'M. Borsuk, "Japan/Interest Rates: Consequences of Rates Sen-sitivity," Far Eastern Economic Review, Mar. 26, 1982, p. 59.

48K i Thy, op. cit.; p. 88.

2 2

2&6 International Competitiveness in Electronics

Japanese banks, after much negotiation and theacceptance of a variety of restrictions, werepermitted to issue CDs beginning in mid-1979.By now, movement toward more flexible finan-cial markets would be hard to stop. If interestrates are decontrolled in a portion of the econ-omy, pressures elsewhere will lead to a paral-lel freeing of rates or else to severe distortions.Such forces are particularly potent given thewidespread, desire to see Japan become an in-ternational financial center.

In terms of the U.S-:--electronics induE.try,movement in Japan toward market:determinedinterest rates should help defuse concern over:government-subsidized financing. Furthermore, as Japanese capital markets become bet-ter developed, new financial instruments willcome into play. Firms will be able to secur2capital from a wider variety of sources, at leastsome of which will be less susceptible to gov-ernment pressure." In short, investment deci-sionmaking will become more decentralized,as in other highly industrialized, capitalisticcountries. Both business executives and gov-ernment officials in Japan have been concerned over the high rates of bankruptcy of re-cent years. Many of these failures have beencaused at least in part by the highly leveragedpositions of Japanese corporations. As a con-sequence, companies have been attempting tobroaden their sources of financing in bothnumber and kindone of the reasons somecompanies are marketing securities overseas.To attract foreign investment, Japanese com-panies will have to offer rates of return com-petitive with those in other countries and othercurrencies. This suggests that capital costs injapan are unlikely to diverge very far fromthose in other parts of the world.

Finally, the orientations and strategies of themajor commercial banks in Japan are shifting.As corporations seek more broadly based fi-nancing, and as the profit levels of banks de-cline, bank managers have been forced to re-

"Japanese industrial firms floated more than twice as manybond issues in foreign markets as domestically during 1980. SeeNI. Kanabayashi, "Japanese Business Borrowings Abroad Surgedto Record In Year Ended March 31," Wall Street Journal. May12, 1981. p. 36.

evaluate their own portfolios. Many are at-tempting to develop alternative markets forloans, including foreign lending and expandedconsumer credit. The Ministry of Finance hasrecently given banks a good deal more latitudein making overseas loans, although informalquotas still exist.5° Symptomatic of the changeis the announcement of a loan at favorablerates to Fairchild for the construction of asemiconductor plant in Japan.51

Lending to households is also on the up-swing. Mortgages, installment buying, andother forms of consumer credit have been more

the'exception than thu rule, but bank loans forhousing expanded fivefold during the decadeof the 1970%s, and now account for same 10 per-cent of total bank credit.52 Today., as table 65indicates, households still borrow much lessin japan than in the United States. Continuingmovement to.vard greater consumer lendingand more diversified bank portfolios againpoints toward capital markets in which theprimary allocative mechanism will be themarket-determined interest rate.

Internationally, Japan has made an explicitde,;sioninvolving both government and thefinancial. communityto take a more promi-nent role in matters affecting the world'scurrencies.53 This shift reflects a number of

Marcom; Ir.. "Borrowers Are Eager 7o G,M Yen Loans ButMust Grapple With Japan's Delays, Wall Street Journal, July7. 1902, p. 24.

""Japan Offers Loan to Fairchild for IC Plant ... " Electronics,June 2. 1982, p. 73.

"Kirby. op. cit., p. 91.""(apanese Official Says Government Wants the Yen to

Become Major World Currency." U.S. Import Weekly. Feb. 2,1983, p. 572.

Table 65.Household Borrowing in Japanand the United States

Mortgages and consumer installmentloans outstanding as a percentage of GNP

Japan Unaed States1965 2.3% 60.8%1970 4.6 59.41975 12.1 6311978 . 17.5 68.1SOURCE: E.SakOlhara, R. Feldman, and Y. llarada, The Japanese Financial Sys

tem in Comparative Perspective. Joir.it Economic Committee. Congressof the United States. Mar. 1:, l682, p. 21.

el I'1

Ch. 7Financing: Its Role in Competitiveness in Electrzmics 285

ti ms7-"".7:-

Photo credit: Perkin-Elmer

Electronbeam lithography system used for makingintegrated circuits and masks

converging events, the most important ofwhich have been Japan's continuing trade sur-pluses. These surpluses have led to demandsin other parts of the world that the yen be usedas an official reserve and unit of account inprivate transactions. Foreign-held balanceshave been inc:reasing rapidly as an actin e in-ternational market in yen develops.54 Finally,foreign investment by Japanese corporationsis expanding swiftly. Japanese foreign directinvestment nearly doubled between 1979 and1981, reaching a level of $8.9 billion in fiscal1981.55 The international position of theJapanese is beginning to look strikingly like thatof the United States 20 years ago, with currentaccount surpluses offset by outflows of directnve :.,tment funds.

54Foreign hole-kings of yen reached the equivalent of about $10billion by mid- 1930still small in absolute terms but doublingover a period t f 11/2 years. See E. W. Hayden, "Tokyo as an In-ternational Fla anciel Center," Atlantic Community Quarterly,vol. 19, fall 1981, p. 351, which also outlines forces contributingto slackening government influence over Japan's financial mar-kets. Hayden emphasizes that continuing distortions can be ex-pected for some years to coma.

" "Japan Capital Abroad Reaches Record in FY 81." Japan Re-port. joint Publications Research Service IPRS 1/10( 7r, June 7n,1982, p. 10. For examples of Japanese Invet.:.me:,:!i !., Fsee A. L. Otten, "Japanese Firms Press European VF TnHelp Profits and Deter Protectionism." Wall St:-eat'16, 1982, p. 54.

The internationalization of the yen will havea major effect on Japan's financing practices.As long as the economy could be insulatedfrom outside impacts, the government couldsuccessfully hold down interest rates. This in-sularity is breaking down as banks, privatebusinesses, and individuals take advantage oftl,,, ,sange of financial opportunities nowopen to them. To the extent that Japan becomesa center of international financea processalready well underwaythe domestic financialsystem will become irrevocably linked to largerworld markets. And, as the United States andother Western nations have discovered, an in-tegrated international market renders an inde-pendent monetary policy aimed at controllinginterest rates virtually impossible.

France

In general, French electronics firms have notemerged as strong international competitorsnor have U.S. companies confronted insur-mountable difficillties in competing withinFrance, although the ingenuity of the Frenchin creating nontariff barriers may match thatof the Japanese. Therefore this/ sectionandthat following on West Geimanyoutlinesfinancing methods much more briefly than forJapan.

In some ways the French financial system -more closely resembles the Japanese case thanthe American.58 Like Japan, France has a tradi-tion of government intervention in what wouldbe private investment decisions here, and thegovernor good deal of control overallocat

ui dWdY from some of its earlierpractices in the late 1970's, and toward freermarkets for capital as well as goods. Althoughthe outlines of Mitterrand's industrial policyremain somewhat vague more in terms ofmechanisms than objectivesthe SocialistGovernment has begun to, alter a number of thespecific practices inherited from earlier ad.

41 is largely based on 1. Zysman, Governments,Markets, ,..?ad Growth: Financial Systems and the Politics t, In-dustrial Change (Ithaca, N Cornell University Press. J83).ch. 3.

294

293 Interna:ionar Compelkiver.ie:ss ;n Elec:ronfcs

ministrations. But regardless of swings of thependulum, government intervention in capital(and other) markets is a longstanding traditionin France, and the thrust of French policies asthey relate to investment is not likely to changeradically_

On the whole, the French financial systemseems more tightly controlled by governmentthan the 'Japanese system, and certainly farmore so than in the United Statesin partbecause .so much of French industry is natio.n-alized. One major d.fference between theFrench and Japanese cases is that France hasbeen a great ideal more tolerant ..of foreign-owned gnterprises. One-fifth of the .couratry's200 largest firms fall in this category. Whenfaced with weak domestic indwAries.includ-ing, at various times, semiconductors and com-puters the French Government has sometimeschosen to allow, even encourage, foreignownership.

Industry in France, as in Japan, gets most ofits external financing from institutionallenders, generally banks. Relative to such coun-tries as the United Stares or Great Britain,securities markets are small; within thesemarkets, sales of bonds far outweigh equities:Even the comparatively small French bondmarket serves mostly as a source of funds forgovernment and far nationalized companies,plus financial institutions of various types.Manufacturing firms look predomiriantly tobanks for their credit needs, roughly parallel-ing the situation in japan.

French industrial policy as it relates to invest-ment is based on underpricing capital, thenusing a mixture of formal and informal mech-anisms to ration funds to investors. Mar-ket-determined interest rates play a distinctlysecondary role. To some extent, lending insti-tutionswhich must restrict total loans 4'1

prescribed amount or be penalizedare freeto choose recipients of funds. But the gt.rM.meat also has a voice, primarily through theMinistry of Finance.

The Ministry can act in -a number of ways.At one level, its influence is informally exer-cised through a network of contacts within the

financial community. At another level, the gov-ernment intervenes more directly. For exam-ple, a particular bank's Joan limits might berelaxed to allow /funds to flow to a favored firmor industry. Sometimes, companies receivegraats or loans directly from the government.57Selected firms and industries may also benefitfrom low interest rate or loan guarantees.

The ability of the French Government to in-:terve in capital allocation decisions isfacilita._ed becausemuch more than in theUnited Statesthe savings function is splitfrom the investing function. That is, the finan-cial institutions that take deposits differ fromthose that lend money. Typically, funds col-lected by deposit-takers are first lent to in-termediaries specializing in longer term. in7vestments. These intermediaries, in turn, lendto private enterprises. Direct transactions be-tween deposit-taking institutions and corporateborrowers are infrequc at. Table 66 illustratesthe contrast with the United States. On theaverage, the deposits and loans of Americanfinancial institutions are much more nearlybalanced. iV.most one-third. of the total valueof loans to nonfinancial concerns in France

"For examples in electronics. see, E. DiNlaria. "French Govt.to Bolder IC Industry With Grants and Loans.- Electronic News,Apr.. 27. 1981, p. J.

Table 66.Holders of Liabilities and Claims With theNonfinancial Sector in Mited Stn' France°

United States France

Deposittaxing institutions ks,savings and /loan*:

Propoittion of total deposits 57.3% 71.7%Proportion of total loansrellaimis .. 51.8 48.3

Long-term credi't institutions:Proportion Of total IN)05i'ls 8.2Proporlicl 7 9 32.9

Investing T-F,iltutruns hr....surance companies,pension `ands):

of total deposits 32.3 11:3}-,rt.:ition of total loanslaalms 31.2 9.3

Other financial Institistior6:Proportion of total deposits q 8.8Proportion of total toans,(clairns VI 9.5

8As of the) end of 1975.

SOURCE: J. Zysman, Governments. Markets, and Growth: Financial Syolems andthe Politics rl Industrial Change (Ithaca, N.Y.: Cornell Universfry Press.1983),

2,9)


flow through specialized lona-term lending in-stitutions. an indication of their importance.

Many of the institutions that take-depositsor fend funds in France are at least quasi-public; with the completion of the Mitterrandgovernment's programwhich includes the na-tionalization of an additional 18 commercialbarks, plus the country's two largest invest-mfmt banking houses three - quarters of alldeposits pass through publicly owned banks.58This gives the bureaucracy many tools for in-fluencing investment decisions. Even wherefinancial institutions are private, the govern-ment can mediate between savers and invest-ors: its most powerful weaponeven if seldomcalled onis simply the ability to undercutprivate lenders with public funds.

Despite the pervasiveness of its influenceover investment decisions, the French Govei n-ment faces severe eznstraints in employing thistool of industrial policy. French industrypar-ticularly in high-technology sectors likeelectronicshas seldom been as competitivein international markets as West German orJapanese industry, much less American. Whileexceptions such as aerospace exist, the com-parative weakness of French corporationslimits their ability to operate autonomously inworld markets The French have lagged con-spicuously in MOS ICs and minicomputers;low-cost capital has not proved Tr ..ah help inbuilding a strong technological base for thecountry's electronics industry. Indeed, severalof the larger French manufacturers are con-trolled by foreign multinationals. Many moreare partly owned by foralgilAmerican. l dii:1:=es have itadoded CII -Hon-eywell Bull and Matra-Harris Semicon-ducteurs. Subsidiaries of foreign-owned enter-prises need not depend on French financial in-stitutions for capital; even minority ownership,which is often coupled with dependence enforeign technology, gives substantial leve'ragein dealing with the bureaucracy or with gov-ernment-controlled financial institutions. Fur-thermore, French companies that prosper he-

"P. Lewis. r-rance Begins $8 Billion Takeover of Private In-dustry and Banking." New York Times, Feb. 1 1982, p.

come less subject to the allocative dictates ofthe government. Not only do successful firmsget less attention simply by virtue of their com-petitiveness, but such companies can more eas-ily generate capital internally, or tap interna-tional sources.

Operating within these constraints, theFrench Government has used the financial sys-tem to affect the country's electronics industryin two basic ways. Not only has the govern-ment supported the industry with both directand indirect financial assistance, butoften asa precondition for loans or grantsthe govern-ment has sometimes insisted that the industryrestructure. While restructuring has frequent-ly been aimed at fostering competitiveness,maintaining employment has also been amotive.

In electronics, the best known examples ofgovernment- directed restructuring have beenassociated with the Plan Calcul, which effec-tively ended in 1976 with a merger of the com-puter firms Compagnie Internationale pourL'Informatique (CII) and Honeywell Bull, thelatter partially U.S.-owned. The French Gov-ernment promised as much as $700 millionover a 5-year period to the new concern, alongwith further assistance throw-1 purr' lases ofboth hardware an'15, fi 10).14 such

privy '; in France are encour-ageu uehaN,e y s consistent with thegoals of the government; the carrots and stickstend to be more visible than in Japan.

Franc i5 now pursuing a similar strategy inlicm atronics. As discussed in detail in

aliapte. 10, the aim is to develop an indigenous,internationally competitive industryin partby encouraging joint ventures through whichAmerican firms transfer technology to aFrench partner. One carrot here is the promiseof R&D funding reported to total more than$300 million over a 5 year period. The U.S.partners have tended to view this as perhapstheir only route to sales within the large,lucrative, and closed French telecommunica-ions market.

''')uite apart from direct government aid, elec-tronics firms find that their status as teChno-',Ogical leaders compared to the rest of French

2 Fr:

292 lnternatic qal Competitiveness in Electronics

industry makes it relatively easy to locate fundsfor R&D or expansion. Through its pervasiveinfluence, France's bureaucracy can assurefavored industries funding from either privateor quasi-public sources.

But again as in Japan, the extent of govern-ment influence on corporate financing has di-minished over the years. During the 1970's,companies meeting with success international-ly could deal with French banks and capitalmarkets largely free of government interven-tion. The government was more concernedwith firms and industries no longer able tocompete; to considerable extent, French in-dustrial policy has been preoccupied with suchsectors as textiles, steel, and shipbuilding. Agood deal of assistance has gone to firms inthese industries, which have been depressedin France as in much of the developed world.In this respect, the French Government has be-haved like that in other industrialized nations,including the United States.59

The French are now aggressively promo-lit,potential technoln laders like electronics,hoping to e firms that might proveable to com,, internationaiiy (ch. 10). Theplans of the Mitterrand government are ex-traordinarily ambitious in the spending levelsproposed for the support of new industrialtechnologies, with iw.ich of the effort in elec-tronics focused on semiconductors. France'srecord in attempting to promote technologi-cally advanced industries has thus far beenmixed: disaster with the Concorde; successwith the Airbus and helicopters, as well asnuclear power; notable progress in telecom-munications; little relative movement as yet inmicroelectronics. Direct and indirect financialsubsidies may contribute to technological suc-cess, but by the recent history in France arefar from sufficient.

West Germany

Financial mediation in the Federal Republicagain involves parties having much closer

"See 7. C. Price, IndUstriai Policies in the European Commun-ity (London: . Martin's Press. 1981), or an excellent discussionof how governments in Western Europe have attempted to dealwith the problems of distressed industries.

working relationships than typical in theUnited States. In particular, the stockholdingsof banks are a major source of their very con-siderable leverage over German companies ofthe Aktiengesellschaft (AG) variety. The AG,or joint stock companies, were once far morenumerous in West German industry, and mostof the large concerns are still organized in thisfashion. Over the postwar period, the numberof Gesellschaften mit beschrinkter Haftung(GmbH)privately held, limited liability orga-nizationsgrew rapidly; banks interact lessclosely with these.

One source of banking influence over the AGform of corporation stems from its two boardsof directors: the shareholders elect a super-visory board, which in turn appoints an ex-ecutive board, the latter responsible foroperating management.'" The supervisoryboard makes major decisions on matters suchas new investments or product lines. While noindividual can belong to both the boards of asingle company, there are no bars to simul-taneous service on the supervisory boards ofseveral companies, even if they are com-petitors. Officers of banks holding equity in aWest German firm often become di, ectors, anda single bank may be represented on the super-visory boards of a number of competing enter-prises. About 10 percent of the members of thesupervisory boards of the 100 largest AG firmsare bank officersnor is this the only mech-anism by which West German banks influencebusiness activities.51

Indeed, the role of banks is even more cen-tral in West Germany than in Japan. There arethree major reasons. The first is simply thatGerman banks are allowed to hold commonstock, as they can in Japan though not the

Konica, "The Rise of the Modern Industrial Enterprise inGermany," Managerial Heirarchies: Comparative Perspectiveson the Rise of the Modern Industrial Enterprise, A. D. Chandler.jr. ant! H. Daems (eds.) (Cambridge, Mass.: Hnrvard UniversityPress, 1980), p. 77.

6,See E. Hartrich. The Fourth and Richest Reich (New York:Macmillan, 1980), ch. 13; also IL Medley. "Monetary Stabilityand Industrial Adaptation in West Germany," Monetary Policy.Selective Credit Policy, and Industrial Policy in Franca, Britain,West Germany, and Sweden, staff study, joint Economic Com-mittee, Congress of the United States, June 26, 1981, r. 92.

29 y


United States. But, whereas Japanese banks arelimited to a maximum of 5 percent ownershipin a single company, the holdings of Germanbanks have not thus far been restricted(although such legislation has been consideredby the parliament in recent years). The secondreason relates to financial structure. Germanindustrial companies, again like their Japanesecounterparts, are highly leveraged, tending torely on bank loans rather than bonds. On theaverage, firms in the Federal Republic carryabout four times as much debt as equity ontheir books.e2 The high proportion of debt iseven more striking in light of the third char-acteristic of West German financial practice:this debt is heavily concentrated in the port-folios of only three banksthe Deutsche Bank,Dresdner Bank, and Commerzbank, all need-less to say very large. These three banks holdseats on the boards of 70 of the 100 largest Ger-man corporations."

Beyond the shares they own, banks in theFederal Republic frequently control proxyrights on privately owned shares' carried intheir vaults. These shares combine with Directequity ownership to create an impressive con-centration of voting power. The banking com-munity can vote more than 90 percent of theshares for many of the large publicly held cor-porations in West Germany." Because majordecisions must be approved by at least 75 per-cent of the shareholders, an effective veto isheld by one or more banks if they control only-,a quarter of a company's common stock. Evenmore so than in Japan, banks in the FederalRepublic absorb clear and explicit equity risks.

West Germany's competitors in Europe fre-quently complain over the relatively highleverage of German corporations, focusing oncapital costs, together with the major holdingsby banks of both equity and debt.65 Their rea-

"J. Ross-Skinner, "Germany's Powerful Banks," Dun's Review,January 1979, p. 68.

"Medley, op. cit., p. 115."M. Kreile, "West Germany: The Dynamics of Expansion,"

Between Power and Plenty: Foreign Economic Policies of Ad-vanced industrial States, P. J. Katzenstein (ed.) (Madison, Wis.:University of Wisconsin Press, 1978), p. 191.

"E. V. Morgan and R. Harrington, Capital Markets in the EEC:The Sources and Uses of Medium- and Long-Term Finance(Boulder, Colo.: Westview Press, 1977), p. 323.

soning is virtually identical to that now heardin the United States concerning japan: financ-ing costs are lower because the major Germanfirms make heavy use of low-cost bank loans.

As for highly L--veraged Japanese firms,however, the magnitude of any advantage incost of capital will depend largely on the taxbenefits accompanying a high proportion ofdebt in a German company's financial struc-ture. To the degree that inside knowledge andcontrol provide better information, the involve-ment of banks in the management of Germancompanies may also lower their perceivedrisks. If so, the banks might choose to lend onbetter terms. The same could be said aboutJapan, although the limits on direct ownershipmake it a less significant factor. In any caseWest German or Japanonly small reductionsin interest rates could plausibly flow from suchsources.

The deep involvement of West German banksin corporate financing has its corollary in arelatively underdeveloped capital market.While stock exchanges exist, trading volumesare much lower than in the United States. Sec-ondary markets for other types of financial in-struments are rare. In fact, for most transac-tions in German financial markets, orders mustbe placed with the banking community. Theusual modes of personal saving are bank ac-counts, insurance, and pension funds. Bonds,including those of financial institutions, ac-count for less than 15 percent of savings, equityownership less than 1 percent.

While banks carry great weight in the FederalRepublic, government influence on the financ-ing of private business is far less pervasive thanin France or japan. Although government own-ership of business accounts for about one-fifthof all fixed capital investmentabout the sameas in the United Kingdompublic ownershipin Germany is largely confined to the transpor-tation, electric power generation, coal, chem-icals, and shipbuilding sectors." In recent

E. Owen-Smith, "Government Intervention in the Econcmyof the Federal Republic of Germany," Governmental Interventionin the Developed Economy. P. Maunder (ed.) (London: GroomHelm, 1979), p. 160. The investment figures were 21 percent inGermany for 1972 and 19 percent in the United Kingdom for

298


years, some of the government's holdings havebeen sold to private interests. Both Federal andLiinder (state) Governments maintain owner-ship interests in banks, but their primary con-cern appears to be financing projects involv-ing housing, agriculture, and small business.Government-owned banks do not compete forbusiness with private banks. While direct gov-ernment subsidies to industry are substantial,amounting to several billion deutsche marksannually (about half as .much in dollar:.), mostof the subsidies have been directed toward so-cial welfare objectives or the support of indus-try in West Berlin and the border areas. Still,Federal funds have aided aircraft design andproduction, along with shipbuilding, coal min-ing, and housing construction.

(footnote continued from p. 293)1975. See Public Enterprise in the EEC. 1Y. Keyser and R. Windle(eds.) (Alphen aan den Rijn. The Netherlands: Sijthoff and Noord-hoff. 1978). Part 111 Federal Republic of Germany. p. 3 and PartVII United Kingdom and Ireland. p.

The picture that emerges, therefore, is oneof close working relationships between WestGerman industry and the banking sector. Com-mercial banks provide the bulk of externalfinancing for companies in industries like elec-tronics, and take a correspondingly active rolein management. In terms of control over thenation's economic activities, the banks occupya central position and wield considerablepower. While a variety of political interestshave recently pressed for reductions in thepresence of the banks, thus far change has beenminimal. Government, in contrast, plays a lessdominant role than in many other industrial-ized nations. Recently, the willingness of theWest German Government to let market forcesdetermine economic direction has come understrain. The eventual consequence may be agreater degree of intervention in microeconom-ic matters (ch.10).

Summary and ConclusionsThis chapteras several othersconcen-

trated on the U.S.-Japan comparison becauseJapanese companies are the most effective andaggressive competitors in the world electronicsindustry outside the United States.

From the narrow viewpoint of financingcosts, it appears that government support forJapan's electronics firmsnow manifestingitself particularly in semiconductors andcomputershas resulted in somewhat lowercosts of capital than would otherwise be thecase. But in real, inflation-adjusted terms it isunlikely that this cost of capital advantage ex-ceeds a few percentage pointsalmost certain-ly less than 5. By itstif, the effect would be tomake expansion somewhat less costly for Jap-anese firms, but the competitive advantagegained from this source alone would be small.Differences in costs of capital faced by firmswithin the industries of the United States orJapan are larger than the differences betweenaverage costs of capital in the two countries.

Although Japanese electronics companies con-tinue to utilize greater financial leverage thanAmerican firms, the advantages of this prac-tice are marginal at best. Higher debt-equityratios do not give Japanese electronics firmssignific: ;e, benefits in financing compared toAmer'''. 3n manufacturers.

Other analyses have resulted in estimates forthe difference in cost if capital between theUnited States and Japan that are considerablylarger. The explanation lies in expectationsconcerning future price inflation in the twocountries, which other studies have not fullyconsidered. To gain "real" returns, lendersmust demand interest rates in excess of the in-flation rate. Price inflation in the United Stateshas exceeded that in Japan by considerablemargins over the past few years, and the in-flationary expectations of lenders have re-flected this history. The direct consequencesfor costs of capital are perhaps less importantthan the effects on time horizons for corporate

29,


investment decisions. High and uncertain fu-ture inflation rates lead managements to an-ticipate large swings in the cash. flow returnsfrom capital. The longer the time horizon, thegreater the possibility that, at some futurepoint, the returns will be insufficient to coverinterest expenses. Such uncertainties bias in-vestment decisions in the United States towardthe short term.

Still, even if the real, inflation-adjusted costsof financing are not that much higher here thanin other parts of the world, costs of capital aregreat enough to create serious dilemmas for thefinancial managers of American electronicsfirms. These dilemmas stem from the limits ondebt broadly acceptable within U.S. financialmarkets, the need for rapid growth in capital-ization to meet expanding market demand forelectronics products, rising capital-intensity insome parts of the industry, and heightenedforeign competition.

In addition to greater capital equipmentcosts, manufacturers of computer3, semicon-ductor devices, and related products facemounting expenses for R&D and product de-velopment simply as a result of advances intechnology and the increasing complexity ofelectronic systems. As in the past, competitionwill be strong even among domestic firms;added competitive pressures will come fromthe Japanese. When the industry was small, andnew startups drove the technology, competi-tion was a vital source of U.S. strength. Nowthat the industry is maturing, the ingredientsof success are changing. Managers of Ameri-can firms are reassessing their businessstrategiesparticularly in terms of pricingwhile finding themselves hard-pressed tofinance R&D and new production facilities.

In :ecent years, American industry has notraise:i much capital from sales of stock. To in-crease equit, without selling stock, electronicsfirms must generate substantial flows of re-tained earnings from profits and depreciation.To some extent, the depreciation schedules im-plemented by the Economic Recovery Tax Actof 1981 will increase cash flows available forreinvestment, as will other changes in the tax

law. But competition is likely to hold down pro-fitability, thus limiting the ability of Americanelectronics manufacturers to finance rapid ex-pansion through reinvested earnings.

Compounding the difficulty for firms inmany portions of the industry is the rising levelof capital intensity. More expensive productionequipment is a fact of life for semiconductorfirms. Equipment used for R&D as well asmanufacturing rapidly becomes outdated. Thisis not necessarily a problem so long as writeoffskeep pace. But the changes in depreciationschedules adopted as part of the 1981 TaxActwhich fix depreciation on most equip-ment at 5 years for all industriescould disad-vantage electronics firms. In the past, deprecia-tion schedules were based, at least nominally,on the actual useful life of the investment. Thenew law shortens depreciation schedules forother industries, where plant and equipmentoften have much longer useful lifetimes. Withall industries now depreciating on essentiallyequivalent schedules, firms in electronics andother technologically dynamic industries arelikely to find themselves at a relative disadvan-tage. Their domestic rivals for investmentfunds benefit from greater increments in de-preciation rates and hence in cash flows,augmenting their ability to attract capital fornt:w investment.

U.S. electronics firms obtain financing froma variety of sources, depending largely on theirsize and stage of development. For many of theyounger companies, the original source wasthe venture capital marketwhere investorsprovide money to infant businesses in the hopeof greater returns than safer investments wouldyield. Venture organizations generally expec'tmost or their return in the form of capital ap-preciation; as a consequence, their investmentdecisions are sensitive to taxation of capitalgains. The 1978 reductions in capital gainstaxes were an important force in enlarging thepool of funds avail:ible for new ventures inelectronics.

As successful firms grow beyond the infantstage, they gain access to a greater variety ofsources of capitale.g., lines of credit with


banks, bond markets. They may also be ableto float public stock offerings. But manage-ments of electronics companies, following theprevailing American pattern, have stronglypreferred internal funding of growth. Somecompanies utilize considerable debt (table 53),but leverage in U.S. electronics firms for themost part remains low.

Such financial patternsparticularly thoseestablished by merchant semiconductor man-ufacturers during the 1960's and 1970'swillnot be easy to maintain during the 1980's.Greater demands for capital equipment and forR&D are combining with intense foreign com-petition to make the finar,cing of growth bysmall, independent companies more prob-lematical. But despite the attention focused oninternational competition, the fundamentalproblem is growthtogether with the upswingin capital intensity. Many once-independentfirms have already been acquired by larger cor-porations, at least partially in consequence oftheir needs for capital.

As a result of these forces, the U.S. elec-tronics industry will almost certainly be com-pelled to rely more heavily on external funds.This is one of the reasons for the concern overcosts of capital. Many industry leaders haveexpressed doubts that supplies of capital willbe adequateor that, if capital is available atall, the costs will be too high, particularly com-pared with costs of capital in Japan. Of course,funds will always be available for investmentsthat promise sufficiently high returns. Freecapital markets will clear at some interest rate,and it is this interest rateor pricethat servesas the allocative mechanism in the U.S. finan-cial system. But it is quite possible thatfroma broader perspective than simply returns tocapitalprojects that are otherwise desirablewill not be funded. Examples include the long-term basic research that undergirds a high-technology industry like electronics.

As a result, the question of whether interestrates in the United States may be prohibitive-ly high is a difficult one. The supply-side thrustof programs initiated by the Reagan adminis-tration was intended to produce significant

growth in the pool of capital available for in-vestment. If effective, such programs shouldresult in lower market interest rates. But manyof the steps taken will also stimulate demandfor funds. It has not yet been possible toseparate the effects on supply and demandflowing from these policies; the vital questionof how future U.S. costs of capital will com-pare with costs in other countries, also uncer-tain, takes the whole matter a step further.

Turning to japan, in years past capital cosh:for electronics firms there may have been sig-nificantly lower than in the United States. Thereasons are severai. Capital has, at varioustimes, been channeled to favored industrial sec-tors, including electronics. By controlling in-terest rates, the Japanese Government effective-ly circumvented the market as a mechanismfor allocating funds. But while remnants of thiscontrol remain, government influence over fi-nancing decisions by banks is now muchweaker than 20 or even 10 years agO. Further-more, as capital markets in Japan continue toevolve they will take forms more like those inother industrialized countriesi.e., market-determined interest rates will become the ma-jor mechanism for capital allocation. Strongerlinkages with capital markets in other countrieswill mean that rates of returnand hence costsof capitalwill not be much different in Japanthan in the United States. Thus, even if Japa-nese electronics firms enjoyed lower costs ofcapital in years past, these advantages are likelyto diminish in the future.

At the same time, the savings rate in Japancontinues to be extraordinarily high, thoughdeclining somewhat at the end of the 1970's.It may continue to gradually fall, but rates'ofcapital formation remain much greater than inthe United States even considering increasingpublic sector demands as a result of largebudget deficits. Moreover, Japan's Governmenthas a well-practiced capability for interveningin capital markets and steering investmenttoward favored sectors of the economy. Whatthe Japanese have called "administrative guid-ance" will not simply disappear. Still, thechanging character of the country's financial

Ul

Ch. 7 Financi'sg: Its Role in Competitiveness in Electronics 297

market's means that Japan's electronics firmswill have less of an advantage in the futurecompared to their rivals in the United Statesand Europe--some of which may even findthemselves tapping sources of capital in Japanto finance their own expansion.

As these trends proceed, major Japanese cor-porations will no doubt continue to diversifytheir capital structures, relying less heavily oncommercial banks. Corporate borrowing inJapan as a percentage of gross national prod-uct is declining as firms diversify their sourcesof funds. The leverage of Japanese electronicscompanies gradually decreased over the 1970's,as figure 52 showed, while public sector bor-

99-111 0 - 83 - 20

rowing has risen. Government bonds are be-coming major long-term tradable securities.Securities of many types, both domestic andinternational, are becoming more widely avail-able in Japan, and market forces are havingtheir effects on interest rates. The governmentwill have more difficulty in managing invest-ment flows, and the capital structures ofJapanese electronics firms will continue mov-ing closer to those in the United States. Assum-ing these trends continueand there is everyreason to expect them toat least some of theconcerns of American industry with respect ioJapanese financial practices should diminish.

302

CHAPTER 8

Human Resources: Education,Training, Manag -mend

Contents

PageOverview 301

Educatipn and Training 302U.S. Secondary Sch 001 Education in Science and Mathematics 303Secondary Schooling Abroad, Especially in Japan 304University and Continuing Education in the United States 305University and Continuing Education in Japan 314International Differences in Education and Training 317

Supply and Mobility of Labor 318Overall Labor Market Trends, 319Personnel Supplies for the U.S. Electronics Industry 320The Question of Mobility 321

Organization and Management / 326Organizational Types and Management Styles 327Worker Participation 330Japanese and American Management Styles: How Much Difference? 333

Summary and Conclusions ; , 334

Appendix 8A. Japanese and AMerican Managethent Styles: A Comparison 33')

List of Tables

Table No. Page67. Engineering Graduates asja Percentage of Their Age Group 306 .

68. U.S. Degrees warded by Field 30669. Enrollments in Japanese/Colleges and Universities 31570. Distribution by Size of Japanese Firms Providing In-House Training 31671.- Labor Force Growth in/Several Countries 31972 Work Stoppages Due to Labor Disputes in Several Countries 32873. Rankings by American Managers of Factors Contributing to Productivity 3358A-1. Responses of Middle and Upper Managers to Questions Dealing

With Communications and Decisionmaking Styles 337RA-2. Data Related to Employee Satisfaction 338

List of Figures

Figlin? No. Page53. R&D Engineers and Scientists in the Labor Force 30554. Engineering Graduates of American Universities 306

201

CHAPTER 8

Human Resources: education,Training, Management

Overview

Among the questions this chapter addressesare: How good are the people an industry de-pends on? Is the pool from which they aredrawn big enough? How do they get their train-ing? And, the mirror image of these: Does in-dustry use their abilities wisely?

Countries without adequate human resourcescannot hope to design and manufacture prod-ucts like computers; even televisions are be-yond the capabilities of many developing econ-omies. In the-United States, peopleunskilledor skilled workers, engineers and technicians,managersare a vital resource for electronicsfirms; thriving semiconductor companies havebeen built around the talents of three or fourengineers.

But people are only the starting point. Howtalents are developed, skills utilized, dependslargely on management: managers shape theorganization, decide on policies, set the styleand tone. The sections that follow examinehuman resources as a factor in competitive-ness, primarily from the standpoint of elec-tronics in the United States. Matters of educa-tion and training are followed by an examine..tion of management practices. One of the ques-tions addressed is: To what extent does thevogue for Japanese management represent any-thing-new and different in the American con-text, as opposed to a reemphasis of themes thathave always been present? The comparisonson education also focus on Japan, in part be-cause of the recent publicity given to that coun-try's lead over the United States in numbersof engineers graduated.

Such topics are particularly appropriate ata time when rates of productivity growth haveslowed in the United States. Is the educationand training of American workers appropriate

for technology-intensive industries like elec-tronics? Do managements make the best useof the talents and abilities of the labor force?Are countries like Japan doing anything thatis really differentor better? In the early partof the century, these .questions were, alreadybeing asked, as part of the "scientific" studyof management. It is no coincidence thatAmerican management experts schooled Japa-nese executives now known for their dedir!a.tion to quality (ch. 6).

The popular press tends to oversimplify theset of issues covered by "human resources."Some commentators define human resourcesnarrowly, as encompassing the skills and at-titudes of the work force; this approach oftenleads to stereotyping of employees in countrieslike Japan Or West Germany. Seeing the Japa-nese worker as the product of a culture thatrewards hard work and diligence captures partof the truth but obscures the larger institutionaland economic context. Others stress rtianagii-ment techniques, often narrowly defined, as F.key to labor productivity. Quality controlcies are the best-publicized current exam;.!?.While certainly critical in the utilization cfirm's human resources, management shouldalso be viewed as part of a broader picture.Management practices themselves reflect amix of schooling and experience shaped by t-th.structure of work and organization within asociety.

This chapter views the American labor forceand the electronics industryfrom twodamental perspectives. First, workers brir48with them a set of skills largely acquired pcitirto joining a company. The question then is tocompare education and training in the UnitedStatesparticularly of white-collar personnel,

305 301 .

302 Internation veness in Electronics

but also blue-collar employeeswith that of themen and women who staff foreign electronicsfirms. Second: Will there be enough appropri-ately trained people to meet the needs of a rap-idly expanding U.S. electronics industry?

Labs': -.T.obility is a separate but related is-sue. A growing industry, such as semiconduc-tor manufacturing, may be able to meet itsmanpower needs by attracting workers fromother parts of the economy. Within the in-dustry, one semiconductor firm may be ableto lure employees from its competitors. Mobili-ty has traditionally been high in the UnitedStates for those with knowledge and experi-ence.

But what of those left behind by technologi-cal change? To a considerable extent, other na-tions have used retraining programs as, instru-ments of public policy for enhancing employeemobility and aiding those whose skills are out-of-date. This has been less common in theUnited States, where mobility and continuingeducation depend on individual initiative.Leaving aside questions of remedial educationand the training necessary for entry level jobs,

with which the United States has experimentedlargely for reasons of social welfare, a strongcase can be made, for an enhanced Federal rolein training and retraining programs to supportthe competitiveness of growing high-technol-ogy industries like electronics.

The other perspective on human resourcesin this chapter relates to corporate manage-ment. Contrasting the practices of Japaneseand American managers shows many of thelessons of effective management to be univer-sal, the unique character of Japanese manage-ment sorneffiing of a myth. Nonetheless, thereare lessons to be learned from firms in Japan,as well as from successful organizations in theUnited States. Competitive firms here andabroad tend to share a common trait: man-agement practices that give employees a sayindecisions affecting their work, along withsupport for skill development. Emphasis onemployee participation and human relationscan contribute to productivity and worker sat-' isfaction, but conclusive evidence linking par-

, ticular management techniques (such as qualitycontrol circles)here or in any countrytocompetitive success is conspicuously lacking.

Education and TrainingThe U.S. electronics industry is built on the

capabilities of production workers, skilled tech-nicians, and white-collar managers and profes-sionals. On the shop floor, blue-collar employ-ees operate semiconductor fabrication equip-ment, assemble computers or TV sets. Muchof this work is essentially unslcilled, meaningthat a typical job can be learned in a few hours.Techniciansgrey-collar employeesoftenwith vocational school training, play an impor-tant role both on the factory floor and in re-search and development (R&D) laboratories.They maintain, troubleshoot, and repair sophis-ticated equipmentand sometimes fabricateitas well as testing and inspecting com-ponents and systems. Technicians also buildand help develop prototypes of new products.Other employees with specialized skills include

draftsmen and nondegree designers, 'produc-tion forelnen, field service installers and repair-men, computer system operators, and techni-cal writers. White-collar workersmany withcollege degreesperform functions rangingfrom plant management to accounting and fi-nancial control, business planning, and legaladvising. Engineers and scientistssome withadvanced degreesdesign and develop prod-ucts, plan manufacturing processes, specifyproduction equipment, and carry out R&D proj-ects in fields ranging from solid-state physicsto computer architectures. All of these skillsare essential to a competitive industry, not justthose of the well-educated and well-paid pro-fessionals; grey-collar technical workers, inparticular, have a critical place in technol-ogy-based organizations. Some jobs depend

Ch. 8Human Resources: Educatioa, Training, Management 303

much more heavily on formal education andtraining than others, but it is fair to say thatbetter skills and abilities at all levels will addto the competitive ability of an enterprise, aswell as adding to peoples' upward mobility.

The United States has maintained a lead inmany fields of technology and science sinceWorld War II, in large part because of theexcellence of the educational system here.Nonetheless, other advanced industrial nationsprovide their work forces with training in tech-nology, mathematics, and science that on theaverage is probably more intensive. It is easyto forget, in the publicity that surrounds NobelPrizes, the Apollo program, or the nascent bio-technology industry, that competitiveness restson the skills and abilities of great numbers ofpeople whose contributions will never be pub-alicized or even acknowledged. At a time whenliteracy levels in the United States decline asthose elsewhere rise, and the Soviet Uniongraduates five times as many engineers, itmakes sense to look at the fcundations for theNation's human resources as well as the pin-nacles of its achievements.

In fact, the evidence of U.S. weakness intechnical education and training is strong andcontinuing to mount.1 The best people and besteducational institutions in the United States areprobably as good as ever, maybe beter. But thebreadth of capability that once distinguishedthe U.S. labor force may be diminishing. TheNational Science Foundation/Department ofEducation (NSF/DOE) report cited above con-cludes that American achievements in basic re-search remain unchallenged, but that the aver-age high school or college graduate in thiscountry has only the most rudimentary knowl-edge of mathematics or science., The trends are

,"Science and Engineering Education for the 1980's and Be-yond." National Science Foundation and Department of Educa-tion. October 1980. See also Today's Problems. Tomorrow'sCrises: A Report of the Natiohal Science Board Commission onPrecollege Education in Mathematics. Science and Technology(Washington. D.C.: National Science Foundation, Oct. 18. 1982):Science and Engineering Education: Data and Information, NSF82.30 (Washington, D.C.: National Science Foundation. 1982),.and Science Indicx ',r5-1980 (Washington. D.C.: NationalScience Board. Natio, Science Foundation, 1981), chs. 1 and5. The U.S.-U.S.S.R. comparison in engineeringswaduates comesfrom p. 209 of the last-mentioned report. /

clear, beginning at secondary levels wherestudents avoid courses in these subjecto. Onlyone-sixth of U.S. secondary school stu( -nts, forexample, take courses in science or mathemat-ics past the 10th grade. Technology, as opposedto science, is totally lacking in secondaryschools, despite the abundant evido:ace of pub-lic fascination with technological achieve-ments. Indeed, few people seem to distinguishtechnology from science, hence misnomerssuch as science fiction.

The NSF/DOE report. along with manyothers, also points to apparent shortages ofentry-level computer professionals and severaltypes of engineers, and the difficulties of sec-ondary schools, vocational institutes, commu-nity colleges, and universities in finding andretaining qualified teachers in the physical sci-ences, mathematics, engineering and computerscience-, and in vocational programs. More-over, equipment used for teaching laboratorycourses in engineering and the sciences isyears out of date and in short supply. In thefuture, American industry, T ,cularly high-technology sectors like electr,, A;s, may sim-ply not have an adequate supply of employeeswith the kinds of skills needed 1,) inaintain U.S.competitiveness.

U. Secondary School Educationin Science and Mathematics

Falling mathematics, and science enrollmentsin American high schools indicate that, whilethere is a small group of students who wantand get advanced courses, the great majorityavoid these subjects when they can. Averagescores on national tests of achievement inmathematics and the sciences are lower thana decade ago. Students who elect to take Ad-vanced Placement Tests in science or mathe-matics make about the same scores as in thepast, indicating that the core of serious stu-dents gets good preparation; but overall, Scho-lastic Aptitude Test (SAT) scores fell for 18 con-secutive years until holding steady in 1981.*

According to the Educational Testing Service, Princeton. N.J.,mean scores in 1981 for co./lege-bound high school seniors were424 for the verbal portion of the SAT and 466 for the mathematics

301


Some of the de ;line can be attributed to thegreater percentage of students who now attendcollege and thus take .tie tests, but an advisory'panel convened to examine the SAT concludedthat, since 1970, other factorsincluding lowereducational standards and diminishing motiva-tion on the part of studentshave been muchmore important.2

Fewer American high school students areelecting mathematics and science courses, par-ticularly the two fundamental physical sci-ences, chemistry and physics; of those who doelect science, more chose the life sciences.While the majority of U.S. high school grad-uates have taken biology, only about a thirdhave had cheinistry; the fraction drops to aboutone-tenth for physics.3 The situation is replicated in high school mathematics, where onlyOne-third of U.S. graduates take 3 years ofcoursework. Regardless of how good theirgrades may be, three-quarters of Americanhigh school graduates do not have the prerequi-site ,bourses to enter a college engineeringprogram.4 What this means for industries likeelectronics is not only that the average i';(711school graduate is unprepared to v

neering or one of the physical :.)ciences ,n col-lege, but may be unable to enter a career call-ing for middle-level technical skills without agood deal of additional training.

Secondary Schooling Abroad,Especially in Japan

U.S. enrollments in science and mathematicscontrast starkly with the picture in japan. Notonly do about 90 percent of Japanese highschool students graduatecompared with 75to 80 percent in this countryhut all are re-

(footnote continued from p. 303)pertion, identical to 1980 scores. In 1966. the means were 466for verbal and 492 for mathematics. While testing criteria maynot have remained precisely the same over this period, the down-ward trend is unambiguo:Is.

2"Science and Engineering Education for the 1980's and Be-yond,- op. cit.. pp. 107-108.

'P. D. Hurd. "Falling Behind in Math and Science," Washing.:ton Post, May 16, 1982, p. C7. See also Science and Engineer-ing Education: Data and Information. op. cit., pp. 57, 59.

"Engineering: Education. Supply/Demand and Job Opportu-nities." Electronic Industries Association, Washington. D.C., Oc-tober 1982.

30d

quired to complete 2 years of mathematics plus2 years of science. Competition for entry intothe best colleges is intense; Japanese studentschoose rigorous electives and spend muchmore time on homework than their Americancounterparts. Those who wish to attend collegestudy mathematics each year, moving beyondtrigonometrythe point where many U.S. highschool curricula still stop.5 The stress in Japa-nese secondary schools on science and mathe-matics for all students is far from unique. TheSoviet equivalent of the. American high schoolcurriculum includes a heavy dose of course-work in these areasfor instance, 2 years ofcalculus. West German secondary school stu-dents, even those who wish to specialize infields such as the classics or modern languages,get extensive training in mathematics and sci-ence; by the same token, those planning tech-nical careers receive their liberal arts educa-tion in high school. Neither curricula, nor aca-demic standards vary as widely among WestGerman schools as in the United States.°

In Japan, large numbers of students v. ho donot go to college get technical, vocational, orsemiprofessional schooling as preparation forjobs in industry where they will work with andprovide support for engineers and scientists.The result is a large pool of well-prepared can-didates for entry-level grey-collar jobs.?

The investments that students in Japan makein science and mathematics yield measurablebenefits. On a number of international achieve-ment tests, Japanese students score consistent-ly above their counterparts in other industrialnations.° Nonetheless, secondary education in

°M. W. Kirst, "Japanese Education: Its Implications for Eco-nomic Competition in the 1980's." Phi Delta Kappan, June 1981,p. 707. Only about 30 percent of U.S. high schools offer cLIculus,and fewer than 10 percent of American high school studentstake the subject; see Hurd. op. cit.. and Science and Engineer-ing Education: Data and Information. op. cit., p. 59.

6Engineering Our Future: Report of the Committee of Inquiryinto the Engineering Profession (London: Her Majesty's Station-ery Office. January 1980), p. 219. Also. D. W. Sallet, "Educationof the Diplom Inge-nieur," Journal of Engineering Education,vol. 59, June 1969, p. 1105.

'S. B. Levine and I Kawada, Human Resources in JapaneseIndustrial Developmenat (Princeton, N.J.: Princeton UniversityPress, 1980), pp. 74, 80. Engineers in Japan are evidently sup-ported by many more technicians than in the United States.-

°R. S. Anderson, Educatit.:1 in Japan (Washington. D.C.: U.S.Government Printing Office, 1975), p. 130.

Ch. 8Human Resources: Education, Training, Management 305

japan has major weaknesses. The most obviousis the strong traditional emphasis on rote learn-ing and imitation, coupled with a depenr!enceon textbooks and lectures rather than demon-stration and learning-by-doing (in reality, U.S.education is probably no better in this regard).Critics of the system argue that this stunts thedeveloprnen' of creative abilities.° Academiccompetition L-t hpan is, furthermore, so in-tense that the Jar nese. Ministryof Education

exprc3sed concern that other aspects ofchild development are being negleCted.- Despitethe undoubted validity of some of these Criti7cisms, the fact remains that high school stu-dents in Japan receive training in science andmathematics that is, on average, more exten-sive than in the United States. Even for stu-dents who do not go on to technical or profes-sional jobs, such training contributes to quanti-tative skills, precision in thinking, and to anunderstanding of the physical world. ,;.it;11 ttbackgroun, 1 helps people to comprehend thetechnologies that their dEily lives depend 'nn.In the future, their employment opportunitiesmay depend on this as well.

University and Continuing Educationin the United States

In some respects, the Japanese and Americaneducational systems are opposites. The Japa-nese concentrate their efforts on precollegetraining where the United States is weak. Onthe other hand, the quality of university educu-tion in Japan is much inferior. In a very realsense, the American system of higher educa-tion must compensate for secondary schoolingthat is generally poor.

Although this comparison may be qualitative-ly valid, it begins to break down in terms ofnumbers. While the United States continues toproduce more Ph. D.s in science and mathe-matics than Japan, Japanese undergraduate

°See, for example, the assessment of M. IV3gai, former JapaneseMinister for Education: "Higher Education in Japan," JapanQuarterly, vol. 24, 1977. p. 308. While many Japanese are quiteself-conscioLs about their country's supposed lack of innova-tion and originality in engineering and the sciences, the prod-uct developmeids flowing in recent years from Japan's industriesshow great creativity in the application of technology.

programs have been turning out greater num-bers of engineers since 1967. In 1981, Ja-pan graduated 75,000 engineers compared to63,000 here, despite a population half that ofthe United States. The margin is a little greaterfor electrical engineering graduates-25 per-cent.1°

As figure 53 shows, the United States onceheld a commanding lead in the proportion ofengineers and scientists in the work force.While the advantage over other Western na-tions probably still exists (various countries cat-egorize scientists, engineers, and techniciansdifterently, making comparisons. ambiguous),it has narrowed greatly, And, as table 67demonstrates, engineering graduates are nowa smaller proportion of their age group in firUnited States than in Japan or West Ger-Inatlycountries where a far greater fractionof engineers in any case devote their efforts tocommercial rather than defense industries.

'°The 1981 breakdown by disciplines is not available fc,r Japan,but in 1980, 19,355 B.S level degrees were awarded in Electricaland computer engineering, compared to 15,410 in the UnitedStates. Figures for Japan are from the Ministry of Education,those for the United States from P. Doigan, "Engineering andTechnology Degrees, 1981," Engineering Education, April 1982,p. 704, and P. Sheridan, "Engineering and Technology Degrees,1980," Engineering Education, April 1981, p. 713.

Figure 53.R&D Engineers and Scientistsin the Labor Force

U,

"2 1422°3ovr-3, 0.7

0"2 0.6

N CDfr; 0.5a) C

Em0a .4

a) cr 0.3"6 .ECM .0 0.2Cha

0.1CD CM

1-2`mCD

a- 1967

West Germany

1969 1971 1973 1975Year

1977 1979 1981

aLower bound estimate.

SOURCE: National Patterns of Science and Technology Resources, 1982(Washington, DC: National Science Foundation. 1984 P. 33.

3 0J


Table 67.Engineering Graduates as a Percentageof Their. Age Groups

United States 1.6%Japan 4.2West Germany' 2.3France 1.3United Kingdom 1,7

a First devee graduates, Including foreign nationals, In 1978, except for WestGerrn.Jiny and France, where the percentages refer to 1977. In the United States,a significant fraction of engineering graduates are from overseas: in 1982, 8 per-cent of B S. degrees in engineering wont to foreign students, 29 percent of M.S.iiiigiees, and 40 percent of Ph. D. degrees. See P. J. Sheridan,'"Engineering andTechnology Degrees. 1982," Engineering Education. April 1983, P. 715.

SOURCE: Engineering Our Future: Report of the Committee of Inquiry into theEngineering Profession (London: Her Majesty's Stationery Office.January 19801, p. 83.

Engineering EluctitionAs table 68 indicates, graduates in engineer-

ing, the physical sc:iences, and mathematics inthe United States accounted for steadily fall-ing proportions of new degrees at both under-graduate and graduate levels during the 1970's.The number of degrees in the mathematical sci-ences, including statistics and computer sci-ence, actually fell between 1970 and 1980.

In engineering, undergraduate enrollmentshave jumped since the mid-1970'sand thenumber of graduates has followed, as shownin figure 54leading to overcrowded classes,overloaded faculty, and severe pressures on thequality of education. The number of full-timeundergraduates enrolled in U.S. engineeringschools went from about 20,000 in the early1970's town all-time high of more than 400,000

70.000

60.000

50,000

73 40.000

30,000

n'

20 000

(:.&00

01965 1970

Year

Figure 54.Engineering Graduates ofAmerican Universities

Doctoral degrees

1t:5 1980 1982

SOURCES: 1965-79"Data Related to the Crisis in Engineering Education,"American Association of Engineering Societies, March 1981, p.17.

1980P. J. Sheridan, "Engineering and Technology Degrees, 1980,"Engineering Education, April 1981, p. 713.

1981P. Dolgan, "Engineering and Technology Degrees, 1981,"'7ngineering Education, April 1982, p. 704.

198ZP. J. Sheridan, "Engineering and Technology Degrees, 1992,"Engineering Education, April 1983, p. 715.

in 1982." At the graduate level, the trends arequite differentbut not encouraging. The num-ber of master's degrees in engineering has in-creased slightly over the past decade, but thenumb.n. of Ph. D.s has declinedone reasonfor faculty shortages in engineering schools..Figure 54 illustrates the trends at both B.S. and

Doigan, "Engineering Enrollments, Fall 1982," EnginCer-ins.; Education, October 1983, p, 18. At the bottom of the mostrecent trough, in 1973, 187,000 students were enrolled in engiveering; by 1982, the total was 403,000.

Table 68.U.S. Degrees Awarded by Field

EngineeringPhysicalsciences Mathematicsa

Total as percentageof degrees awarded

in all fields

1960: B.S. 37,808 16,057 11,437 175/D

M.S 6,989 3,387 1,765 17

Ph. D. 786 NA NA NA

1970: B.S. 42,966 21,551 29,109 12%M.S 15,548 5,948 7,107 13

Ph. D. 3,620 4,400 1,222 31

1980: B.S. 58.742 23,661 22,686 10%M.S 17,243 5,233 6,515 10

Ph. D. 2,751 3,151 963 21

NA ., Not Available.alncluding statistics and computer science.SOURCE: EngineeringData Related to the Crisis in Engineering Education," American Association of Engineering Societies,

March 1981, p. 17; Physical Sciences and MathematicsNational Patterns of Science and Technology Resources1981, NSF 81-311 (Washington, D.C.: National Science Foundation, 1981j, pp. 78-80.

, t 1


Ph. D. levels. Not only have doctoral enroll-ments failed to keep up, but about half of allPh. D. engineering candidates are now foreignnationals; many of them leave the United Statesafter graduation.*

An important cause of declining enrollmentsof Ph. D. candidates in engineering has beanthe high starting salaries that holders of newbachelor's degrees commandin 1082, about$26,000. Swelling demand by industry for engi-neers has attracted undergraduates to the field,at the same time siphoning many off from thepool of prospective graduate students. Tosomeone who might otherwise consider aPh. D. followed by a teaching career, the re-wards of immediate employment can seemmuch more attractive than several years of low-paying stipends or graduate assistantships,then the salary of a junior faculty member.While pay for college teachers has always beenwell below that in industry, the other attrac-tions of an academic career have diminishedin these days of overcrowded classrooms, out-dated equipment, and limited research funding.

Poor facilities and an escalating student-to-faculty ratio are leading to declines in the qual-ity of education provided in American engi-neering schools. For many years, the propor-tion of programs in engineering and computerscience that were unconditionally reaccreditedduring periodic reviews held steady at about70 percent, but in 1981 only 50 percent of theprograms examined received full accredita-tion.12 This sudden change indicates the gravityof the problems facing engineering educationin the United States.

The most common and most serious causesof declining educational quality are facultyshortages and obsolete laboratory equipment.

'See note to table 67. In 1982, 1,167 of 2,887 engineeringPh. D.s went to foreign nationals; both industry and universitieshave b come heavily dependent on foreign-born engineers, es-pecially at the doctoral level. Figures on graduates reflect earlierenrollments; currently, nearly 50 percent of Ph. D. candidatesin U.S. engineering schools are foreign nationals.

""Adequacy of U.S. Engineering Education." Emerging Issuesin Science and Technology, 1981 (Washington. D.C.: NationalScience Foundation. June 1982), p. 60. Programs with deficien-cies may be reexamined after a shorter than normal interval orplaced on probation.


Engineer holding bracket designed withcomputer assistance

Even when funds hive been available to hirenew faculty, good candidates are rare becauseof the low numbers of new Ph. D.s and the un-competitive salaries offered by universities.Estimates of the number of unfilled facultypositions in U.S. engineering schools have beenin the range of 1,400 to 2,000about 10 per-cent of the total number of faculty positions inengineering." Furthermore, universities can nolonger depend on graduate students to relievesome of the load on regular faculty by assistingin classroom teaching and laboratory instruc-tion.

The equipment problem is equally serious.While faculties do their best with the resourcesavailable, it is difficult to teach a digital designlaboratory with equipment from the analog era.And laboratories, as well as classrooms, havebecome overcrowded as undergraduate enroll-ments have climbed. Quality suffers when stu-dents have less contact with faculty, as well asless exposure to up-to-date laboratory equip-ment and computing facilities. Many univer-

"As of the fall of 1982. a survey of U.S. engineering schoolsreported 1.400 authorized and budgeted faculty positions vacant.of a total of about 18,000. The number should be regarded asa lower bound because few universities have increased the num-ber of authorized faculty positions at rates commensurate withgrowth in undergraduate enrollments. The most severe problemsare in computer specialties. See J. W. Geils, "The Faculty Short-age: The 1982 Survey." Engineering Education, October 1983.p. 47.

311


cities, hurt by past slumps in engineering en-rollir.,.mts, are reluctant to put scarce funds into,expansion to meet what may be a transient de-mand. More fundamentally, universities havehad great difficulty in adjusting to shifting stu-dent choices at a time when total enrollmentshave stopped rising. Tight budgets have causedprograms in the sciences as well as engineer-ing to fall behind the times."

The well-publicized situations at large, State-supported schools such as Iowa State Univer-sity and the University of Illinois, typical of theinstitutions that form the core of the U.S. sys-tem of engineering education, are representa-tive.15 Iowa State simply ran out of facilities tohandle enrollment increases in computer engi-neering, despite operatiny on a 6-day schedule.Because of overcrowding, students werewarned that they might not be able to completetheir programs in 4 years. Transfer studentsat Illinois must have a grade-point average of4.2 on a scale of 5 to enter engineering, whilethe universitywide requirement is only 3.25.Shortages of facilities and teaching facultyforced 16 of 30 large American engineeringschools to adopt some form of restriction onthe number of students they admit."

Only the elite universities have been largelyspared such problems, and even these have hadtrouble attracting enough good graduate stu-dents. But because the best schools have always

""Science and Engineering Education for the 1980's and Be-yond," op. cit., pp. 68-69. Courses in physics and chemistry alsodepend on laboratory and computer facilities. For a discussionof laboratory equipment shortages with the emphasis on researchneeds, see "Obsolesrmnce of Scientific Instrumentation in Re-search Universities," Emerging Issues in Science and Technol-ogy, 1981, op. cit.. p. 49.

The nine State-supported engineering schools in Texas havereported equipment needs totaling $88 million, about 70 percentof this for undergraduate teaching laboratories. The situationin Texas is probably fairly typical; an extrapolation to Ole UnitedStates as a whole risults in an estimate of about $1 billion fornew laboratory equipment in engineering alone. See "$1 Billionfor IrMructional Equipment," Engineering Education News,June 1982, p. 1.

""Engineering Education Under Stress," Science,1Sept. 25.1981. p. 1479; C. Phillips, "Universities in U.S. LosingGround in Computer Education," Wall Street Jan. 14,1983. p. 1.

""Universities Limiting Engineering Enrollments," Engineer-ing Education News, March 1981. The limitation's are based sim-ply on numhers; as at Illinois, qualified student. are being turnedaway.

limited their enrollments, they have been ableto raise Ale average quality of incoming stu-dents while keeping expansion to manageablerates. Engineering departments at schools likeStanford or MIT have also been able to retaintheir faculties. One of the dangers implicit inresponses by industry or the Federal Govern-ment to the problems afflicting engineeringeducation is that resources may flow dispro-portionately to the top-ranked, research-ori-ented universities. Of the nearly 300 collegesand universities that offer engineering in theUnited States, it is the middle tierboth publicand privatethat turns c,-ut the vast majorityof graduates and faces the most serious prob-lems.

Supply and Demand

Even though enrollments are still climbing,and the number of B.S. graduates in engineer-ing has been going up at about 10 percent peryear (fig. 54), it is not at all certain that thenumber of engineering graduates in the UnitedStates will meet future needs. As discussed inmore detail later in the chapter, there will al-most certainly be entry-level shortages at sometimes in some specialtiese.g., computer engi-neeringand the shortfall in Ph. as for teach-ing is bound to continue; according to one esti-mate, there is a current shortage of 3,500 doc-toral-level engineers in industry beyond that ofPh. D.s for university faculties.17

While the rapid rise in engine ering enroll-ments has led to fears by some that the UnitedStates might be headed for an oversupply bythe 1990's, such concerns seem overstated ifonly because many graduates of engineeringprograms move on to other fields. Competentengineers have virtually always been employ-able in the United States, regardless of econom-ic conditions. Nevertheless, the American laborforce contains nearly 11/2 million engineers,and some portions of the engineering community deny the reality of the current "shortage,"claiming that what industry really wants is alarge pool of entry-level people to help keep

""National Engineering Action Conference." EngMe.?ringEducation News, April 197,2, p. 1.

Ch. 8--Human Resources: Education, Training, Management 309

salaries of mitkareer engineers low. There isa good deal of truth to this. Entry-level short-ages arise in part because employers prefer toIliNe new engineers with fresh skills at lowerpay. This is an easier and perhaps cheaper wayof meeting their needs than coupling theexperience of midcareer engineersmany ofwhom find themselves with increasingly ob-solescent skillswith well-designed continuingeducation programs.

Regardless, at least some specialties seembound to face continuing shortages by almostany criterion. These 'specialties include anumber that are particularly relevant for thefuture competitiveness of the U.S. electronicsindustry; most notably, entry-level computerprofessionals are expected to be in high de-mand well into the 1990's. Programs of instruc-tion in computer science and engineering stilltend to be small and underdeveloped. Some arein engineering schoolsoften within electricalengineering departmentsothers in schools ofarts and sciences, where computer science maybe associated with mathematics departments.Many teaching departments lack the criticalmass that would help them thrive, not surpris-ing in a field which did not exist 25 years ago.In computer science, the United States gradu-ates only 250 Ph. D.s each year, a numberWhich has been decliningone reason comput-er science and engineering faculties are suffer-ing greater proportional shortages of facultythan (other) engineering departments." Atpresent, qualified software engineers are inshort supply; although people with many typesof training can fill jobs as applications pro-gramers, there are far fewer candidates for jobsin the design and development of computer-based systems themselves.

Other new and/or specialized fields suffersimilar problems. Perhaps half-a-dozen Ameri-can universities have the facilities needed todesign and fabricate large-scale integrated cir-

"Seventeen percent of faculty positions in computer scienceand engineering were vacant at the beginning of 1982, versusabout 9 J:cent for engineering as a whole. See "Universitiesin U.S. Are Losing Ground in Computer Education," op. cit.;"The Faculty Shortage: The 1982 Survey," op. cit.

suits. Microprocessor applications coursesmay require equipment that schools cannot af-ford. Few universities have adequate resourcesfor computer-aided design in any of the fieldsof engineering. At the same time, such diffi-culties can be viewed as similar to those thathave always existed. It has never been easy togive students a sense of the development effortthat goes into an airplane or a nuclear power-plant. In this sense, the adaptations requiredby the emergence of large-scale integrated cir-

.cuits or cheap computing power are nothingnew.

Industry Initiatives

To help meet the needs of their members, twoof the trade associations in electronics haveestablished programs to support engineeringeducation. The American ElectrOnics Associa-tion has asked for money to augment facultysalaries and establish chairs in electrical engi-neering, as well as to expand fellowship pro-grams for students; the Semiconductor Indus-try Association is funding research, thus pro-viding indirect support to both students andfaculty members through stipends and salaries,as well as money for equipment."

A different approach has been taken byWang Laboratories, which manufactures mini-computers and office automation equipment.The Wang Institute, located near Boston, of-fers a master's degree in software engineeringthrough its School of Information Technology.Initially endowed by An Wang, the company'sfounder, the Institute is now an independent,nonprofit organization. With a curriculum de-signed to give training both in the technicalaspects of computer software and in planning,management, and human relations, the schoolwhich graduated its first students in :1982grew directly out of the inability of companieslike Wang to meet their personnel needs. Tui-tion is about $8,000 per year, less than half theactual cost of the program; the difference is

""AEA, SIA to Vie in Fund-Raising Efforts," Electronic News.Nov. 16. 1981, p. 36. Companies in other industries have bbgunparallel efforts.

313


covered by endowments and cont1'ibutions.20Because of the emphasis on job-related skills,candidates for admission must have at least 2years of professional experience, in - additionto a B.S. degree in an appropriate field. TheWang Institute is one of a number of experi-ments presently underway in nontraditionaltraining in specialized technical fields.

University-Industry RelationsDespite these and other examples of new and

close relationships between business and edu-cational institutionsfor instance, the indus-try-supported Center for Integrated Systems atStanforduniversity-industry relations, in theUnited States as in most countries, tend to beuneasy. Tensions between the theoretical learn-ings of faculty members and the more practicalconcerns of private firms, particularly smallercompanies that do not engage in much re-search, are common. This also holds for pro-fessions such as businos administration; tosome extent it applies to the sciences as well.

In engineering, these tensions have deep his-torical roots; by the first decade of the 20th cen-tury, the academic perspective had largely wonout over the shopfloor orientation that manyin industry had advocated.21 Later, between thewars, U.S. engineering education began to stag-nate. During World War II, numerous R&D ef-fortsincluding many in electronicswerespearheaded by scientists (particularly physi-cists) with engineers filling subordinate roles.This lesson was one of several pointing to theneed for reevaluations of engineering educa-tion.

The resulting turn toward theory led to "engi-neering science" as the core of undergraduatecurricula. In the post-Sputnik period beginningin the late 1950's, engineering schools empha-sized quantitative, analytical skills even more.Accompanied by a strengthening of mathemat-

toInformatiqn supplied by Wang Institute. See also M. A.Bengs. "A Unique Institution for High-Tech Training." BostonGlobe, Mar. 2, 1980. and "Institutionalizing the Students of Soft-ware." Computer Design, December 1982. p. 189.

Calvert, The Mechanical Engineer in America. 1830 -1910:Professional Cultures in Conflict (Baltimore. Md.: Johns HopkinsUniversity Press. 1967).

ics requirements, the focus on engineering sci-ence came at the expense of engineering de-sign, as well as manufacturing and productionprocesses. This was also a time when thespread of digital computers made numericalsolutions to many previously intractable prob-lems a reality, further strengthening the move-ment toward analysis at the expense of synthe-sis. In recent year there has been somethingof a swing back.

Industry has always wanted graduates whocan go to work immediately, while acknowledging tha virtues of theoretical preparation inthe engineering sciences as preparation for ad-vancement and for continuing education. De-mand for the "old-fashioned," practically ori-ented engineer has led to a proliferation of cur-ricula in what isusually called engineeringtechnology.

Engineering TechnologyTechnology Programssome 2 years in

length and leading to an associate degree,others full 4-year B.S. courses of studyrepre-sent an attempt by American colleges and uni-versities to equip entry-level employees withimmediately applicable job skills. Graduates ofthese programsmore than 26,000 at the asso-ciate and bachelor's levels in 1981 (40 percentof the total in engineering)can be thought ofas paraengineers; they get less extensive andless rigorous training in mathematics, the sci-ences, and in engineering science, but consid-erable exposure to routine technical prob-lems.22 While B.S. technology graduates are bet-

"in 1976, 16,685 associate degrees and 5.721 B.S. degrees inengineering technology were granted in the United States: in1982, the figures were 17.198 for associate degrees, and 8.325at the bachelor's level. The 106 data are from "Engineering andTechnology Degrees, 1976," the 1982 from "Engineering andTechnology Degrees, 1982," op. cit.

Well over 200 schools have technology programs, about thesame number as for engineering. of the associate degreesare awarded by community college and vocational-technicalschools. In a university, B.S. programs in technology and engi-neering may be offered by different colleges, particularly wherethe technology curriculum has grown out of an industrial artssetting: alternatively, both programs may be found in a "Col-lege of Engineering and Technology." Many faculty membersin engineering have resisted the movement mc'vard technologyeducation. feeling that it detracts from the proiBssion and threat-ens their own image. Outside the community of educators, a good


ter prepared for advancement and for creativework than technicians, their upward mobilityis considerably less than for engineers. In atleast one sense, the problems that have hit en-gineering education are more serious still fortechnology: the practical, hands-on experiencethat these programs seek to provide dependsheavily on equipment similar to that actuallyused in industry.

Community Colleges and Local InitiativesIn recent years, community colleges have ex-

panded more rapidly than any other segmentof higher education. Many offer engineeringtechnology, as well as preengineering pro-,grams that send students on to universities.Moreover, - ommunity or junior colleges andvocational-technical schools train many of thetechnicians who take jobs in U.S. industrieslike electronics. While the number of studentsearning associate degrees in technical fieldshas grown in the last decade, there is little in-formation on the quality of these programs.

deal of confusion persists concerning the role and functilea vFtechnology educationnot surprising when associate programsgraduate men and women trained essentially as techni.zians,While a B.S.-level technologist is much closer to an engineer.

Differences in academic standards among technology pro-grams may be even greater than in engineering. In contrast tocountries like West Germany, where all I:echnical universitiesare held to similar standards, ';uality in American engineeringand technology programs var,r;s widely, even among those thatare fully accredited.

Photo credit: © Ted Spiegel, 1983

Computer-assisted test to determine design dataon metal fatigue

Some offer up-to-date training in needed spe-cialties, while others are accused of turning outpeople for jobs that have already disappeared.

Public 2-year and community colleges facechronic problems in funding their programsand retaining faculty.23 Even in Silicon Valley(the area near San Francisco where so manyelectronics firms have located)which is nowgetting a great deal of attention as a model forindustrial developmentthese institutionshave never been well- integrated into the localenvironment, and seem isolated from industryas well as from the mainstream 3f university -oriented education.24 Despite she concentrationof E.ciectronics companies, the six communitycolleges in the area have faced severe shortagesof equipment for student laboratories, and arelationship with industry in which each groupseems generally supportive of the other but inwhich the various parties do not always man-age to communicate or cooperate effectively.

Only a few States or localities have thus farattempted to meet needs for technology-based

.,"Science and Engineering Education for the 1980's and Be-yond," op. cit.. p. 93.

"E. L. Useem. "Education and High Technology Industry; TheCase of Silicon Valley," Institute for the Interdisciplinary Studyof Education. Northeastern University, Boston, Mass., Septem-ber 1931, pp. 12-18. Useern's study, based on more than 100 inter-views, finds t'iat neither educators nor companies are respond-ing very well to the region's educational needsparticularly at.the secondary level.

I LC.r

. .f t's

_...... _1 1._._ La so 00 1 7." I 5. Z.1.5 331

tc..tto III 1I4 E-02

Photo credit: © Ted Spiegel, 1983

Results of test shown In adjacent photograph. Manyengineering schools have had difficulty In acquiring

modern equipment of this type

315


education and training outside their college oruniversity systems. North Carolina has estab-lished a Microelectronics Center linking sev-eral universities and industry at Research Tri-angle Park near Raleigh; the State has also setup the North Carolina School of Science and.Mathematics for high school students with un-usual ability. But this is more the exceptionthan the rule, and budgetary constraints inmany localities may limit further developmentof nontraditional alternatives.

Continuing EducationOngoing education and training for engi-

neers and other technical workersincludingthose without degreesis a vast and amor-phous activity. Colleges and universities enrolllarge numbers of students in graduate or con-tinuing education programs, some of whomtake only a Env courses while others activelypursue degrees. Such programs, many ofwhich cater to part-time and evening students,face problems paralleling those of undergradu-ate engineering and science curricula. The lowvisibility and lack of prestige of continuingeducation aggravates the difficulties; facultyturnover is high, quality uncertain. For exam-ple, enrollments in New York University'sadult education courses in computer program-ing have been increasing by 20 to 25 percentper year, but budget limitations have made itdifficult to purchase needed equipment, as wellas to find and keep competent faculty.25 Thissituation is replicated in private and public col-leges and universities throughout the country.

Many professional societies are active in con-tinuing education, principally through shortcoursessometimes offered in conjunctionwith universitieson topics of interest to theirmembers. Current favorites in electronics in-clude microprocessor-based systems, and pro-graming in newer computer languages like Pas-cal. Colleges and universities offer their ownshort courses, as do private, profit-seek' ng en-terprises. Some companies have operatededucational armse.g.,, RCA Institute. Short

z5C. Anders, "Colleges Faltering in Effort to Ease Critical Short-age of Programmera," Wall Street Journal, Aug. 24, 1981, p. 15.

courses and related noncredit programs varywidely in quality and rigor. Some hold to highstandards, others offer little more than can begleaned from trade magazines.

Although many firms offer on-the-job train-ing and continuing education for their engi -.neers, scientists, and technicians, it is impossi-ble to generalize concerning the extent and ef-fectiveness of such efforts. Some companiesallow employees to spend several hours perweek taking college courses on company time;others will pay tuition provided the studentgets a good grade. Some develop in-house pro-grams. Others refuse any assistance, relying en-tirely on individual initiative. Some organiza-tions have a companywide policy covering con-tinuing education; in other cases, decisions areleft with lower level supervisors. In many com-panies, internal training programs are intendedprimarily for new employees; in other in-stances, firms organize or support programsaimed at a broader slice of their work force.While extensive in-house training is most com-mon in large corporations, smaller electronicsfirms have also been turning to such efforts tohelp meet their manpower needs.

Beyond case-by-case insights, the overall di-mensions of company-run training programsin the United States are largely unexplored; theAmerican Society for Training and Develop-ment estimates that business and industry alio,cate some $30 billion to $40 billion a year toeducation and training, but little informationis available on how such moneys are spent, andby whom, or just what is counted in arrivingat the total.26

Looking more narrowly at engineers and sci-entists perhaps 10 to 15 percent of those withat least a bachelor's degree are taking further

"Information from the American Society for Training and De-velopment. Also "Addition to the Record: Statement of AnthonyP. Carnevale, for the American Society for Trairtingnd Develop-ment," Projected Changes in the Economy, Population, LaborMarket, and Work Force, and Their Implications for EconomicDevelopment Policy, hearings, Subcommittee on Economic De-velopment, Committee on Public Works and Transportation,House of Representatives, Nov. 18-19, 1981, p. 233. Carnevalenotes that 35 percent of firms surveyed by the Conference Boardoffer remedial programs in reading, writing, and arithmetic fortheir employees.


academic coursework at any given time.27 Thegreat majority are probably recent graduates,.pith many of these pursuing advanced degreeson a part-time basis.

In electronics, a number of larger U.S. semi-conductor firmsfor example, Intel, which hasalso been a leader in the Semiconductor Indus-try Association's research cooperativehaveinstituted programs of on-the-job training forpersonnel it a variety of levels. Almost allsemiconductor manufacturers evidently pro-vide some training, but the intensity and lengthof the programs vary from company to com-pany. Continuing education for electronics andcomputer engineers in Silicon Valley is readilyavailable for students who can meet the en-trance requirements of schools like Stanfordor the University of California at Berkeley;Stanford has also pioneered interactive televi-sion links enabling it to offer in-plant coursesthroughout the area. However, local electronicscompanieswith a few notable exceptions likeHewlett-Packardhave had little involvementwith secondary-level public education in Sili-con Valley, and most of the interactions withuniversities seem to center on the elite institu-tions where faculty consulting strengthens tieswith industry.28

Semiconductor firms are not alone in offer-ing in- housetraining to their employees. Someelectronics manuf,, cturerc report that, duringtheir initial year on the job, new employeesspend up to 7 hours per week in formal course -work.29 Hughes Aircraft's training program fornew graduates in electrical engineering is espe-cially comprehensive.36 Bell Laboratories andIBM are also well-known for continuing educa-tion and training; low employee turnover atlarge companies like Hughes or IBM favors in-vestment in human resources just as it does in

"The figure was 13 percent for 1978. A somewhat larger num-ber took non-credit courses of various kinds. See "Battelle StudyShows 80 Percent of Organizations Support Continuing Educa-tion for Scientists, Engineers," Information From Battelle, Mar.5, 1980.

"Useem, op. cit., pp. 9-12.""Employment in the U.S. Electronics Industry, Volume I,"

prepared for OTA by Sterling Hobe Corp. under contract No.033-1210, p. 234.

"See R. Connolly, "Companies Still Short of EEs," Electronics,May 19, 1982, p. 105.

99-111 0 ='83 - 21

Japanese electronics firms. In contrast, smallerU.S. semiconductor manufacturers may haveannual personnel turnovers of 35 percent ormore; such companies are understandably re-luctant to devote resources to training men andwomen who may then jump to competingfirms. High turnover rates in electronics holdfor grey-collar technicians as well as engi-neers.31

In contrast, organizations like Bell Labora-tories can finance continuing education for em-ployeeseven send them back to school fulltimewith less fear that people will quit oncethey have a new M.S. or Ph. D. in hand. In part,this is simply because many engineers and sci-entists view Bell Labs as an exciting and pres-tigious place to work; for some, it is more a goalthan a stepping stone. Bell supports part-timegraduate study at local universities, as well asfull-time, on-campus programs at a number ofleading engineering schools; the company alsooffers a variety of in-house training activities,plus a tuition reimbursement plan for employ-ees pursuing undergraduate or vocationaltraining.32

While it is thus true that a considerable num-ber of American engineers and scientists electto take additional academic coursework, manyindividuals make such commitments independ-ently, often with little or no encouragementfrom their employers. Smaller companies, andsome large ones, often feel that they cannot af-ford to support such effortsi.e., that the pay-back would be insufficient. As the NSF/DOEstudy notes:33

At present, continuing education is a frac-tionated, uncoordinated set of operations inwhich academia, industry, professional soci-eties, and individual entrepreneurs pursuetheir own individual paths in response to whatthey perceive as their individual needs. Therehas been virtually no Federal support for con-

"See R. W. Comerford, "Automation Promises To Lighten theFieldService Load," Electronics, Apr. 7, 1982, p. 110. These turn-over rates are good evidence that personnel shortages have beenreal and acute.

""Educational Opportunities at Bell Labs," Education Center,Bell Laboratories, Holmdel, N.J., 1981.

""Science end Engineering Education for the 1980's and Be-yond," op. cit., p. 96.

31?


tinuing education, in part because the cysts ofindustrial programs have been regarded asbusiness expenses.

Because retraining for midcareer engineersand skilled workers will be increasingly impor-tant in the years ahead, particularly in view ofthe aging of the American labor force, the Fed-eral Government may need to reconsider its in-volvement.

The Government already plays an importantrole as direct employer. In 1978, nearly 90,000engineers worked for the Federal Government,about 6 percent of the country's total engineer-ing labor force. Many are employed by the De-partments of Defense and Energy, and the Na-tional Aeronautics and Space Administration.The Air Force alone faces a shortage of over1,000 engineers; if the Nation's defense budgetis to increase during the 1980's as planned, de-mand for engineersboth within the Govern-ment and among defense contractorswillswell even further. Some have argued that engi-neering manpower shortages could jeopardizethe country's security.34 As one response, theDefense Communications Agency is planninga National Science Center for Communicationsand Electronics, intended to help cope with theshortfall in the defense community. Fundedwith the aid of corporate contributions, theCenter, to open in 1983, will develop educa-tion and training courses for participating sec-ondary schools and universities.35

Over the past several decades, nearly half ofall U.S. engineering students have receivedfinancial assistan of /7-.1e sort or another fromthe Federal Government. Funds for laboratoryequipment intended for teaching and research,as well as for curriculum development, havecome from Washingtonfor science and en-gineering, principally through the NationalScience Foundation. Tight Federal budgets foreducation may have the unfortunate conse-quence of shrinking the pool of grduates in

"" "Testimony of Gen. R. T. Marsh, Commander of Air ForceSystems Command," Engineering Manpower Concerns, hearing,Committee on Science and Technology, House of Representa-tives, Oct. 6, 1981, p. 11. Perhaps one-quarter of the country'sengineers work in defense-related fields.

""The NSCCE: A New National Program," Electronics, Dec.29, 1981.

engineering and science, and their quality, ata time when the United States already findsthat it does not haVe enough skilled profes-sionals to staff its commercial industries ormeet its military needs.

University and ContinuingEducation In Japan

If Japanese secondary school students studymathematics and the sciences more extensivelythan their counterparts in the United States,at the university level Japan's educational sys-tem is inferior. Postsecondary education ex-panded rapidly over the postwar period; manyprivate colleges were founded, some with lowstandards. While the small group of elite uni-versities provides more rigorous training, theyhave faced the same criticisms as Japanese sec-ondary schoolsexcessive reliance on rotelearning and the acquisition of facts, ratherthan more general skills in analysis and synthe-sis. In neither science and mathematics, norin engineering, does the quality of university-level education in Japan match that in theUnited States.

EngineeringElectrical engineering students in Japan

spend many hours in the classroom and labora-tory, and take a series of courses rather like thatof Americans, but Japanese companies contin-ue to find graduates unprepared to go to work,while faculty members point to major weak-nesses in curricula.35 Programs in engineeringand computer science leave little room for un-structured or independent learning. Electivesare limited. Students tend to work in groups;according to critics, this fosters conformity atthe expense of creativity and individual initia-tive. The education that Japanese college stu-dents receive outside their technical fields,moreover, is less demanding than here. Defi-ciencies in higher education are among the rea-sons that Japanese companies place great stress

"S. Tubbs, "Electrical Engineering at Kyoto University," Engi-neering Education, May 1982, p. 812; T. Sugano, "Preparationof_New_Electronics Professionals in Japan: Note for Presenta-tions Given at the Japan Society Meetings of May 1, 1981, inPalo Alto, Calif., and of May 4, 1981, in New York."


on internal trainingthey must, simply to bringnew employees up to a satisfactory level ofcompetence.

Far fewer engineering students in Japan goon to the graduate leveleither M.S. or Ph. D.than in the United States. Table 69 illustratesthe stress on undergraduate training and thecomparatively small numbers in graduateschool. In 1980, undergs, ,iduate enrollments inJapanese engineering programs were almostthe same as those in the United States, but thenumber of graduate students was about fourtimes less.37 Those students who do choose toattend graduate school find thatas in under-graduate programscourse work and researchare less rigorous than in American universities.

As the figures in table 69 suggest, while aca-demic competition is keen at secondary levels,with Japanese students vying for places in themost prestigious universities, postgraduatetraining brings few rewards. Because corpora-tions in Japan rely heavily on in-house trainingto impart job-related skills, and because re-search does not have the prestige that it car-ries in the United States or Europe, Japaneseengineers have little incentive to go on to grad-uate school. Patterns are similar in other pro-fessions. Graduate work in business or law isa popular road to career advancement in the

"American engineering schools enrolled 72.600 M.S. andPh. D. students in 1980. about 40 percent on a part-time basis"Engineering Enrollments, Fall 1982," op. cit. Although only337.800 undergraduates were enrolled in Japanese engineeringschools compared to nearly 400.000 in the United States, theretention rate is much higher in Japan. Once admitted, Japanesestudents face far fewer hurdles than Americans, and a higherpercentage graduate.

Table 69.Enrollments in JapaneseColleges and Universitites

Number of students, 1980Juniorcollege

Technicalcollege University M.A./M.S. Ph. D.

Engineering ..Physicalscience

All otherprograms ...Total

20,100

346,100

46,300 337,800

54,600

1,349,100

14,900

3,740

17,160

2.400

2,590

13,210

366,200 46,300 1,741,500 35,800 18,200SOURCE: "Kagaku Gijutsu Benran" (Indicators of Science and Technology),

Kagaku Gilutsucho Keikakukyoku (Science and Technology Agency,Planning Bureau). 1981, pp. 100.103.

United States, but not in Japan, where businessschools are virtually nonexistent and lawyersform a miniscule part of the labor force.

Continuing Education and Training

Despite the self-criticism that Japanese levelat their institutions of higle.er education, theperformance of the country's engineers andscientists across many fields, along with thedemonstrated competitiveness of Japanese cor-porations in high-technology industries likeelectronics, demonstrates that the system, tak-en as a whole, functions well. Indeed, some ofthe self-criticism appears to be no more thana mechanism for urging people and organiza-tions to greater efforts.

Deficiencies in universities are at least par-tially offset by, informal mechanisms for self-education, as well as company-run trainingprograms. Western observers repeatedly notethat men and women in Japan are voraciousreaders with a strong penchant for self-study.The average Japanese not only spends moretime reading than the average American, butmore of what he reads is job-related. The spreadof quality control methodologies throughJapanese industry, outlined in chapter 6, de -.pended heavily on self-study through books,magazines, and radio and TV broadcasting.The national broadcasting company, NHK,transmits nearly a hundred educational pro-grams to attentive audiences each week, in-cluding the popular "science classroom"series.

Japan's Government also provides free train-ing for recent high school graduates, as wellas for workers who need improved skills beforethey can join or rejoin the labor force. Over theyears, courses taken by those in the second cat-egoryadults already in the job market--haveexpanded greatly. They fill many of the samefunctions for smaller companies as the ink..,house training programs conducted-by largecorporations. Data collected by the Ministry ofLabor indicate that more than 200,000 traineeswere enrolled in publicly supported vocationalprograms in 1977, although the content of these

313


programs has been criticized for not keepingup with the needs of industry.3°

Company-Run Training ProgramsInternal training and continuing education

comprise an integral part of organization andmanagement in larger Japanese companies.This is perhaps the most fundamental differ-ence between the Japanese and American ap-proaches to technical education. While manyAmerican corporations engage in such activ-ities, Japanese programs are much more com-prehensive. Developed in part to compensatefor deficiencies in formal education, traininghas evolved to complement employment pat-terns in which many employees spend their en-tire careers within a single organization.

Of course, not all Japanese firms or workersfit this pattern. Table 70 shows that big com-panies provide much more training than small.One reason is that managers are generally ro-tated within large organizations, a practiceoften accompanied by study programs. Moreimportant, long-term employment within a sin-gle firmsometimes called "lifetime" employ-mentis the rule only in the major corpora-tions (and then only among male employees).

While training programs within Japanesecompanies generally impart specialized skillse.g., computer programingthey serve otherpurposes as well, purposes that may seempaternalistic or coercive to Western observers.

"H. Shimada, "The Japanese Employment System," Japan In-stitute of Labor, Industrial Relations Series, Tokyo. 1980, p. 21.

Table 70.Distribution by Size of Japanese FirmsProviding InHouse Training

Size of company by Proportion of companies withnumber of employees training programs (as of 1974)

1,000 or more employees ..500.999 ,

300-499100.299

_30-99**- 5-29

All firms.

95.1%85.375.958.826.310.141.3%

SOURCE: H. Shimada, The Japanese 'Employment System, Japan Institute ofLabor, Industrial Relations series, Tokyo, 1990. Based on data fromMinistry of Labor, Jlgyonai Kyolkukunren Jisshi Jokyo Chosa {Surveyof Intra-Firm Vocational Training and Education), 1974.

For example, corporations rely on in-housetraining to help buil a sense of loyalty to thegroup and to the ofg-mization." The widelyremarked cooperative spirit of Japanese em-ployees is no accident.

Well-known features of Japanese organiza-tional structures such as quality control circlesalso serve a training function, one in which theinformal elementsand the stress on inter-group cooperationare at least as importantas any knowledge imparted. In an unusuallycomprehensive program in a Japanese automo-bile plant, engineers teach other employees ina "workshop university."40 After completing anextensive program of after-hours study-2years or more, with no special remunerationthe workshop university graduate is rewardedwith a certificate from his section chief. Theaim is not only to improve individual skills, butto keep employees intimately involved in day-to-day matters that affect productivity andmanufacturing efficiencyranging from work-place organization, job flows, and task descrip-tions to interpersonal relations.

Among the most systematic of the industrialtraining programs in Japan have been those'de-veloped by leading manufacturers of electron-ics and electrical equipment. Since the 1920's,firms such as Mitsubishi Electric and Matsu-shita have been known for recruiting promis-ing young employees directly from high school,and giving them extensive and formalized in-house training.'" Such programs emerged inresponse to shortages of qualified wt. .7'ers inthe ,3ftermath of World War I. Japa:. ,...,overn-mer0- fostered universal primary education, but

3°For a detailed analysis of training within a Japanese bank,see T, P. Rohlen, For Harmony and Strength: Japanese White-Collar Organization in Anthropological Perspective (Berkeley,Calif.: University of California Press, 1974). In the bank studiedby Rohlen, some of the training programs emphasized technicalskills while others were directed at "character building." Bothvarieties were designed to help integrate workers into the cor-porate community. Over the course of a year, about one-thirdof the staff went through one or more programs at the bank'sown training institute.

For a comprehensive treatment of training practices at Toyota,see R. E. Cole, Work Mobility, and Participation (Berkeley, Calif.:University of California Press, 1979).

**Work Mobility, and Participation, op. cit., pp. 183-184."Levine and Kawada, op. cit., p. 267.

Ch. 8Human Resources: Education, Training, Manag&n:ont 317

during that period gave little attention to sec-ondary schooling. Vocational training was !eftto the private sector, where companies de-signed their own programs to train the workersneeded for expansion and industrialization. Ifthe government had pursued a more compre-hensive manpower policy, including the sup-port of secondary and vocational education,Japanese firms probably would not have movedso far in this direction.

Initially, then, internal training was a directresponse to shortages of skilled labor, and ef-fort was directed at blue- and grey-collar work-ers rather than managers or engineers. Despitevast improvements in Japanese secondary ed-ucation since the 1920's, most large companiesretainindeed have continued to developthese programs. Many operate their own edu-cational institutes; Hitachi, for instance, main-tains two, sending graduates of technical highschools for year-long courses of study.42 Thecompany, which is not untypical, also offersa large number of specialized training courseson an ad hoc basis. Hitachi has given morethan 1,000 over the past two decades (somemany times); they include foreign languagesand topics in management, with specializedsubjects such as international business avail-able for executives.43 Students in a typicalcourse spend 30 hours in the classroom andtwice that on outside assignments; the averageskilled worker or technical professional atHitachi takes two such courses a year.

It is difficult to compare the direct costs ofsuch activities with the corresponding benefitsto the firm. But even in large organizationswith extensive training programs, such as Toy-

. ota, expenditures reportedly total less than 1percent of salaries and wages.t, The returns-

-tangible and intangibleappear substantial.

"R. Dore, British FactoryJapanese Factory (Berkeley, Calif.:University of California Press, 1973), p. 651.

"M. A. Maguire, "Personnel in the Electronics Industry:United States and Japan," prepared for OTA ,inder contract No.033-1360, pp. 54-55 On training for managers in Japan, see T.Amaya, "Human Resource Development in Industry," Japan In-stitute of Labor, Industrialized Relations Series, Tokyo, 1963,pp. 21-24.

"Work Mobility, and Participation, op. cit., p. 185.

International Differences inEducation and Training

A principal conclusion from the precedingsections is that, while American universitiescontinue to provide an excellent education forthis country's engineers and scientistsas wit-nessed by the large numbers of foreign gradu-ate students who come here to studythe aver-age American high school or college graduateis poorly prepared to function in a technologi-cally based society. Compared to their counter-parts in a number of other advanced industrialnations, American students get less training inmathematics and science, and even if theystudy these subjects, learn virtually nothingabout technology.

Deficiencies in mathematics are particularlyserious. Mathematics acts as a filter at the en-trance to many careers. Although the impor-tance of mathematics to the practice of engi-neering is sometimes exaggerated, a high levelof competence relative to the average is neededto complete a degree program. A student whodoes not master algebra and trigonometry inhigh school drops immediately into the classof those needing remedial work; he or she willnot be admitted directly into a university pro-gram in engineering or science. Those with de-ficiencies who try to catch up often fail. Partof the problem is simply that as many as one-fourth of high school teaching posts in mathe-matics are currently vacant, and a comparablefraction are filled by individuals only temporarily certified to teach, many of them marginallyqualified at best.45 Industry has hired awaymany high school mathematics teachers at at-tractive salaries, in part to fill vacancies forcomputer programers and systems analysts.

The American educational system also doesa poor job of preparing those who do not goto college. Even among high school graduates,functional illiteracy is common (estimates forthe population as a whole range around 20 per-cent). Vocational education and training vary

"'IA Science Dean iJescribes Teaching as in Sorry State," NewYork Times, Mar. 6,1982, p. Cl. Shortages of teachers in scienceas well as mathematics appear to be worsening; see Science andEngineering Education: Data and Information, op. cit., p. 7.

32i


widely in quality; excellent programs and in-adequate ones can be found virtually side-by-side." Other countries have developed morecoordinated and comprehensive approaches tovocational training, with benefits both to indi-vidual workers and to industry."

Skilled technical workers are a vital resourcefor the U.S. electronics industry, and deficien-cies throughout the middle levels of the Ameri-can labor force could constrain the futuregrowth and development of semiconductor andcomputer firms, as well as cowponies in otherhigh-technology fields. Technicians, designersand draftsmen, and field service personnelmust be literate, have basic quantitative andtechnical skills, and, ideally, understand some-thing of the logic of the systems they work with.Without such abilities, they cannot use ad-vanced production and R&D equipment togreatest effect, nor exercise sound judgmentin the technical problems they face on a day-to-day basis. Individuals without these skillshave little upward mobility; an assembly lineworker needs at least some quantitative facilityto be able to move into jobs such as machinerepair, quality control and inspection, or shop-floor supervision.

'"See, for example. G. W. Wilbur. "Va. Vocational EducationSeen As Hindrance to Development." Washington Post. Nov.29, 1982.

"The extensive system in West Germany is described i 1"-

tional Training in the Federal Republic of Germany (Bri.Commission of the European Communities. 1978). See aib.o S.Hutton and P. Lawrence, German Engineers: The Anatomy ofa Profession (Oxford: Clarendon Press. 1981). pp. 94.95; and I.M. Geddes, "Germany Profits by Apprentice System." WallStreet Journal, Sept. 15. 1981. p. 33.

The demand for grey-collar technical em-ployees in industries like electronics is high;one study has estimated a growth rate in theUnited States of nearly 18 percent per year,faster than the projected growth in demand forengineers." But in pointed contrast to coun-tries like Japan or West Germany, the Ameri-can educational system has not responded inany large-scale fashion to these needs. In Ger-many, fully 60 percent of the labor force hasspecialized training in grey-collar technicalskills, while in the United States the figure maybe as low as 10 percent."

A scarcity of adequately trained technicalworkers could be just as serious a problem forAmerican industries like electronics as con-straints on capital investment or a stagnatingoverall economy. Labor mobility has tradition-ally been a mechanism for opening manpowerbottlenecks; indeed, the U.S. electronics indus-try already depends heavily on foreign-bornthough U.S.-educatedengineers. The nextsection looks more closely at the structure ofthe U.S. labor market, particularly mobility.

""Technical Employment Projections of Professionals andParaprofessionals, 1981-1983-1985," American Electronics Asso-ciation. May 1981; see also "Testimony of Robert P. Hender-son, Chairman and C.E.O., Itek Corp.. Lexington, Mass.," Fore-casting Needs for the High Technology Industry, hearing. Sub-committee on Science, Research, and Technology, Committeeon Science and Technology, House of Representatives, Nov. 24,

981''S. Praia, "Vocational Qualifications of the Labour Force

in Britain and Germany," National Institute Economic Review,November 1981, p.47; response of R. H. Hayes. Business Man-agement Practices and the Productivity of the American Econ-omy, hearings. Joint Economic Committee. May 1 and 11. andlune 1 and 5, 1981, p. 46.

Supply and Mobility of LaborShortages of men and women with knowl-

edge and skills at .a time of high overall unem-ployment poirt to weaknesses in U.S. labormarket policies, including manpower trainingand adjustment assistance." While "full em-

5°For a relatively comprehensive, and critical, analysis of labormarket policies in the United States, see R. I. Vaughn. "The lobDevelopment Administration: A National Employment, Educa-

ployment" has been a policy goal for manyyears, the upward trend of the unemploymentrate over the past decade has combined with

tion and Training Policy," Projected Changes in the Economy,Population, Labor Market, and. Work Force, and Their Implica-tions for Economic Development Policy, op. cit., p. 33. During1981, perhaps I million U.S. jobs wont unfilled, while,10 millionpeople were without work. See K. Sawyer, "Learning lobs inSchool," Washington Post, July 28. 1982, p. 1.

322


sporadic shortages of workers having specificskills to create a new circumstance, one towhich the Federal Government has failed torespond.

Over its history, the United States has seenperiodic labor shortages, for both skilled andunskilled workers. More recently, it has begunto seem thateven if the general quality ofAmerican education were to remain highthelabor market might simply not be able to supplythe right numbers of people, in the right places,at the right time. There are a host of reasonsfor such concerns, ranging from changing at-titudes toward work, to the aging of the U.S.population, to local constraints such as highhousing costs.° As the work force ages, and the.needs of the U.S. economy shift, retraining willbe the only way to utilize people's talents fully.

This section asks whether the developmentof the U.S. electronics industry will be con-strained by limited supplies of engineers andcomputer scientists (overall employment trendsare examined in the next chapter), togetherwith a related question: Are the high levels oflabor mobility that have characterized someparts of the U.S. electronics industry essentialfor continued growth and competitiveness?The comparative neglect of training and re-training in the Unit 'd States stems in part fromthe ease with which companies have been ableto hire new employees with needed skills; thisin turn has reinforced tendencies for workersto move from job to job in search of fresh op-portunities or higher pay. The labor market inJapan functions much differently. There, thesystem emphasizes long-term employment (forsome) and loyalty to the firm; mobility is low.Management practices in Japan have soughtto compensate for the weaknesses of such asystem, while taking advantage of the stabilityit brings; rather than looking for new peopleto revitalize faltering efforts, Japanese firmsredeploy those they have.

*ln Silicon Valley, a housing shortage has driven prices so highthat semiconductor firms have found it difficult to hire from out-side the area; few candidates can afford to move.

Overall Labor Market Trends

The labor forces of Japan and the UnitedStates expanded swiftly during the 1960's,largely as a result of postwar baby booms.Table 71 shows the rates of increase in bothcountries to have been considerably greaterthan in Western Europe. Japan's labor forcegrew from 49 million in 1966 to 56 million in1979, while that in the United States went from79 million to 105 million over tne same peri-od.51 Although Japan has experienced somelabor shortages, the relative abundance ofworking-age men and women in both Japanand the United States contributed to economicexpansion during the postwar period. Youngerworkers made up an especially large propor-tion of Japan's labor force during the 1950's.During the 1970-80 period, both countries con-tinued to experience rapid increases in theirworking-age populations (table 71); growth intheir labor forces will slow during the 1980's.

Rising employment levels in industrializedeconomies over the past two decades have beenaccompanied by shifts toward tilt.; service sec-tor; agricultural employment has declined,with manufacturing roughly stable or declin-ing slowly (see ch. 5, fig. 32). Japan has beensomething of an exception, with rise in in-dustrial employment couple_ All a sharpdrop in agriculture; both the industrial and theservice sector grew as a result of migrationsfrom the farm. Such trends will continue as in-

"Labour Force Statistics 1968-1979 (Paris: Organization forEconomic Cooperation and Development, 1981), pp. 18.19.

Table 71.Labor Force Growth in Several Countries

Average annual increasein labor force

1960-70 1970-80

United States 1.80/0 1.5%Japan 1.3 1.3WeSt Germany 0.3 0.7France 0.8 1.1

United Kingdom 0.2 0.3SOURCE: 1960.70W. Gelenson and K. Odake, "The Japanese Labor Market,"

Asia's New Giant, H. Pc..ick and H. Rosovsky (eds.) (Washington,D.C.: Brookings Institution, 1976) P. 590.

197480World Development Report, 1980 (Washington, D.C.: TheWorld Bank, August 1980), p. 147.

323

320 International Competitiveness in Electroni:s

dustrial employment in the advanced countriesslowly shrinks relative to services.

It is perhaps understandable that, during aperiod of rapid overall labor force expansionand continuing movement into services, theU.S. Government paid little attention to man-power policies: the economy was growing rap-idly; periods of high unemployment wereviewed as transient; people could take advan-tage of a relatively broad range of oppor-tunities. The situation today is much different:aggregate expansion has slowed; the skillsneeded by industry are more specialized; un-employment has become persistent. Currentunemployment is especially troubling becauseit is caused in part by mismatches between thecapabilities of people looking for work and thejobs available; in such circumstances, morerapid, oggregate expansion may do little good,and may even be impossible if growth indus-tries cannot hire the people they need.

Personnel Supplies 'for theU.S. Electronics Industry

In the United States, shortages of softWareengineers and semiconductor designers havebeen heavily publicized over the past few years.Not only has demand been higheven throughthe deep recession of 1982but warnings oflonger term shortfalls have been common. Oneeducator predicted that American schools willgraduate a cumulative total of 70,000 new B.S.r'iegree-holders in electrical engineering and:omputer science over the period 1982-85,while nearly 200,000 will be needed.52 As dis-cussed in the next chapter, demand for com-puter service technicians is expected to dou-ble during the current decade, with job open-

". ings for programers and systems.analysts go-ing up almost as fast.

A number of job-market surveys and esti-mates of aggregate demand for engineers havebeen conducted in the recent past. The LaborDepartment estimated that in 1980 there were17,000 unfilled entry-level engineering posi-

""Congress Warned of Shortages in Electrical, Computer En-gineers," Electronic News, Nov. 23, 1981, p. N. The rather alarm-ist estimates were those of K. Willenbrock, Southern MethodistUniversity.

tions throughout the Nation. Other estimateshave ranged up to 25,000.53 NSF's projectionsfor engineers together with scientists indicatethat `he total supply of new graduates shouldmeei the demand by the end of the decade.However, NSF may be overestimating the ex-tent to which scientists can function in engi-neering jobs; in any case, shortages are an-ticipated even by NSF in the computer field,for statisticians, and in several engineeringspecialties. About one third of the 1.4 millionjob openings in science and engineering overthe 1978-90 period are expected to be computerrelated (including programers). Despite NSF'srelative optimism, other forecastsadmittedlyoften conducted by or for industry, and thusperhaps skewed by the preference of compa-nies to be able to pick and choose when hir-ing new employeeshave projected massiveshortages of engineers, perhaps as many as.300,000 by 1990.54 All such forecasts should beapproached with considerable skepticism.None of the methodologieswhether based onsimple trend analysis, on survey techniques (asfor many of the engineering manpower stud-ies), or on econometric models (as used by theBureau of Labor Statistics)has a good recordfor pfojecting employment; there are too manyimponderables.

While forecasts and projections can warn ofpossible future shortages, insight also comesfrom current levels of unemployment in someoccupations. Unemployment rates have beenremarkably low in technical fields. During19'63, when overall U.S. unemployment aver-aged about 7 percent, only 0.6 percent of com-puter specialists found themselves out ofwork." The unemployment rate for engineers

"Henderson, op. cit., p. 63."Henderson, op. cit., p. 66; "Science and Engineering Educa-

tion for the 1980's and Beyond." op. cit pp. 48-50, 60; M. A.Harris, "Manpower Surveys Continue to Disagree," Electronics,July 28, 1983, p. 108. NSF concludes that interfield mobilityparticularly influxes of those trained in mathematicswill mit-igate but not ellminate the shortage of computer specialists. Onepotential problem is that even if the total supply of engineersroughly meets the demand, small firms with limited resourcesmay still be unable to hire new people.

Wetional Patterns of Science and Technology Resources 1982,NSi 82.319 (Washington, D.C.: National Science Foundation,Marrh 1982), p. 68. This amounts to only 2.000 people. Whileunemployment rates for professionals of all types are normallywell below the overall unemployment rate, the 0.6 percent figurefor computer specialists is unusually low.

32.,


(as a group) in 1980 was less than 1 percent;engineering unemployment averaged 1.8 per-cent over the decade of the 1970's, a period thatincluded the aerospace "collapse," when theunemployment rate for engineers reached 2.9percent." Aggregate unemployment levels dur-ing the decade averaged 6.2 percent, more thanthree times as high.

The persistence of unemployment rates farbelow the national average indicates that an"oversupply" of new graduates in engineeringis unlikely. And, while mathematicians andphysical scientists, as well as engineers, maysometimes have trouble finding the jobs theyconsider most desirable, men and women withtraining in such fields can move into a widevariety of occupations; many scientists even-tually find themselves practicing engineering.It is hard to argue that the United States couldhave too many graduates of science, mathe-matics, or engineering curricula..

. Data on salaries and job offers for new engi-neering graduates provide additional evidenceof high demand. In 1981, engineers made uponly 8 percent of new college graduates, butreceived more than 65 percent of all job of-fersand at starting salaries twice as high asfor those in the humanities.57 Salary offers toengineers and scientists rose at higher ratesthan for other categories of graduates through-out the 1970's. Demand remained high evenduring the recession of 1981-82.58

Another indicator of personnel shortages ismobility across disciplinesthe number of peo-ple who switch to fields other than those inwhich they got their formal education. Muchof the demand for computer specialists has,been filled by men and women with training

"Science Indicators-1980, op. cit., p: 320. The peak year forunemployment among engineers was 1971.

5713. Abelson, "Industrial Recruiting on Campus," Science,sem, ri, 1981. p. 1445. The data comes from a survey by theCollege Placement Council covering more than 60,000 offers tonew recipients of bachelor's degrees, The salary data also pointsout the big differences between indusWel and academic start-ing salaries.

"lr. 1982, two-thirds of computer and office equipment firmssurveyed by NSF reported difficulty in hiring electrical and com-pu'er engineering graduates, as opposed to 95 percent in 1981.See "EEs Still Needed, Though Shortage Has Eased, Says NSF,"Electronics, Jan. 13. 1983, p. 69.

in mathematics, engineering, and the physicalsciences; fewer than one-third of those work-ing as computer professionals have degrees incomputer fields." High turnover rates are partof the same picture- as noted earlier, turnoverhas been rapid among both engineers and tech-nicians in the U.S. electronics industry.

Regardless of uncertainties in the projec-tions, then, few people are worrying that theUnited States will have too many engineers inthe years ahead; capable individuals with train-ing in engineering comprise one of the mostemployable parts of the labor force. The pros-pects of shortage are real in the sense that var-ious projections differ mostly in the magni-tudes of the shortfalls predicted.

In contrast to the wide public awareness ofpotential shortages of engineers and computerscientists, supplies of grey-collar manpowerhave received remarkably little attention. Thus,it is impossible to discuss needs for techni-cians, service personnel, and other skilledworkers in quantitative detail. But the situationfor machinists illustrates the kinds of problemsto be expected. The Bureau of Labor Statisticsestimates that annual job openings will average22,000 over the near future; meanwhile, in 1978only 2,300 machinists completed- registeredprograms of apprentice training.80

The Question of MobilityLateral mobility helps moderate sporadic--

shortages of workers with particular sets ofskills. just as clearly, individuals can only movewithin a limited realm; a surplus of physicistsmight help compensate for a scarcity of com-puter engineers, but few biologists Would beable to function in such jobs.

59"Science and Engineering Education for the 1980's and Be-yond," op. cit., p. 39. This refler_ta in part the slow developmentof academic programs in computer science and engineering.

6°S. Qualtrough and J. Jablonowsld, "Filling the Need for SkilledWorkers," American Machinist, June 1579, p. 131. But see alsoN. H. Rosenthal, "Shortages of Machinists: An Evaluation ofthe Information," Monthly Labor Review, July 1982. p. 31. Al-though the electronics industry employs machinists, far greaternumbers work in heavier manufacturing industries. Regardlessof the statistics, a good deal of anecdotal evidence bears out the,difficulty that manufacturing firms of all types have had in find-ing journeyman machinists, tool and die makers, and otherskilled craftsmen.


American workers move within and acrosstechnical fields further and more frequentlythan their counterparts in other industrial na-tions. Managers and technical professionalschange jobs much more often in the UnitedStates than it_ Japan; mobility is greater inEurope than in Japan, but still considerably lessthan here.'" In the U.S. electronics industry,turnover has been high among unskilled work-ers, where unions have been week, as well asamong those whose abilities have been in highdemand.

The effects of labor mobility cut severalways. It is li!`le solace to a firm losing key peo-ple if they start a new enterprise that con-tributes to U.S. competitiveness. At the sametime, organizations with low rates of person-nel turnoverin any countrymust guardagainst stagnation, find ways to generate newideas; this is one of the reasons for internaltraining and job rotation programs in Japan.The pluses and minuses of high or low ratesof 1 lior mobility depend on factors such asrates of technological change, current eco-nomic conditions, and corporate strategies.

Patterns of mobility across industries andcountries depend, among other things, on in-centives such as promotion policies and wage/benefit packages; managements have consid-erable latitude in tailoring these to enhance ordiscourage turnover. Government programsdealing with adjustmente.g., training andretraining, unemployment assistancealso actas incentives or disincentives. While gener-alizations emphasizing cultural differences aresometimes advanced to explain mobility pat-terns in the United States as compared to Ja-pan, examining incentivesand the ways inwhich public policies affect themprovides asounder basis for understanding. AlthoughJapan's labor force tends to be less mobile thanthat of the United States, a good deal of varia-tion exists across industries and firms in bothcountries.

"'On West Germany, see German Engineers: The Anatomy ofa Profession, op. cit., p. 48ff.

Labor Force Mobility in the United StatesThe United States draws strength from the

mobility of its labor force, not only in mod-erating skill shortages, out as a stimulus to in-novation, technology diffusion, and entrepre-neurship. New firms in rapidly growing seg-ments of electronicssemiconductors, com-puter software and peripheralsare often builtaround engineers and managers who leave onecompany to start another. On the other hand,rapid staff turnover, as pointed out above,works against company-run programs of edu-cation and training. In part to counteract theattractions of entrepreneurial ventures, anumber of large and successful American elec-tronics firmsincluding Hewlett-Packard,Texas Instruments, and IBMhave adoptedpersonnel policies aimed at retaining theiremployees. Likewise, merchant manufacturerssuch as National Semiconductor and Intel at-tempted to maintain staffs and avoid layoffsduring the semiconductor sales slump of 1981-82. In this regard, American electronics man-ufacturers are quite consciously emulatingtheir Japanese competitors.

Still, white-collar mobility has been a sinequa non of the more dynamic merchant semi-conductor firms, which have competed aggres-sively for both technical and managerial talent.Silicon Valley manufacturers have offered awide range of benefits, including extensiverecreational facilities, to recruit white-collarprofessionals. Some have even paid bountiesto employees who bring in new people. Pros-pects for rapid advancementand the lure ofsomeday getting in on the ground floor of anew organizationhave helped attract mana-gers and engineers, as has the California set-ting. The mobility of talented people has helpeddiffuse electronics technology, contributing torapid commercialization of new developmentswhich in turn has helped build an interna-tionally competitive industry.

The lawsuits occasionally filed against ex-employees by firms seeking to prevent leakageof their technology are among the more strik-

326

44'

Photo credit

,,ny electronics technicians get their oin the military

illustrations of the relation bet,mobility and technology diffi's unsuccessful 1968 caseLives who went over to Faircy example. In 1980, Intel suener employees who left to stetied Seeq; the basis of Intel'sed in a negotiated settlement,mployees intended to baseiness on trade secrets dealign and manufacture of larknmable ROMs (read onlyal action to prevent technolog

Russell, "Seeq Loses Bid for Rehearing.":5. 1982, sec. 1, p. 50.


vet, 1983

tining

arson-Moto-;t ex-as anlup ofmanyvhichat thetheir

h thepo-

ies).82)ws is

News,

an extreme case; more commonly, firms adoptpositive programs of rewards and incentivesto keep valuable employees. Again, the semi-conductor industry has been a leaderhelpedby a working environment that many employ-ees find stimulating. Of course, features thathelp retain people also serve a company wellin attracting new employees.

Turnover has also been high among unskilledblue-collar workers in many parts of the U.S.electronics industry. In domestic semiconductor plants, production employees tend to befemale and ethnic. According to one estimate,women make up 40 percent of the SiliconValley work force, and are heavily concen-trated in lower paying jobs; three-quarters ofassemblers are women." In contrast to the'mobility of top-echelon managers aid technicalprofessionals, turnover among unskilled pro-duction workers is associated with a lack ofskills; they can be laid off during businessslumps and replaced later.

Mobility in JapanThe stereotype of Japan's "lifetime" employ-,

ment system contrasts sharply with patternsin the U.S. electronics industry. According tothe popular view, the Japanese system ensuresjob security until retirement. Also part of thestereotype is a sequence of promotions basedlargely on seniority rather than merit, withemployees waiting patiently to move up the payscale, assured of their ultimate reward. Theseaspects of the Japanese system have beenviewed as integral parts of a company-as-familymodel, making unions in the American orEuropean style superfluous. "Enterprise un-ions," organized on a company basis ratherthan by trade or occupation, have been seenas part-and-parcel of a socioeconomic milieucharacterized by harmony among workers andmanagersthis in turn leading to low interfirmmobility coupled with high employee motiva-tion and productivity. While pieces of thismodel are visible within Japan's economy,, it

,312. Howard, "Second Class in Silicon valley," WorkingPap?r,I, September-October 1081, p. 25. See also M. Chase,"Semiconductor Firms Get Mixed Review on Safety in Studyby California Agency," Wall Street Journal. Jan. 1, 1902, p. 6.

327

324 International Cornpet:Vvaness in Electronics

applies to only a minority of the labor force;moreover, the stereotype obscures crucial de-tails that affect the working lives of all Japa-nese.

To begin with, labor relations were far fromsmooth in Japan as recently as the 1950's. Fur-thermore, lifetime employment is typical onlyof large Japanese companies, and many ofthese encourage their employees to retire at arelatively early agecommonly around 55 or60. Afterward, many "retirees" must find newworkwhich may turn out to be a part-time orlower paying job with a subsidiary of theirformer employerbecause retirement benefitsare low.°' Moreover, in small firms especial-ly, but also in large enterprises, Japanese work-ers do leave their jobs. Horizontal mobilityi.e., movement from one firm to another with-out advancementis fairly canmon amongyounger Japanese workers, particularly thosewith low skills. Women seldom have much jobsecurity or upward mobility, much less themany temporary employees that are anotherfeature of Japan's labor market." Women aregenerally encouraged to resign upon mar-riagecertainly at childbirthand if theyreturn have no seniority. The 3.4 million tem-porary and day workers; men and womenac-counting for about 6 percent of the workforceare the first to be let go in the event ofrecession. Temporary employees provide flexi-bility to cope with economic slumps withoutlaying off regular workers. The proportion oftemporary employees in Japanese manufactur-ing firms has increased markedly since the re-cession of the mid-1970's." Furthermore,

"Japanese electronics firms, along with the rest of Japaneseindustry. have been under some pressure to extend retirementages. In the mid-1960's, retirement in the major electronics firmswas generally compulsory at 55 to 57 for men, perhaps 50 forwomen. By the mid-1970's, many companies had extended theseages by about 5 yeirs. See S. Takezawa, et el., Improvementsin the Quality of Working Life in Three Japanese Industries(Geneva: International Labour Office, 1982), pp. 66-87, 95.

"A. H. Cook and H. Hayashi, Working Women in Japan(Ithaca, N.Y.: Cornell University Press, 1980). See also F.Ginsbourger, "Japan's Dark Side," World Press Review, July1981, p. 32.

mC.-y. Lin, "Wage-Price Behavior Under External Price Shocksand Productivity Slowdown: A U.S.-Japan Comparison," Discus-sion Paper No. 402, Economic Growth Center, Yale University,April 1982, p. 22a.

although corporations in Japan attempt to ad-just to business downturns by reducing work-ing hours before laying off regular employees,when recessions deepenas in 1974-75theyreduce employment levels at rates quite corn-parable to those in Europe, if not the UnitedStates. Smaller Japanese companies have sel-dom been reluctant to cut back their laborforces.

Nor does the stereotyped picture of seniority-based wages in Japanese corporations (thenenko system) hold up under scrutiny. Or)?study finds that promotion is based on a "col..promise" between seniority and ability, the par-ticulars varying considerably across firms.°7Smaller, more rapidly growing organizationstend to emphasize meritocratic promotion,while older, established firms remain less will-ing to single out talented individuals fromothers of their age group. Age and ability are,furthermore, weighted differently dependingor level, with progress in the upper ranks astronger function of ability. Clark concludesthat the ambiguity built into Japanese promo-tion practices encourages people to do theirbest: while promotion has generally been auto-matic after a certain period of service, thereis also the possibility that outstanding perform-ance will be rewarded with rapid advance.And, although the nenko system may appearto underpay well-trained and able young work-

., ers, over the longer term they can expect to at-tain salary levels well above those in their agebracket who have lower skills or less education;salary profiles for older male workers in Japanshow considerable spread.

Finally, as the Japanese labor force has aged,employment practices have begun to change.With the fraction of older workers increasing,salary competition for the best qualified recentgraduates will intensify; recent surveys of hir-ing suggest that, in -:ertain fields, includingelectronics, shortage' 'f younger employeesare likely. Data oomph, by the Ministry of

.711. Clark, The Japanese Company Haven: Yale Univer-sity Press, 1979), pp. 115-118. On the nment of the nenkosystem, see T. Inapmi, "Labor-Manage, ' Communicationat the Workshop Level " Japan IMMitute of L. Industrial Rela-tions Series, Tokyo, 1983. InEq.,,e:ni also inclue data on pro-motion patterns (pp. 10-14).

323


Labor indicate that younger Japanese workerscan choose between two or three entry-leveljobs, but those aged 55 and over must win outover 5 to 10 other jobseekers to find a posi-tion." As a result of such trends, wage com-pression for older employees seems likely tointensify, retirement ages may be extended fur-ther, and the role of seniority in promotiondecisions will diminish. Generational conflictbetween younger employees, for whom highdemand will push tip salaries, and older work-ers who stand to lose by comparison, may fol-low." If and when such events come to pass,the features that now. make the Japanese em-ployment system seem unique will stand outless.

The United States and Japan Compared

While the contrasts are often exaggerated,Japanese and American employment practicesdo lead to quite different patterns of mobility.How do these interact with the structures ofthe electronics industries in the two countriesto affect international competitiveness? Firmsin each nation have alternatives in seeking thepeople they need. One option is to hire employ-ees away from other companies, an approachmore likely to be successful in the UnitedStates. An alternative is internal recruitmentiritrafirm mobilityin conjunction with re-training, an avenue particularly appropriate ina system such as Japan's, where people tendto identify more strongly with the corporationthan with a vexation or profession. Still anothermethod of coping with shifting occupationalneeds is to alter or expand the potential poolof new entrants. Despite the vitality that theU.S. electronics industry has drawn from em-ployee mobility, there is no need to associateeither high mobility or a lack of mobilityin

6,Clark. op. cit., p. 32.0,R. E. Cole, "Changing Labor Force Characteristics and Their

Impact on Japanese Industrial Relations," The Paradox of Prog-ress. E. Austin (ed.) (New Haven, Conn.: Yale University Press,1976). p. 194.

and of themselveswith enhanced competi-tiveness; nor should high mobility be consid-ered more "modern" than low (or vice versa).High mobility in the United States goes withother aspects of the U.S. economy, just as lowmobility is consistent with Japan's socioeco-nomic environment.

Public policies influence the choices madeby corporations among the options outlinedabove. Government support for training techni-cians can enlarge the talent pool. Vigorousmanpower policies, designed to support ex-panding sectors of an economy, will stimulateinterfirm and interindustry mobility. Taxwriteoffs for company-run programs of educa-tion and training would encourage intrafirmmobility. High turnover rates have made Amer-ican corporations wary of investments in train-

' ing or retraining that may pay off to their com-petitors. "Talent raiding"so characteristic ofAmerican semiconductor firmsoften be-comes the alternative.

Employment practices in the United Statesmay change as a result of the demographics ofaging, just as the aging of the Japanese lnborforce is altering patterns in that country. Asthe U.S. population grows older, continuingeducation for those in midcareerblue- andgrey-collar workers, as well., as white-collarprofessionalswill become a necessity. Whenthe labor force was expanding rapidly, employ-ers could count on new graduates to fill manyof their needs; this is less true today, particular-ly in light of current inadequacies in technicaleducation. American firms may find them-selves emulating the internal training effortsof their Japanese competitors, with manage-ment practices designed to enhance a com-pany's human resources becoming criticalelements of corporate strategy. The remainderof this chapter turns to questions of manage-ment and the organization of the workplace,askingamong other thingswhether therereally is a uniquely Japanese. approach tomanagement.

323


Organization and ManagementAs the end of the 19th century brought rapid

economic growth and technological change toAmerican, industry, new management tech-niques arose to deal with shopfloor organiza-tion. The old ways, developed during the daysof hand work, proved a poor guide to factoryproduction using mechanized equipment, par-ticularly in the emerging mass production in-dustries

Frederick W. Taylor, founder of the scien-tific management school, is the best remem-bered of those who pioneered new methods.7°Taylor began as an engineer at an ironworks,and his approach to managementincludingplant layout and job flows, and the man/machine interface--reflects the bent for ration-alization associated with his profession. WhileTaylor himself, and the techniques he devel-oped and advocated around the turn of the cen-tury, showed considerable appreciation for thehuman element in factory work, many of hisfollowers carried scientific management to theextreme of treating people as another varietyof machine. Scientific management still bearsthis stigmaand "Taylorisrn" is a dirty wordfor many who associate it with the Chaplin ofModern Times.

Taylor believed that, for every task inmanufacturing, there was an optimum methodthat could be "scientifically" discovered. Byreducing each job to its essential elementsemploying, for instance, the techniques of whathas come to be known as time-and-motionstudythe workplace was to be rationalizedand productivity maximized. Although Taylorthought that this approach should also increasecooperation among workers, one of the chiefcriticisms Ol scientific management has alwaysbeen its rather mechanistic conception of theindividual, leading to an emphasis on simple,repetitive tasks.

70F. W. Taylor, The Principles of Scientific Management (NewYork: Norton & Co.. 1911). N. P. Mouzelis, Organization and,Bureaucracy: An Analysis of Modern Theories (Chicago: AldinePublishing Co., 1967), gives .a useful historical overview of var-ious approaches to organizational management.

.

.77 y

kiz;

Photo credit: Wash

Integration of programmable robots into the-factdry environment poses a new set of probler

for manufacturing industries

The idea of a scientific means for org,ing factory work attracted American busi:men. New machine tools, the assemblymass production of durable goods like bic3home appliances, and automobiles, presea rapid succession of new problems; irtrialists eagerly embraced Taylorismmeans of dealing with them. The manage]science movement springing from Tayearly work has continued to thrive anspread internationally; ft still shapes curriand textbooks in American business scjand industrial engineering programs.

The human relations approach to marment was developed .primarily by induE'psychologists, beginning a decade or two


In contrast to the engineers who espoused Tay-lorisin, the human relations school stressedpeoples' attitudes and motivation as keys toproductivity and manufacturing efficiency.Studies in the human relations vein exploredthe workplace as a social organization and theindividual employee as a member of the group;practitioners saw their goal as fostering an ami-cable working environment, one built aroundthe existing shopflbor culture. While advocatesof scientific management tended to be anti-union, the human relations school accepted

_ unions as an integral part of the social systemof the factory.

Just as the reductionist tendencies of scien-tific management have been criticized, so thehuman relations approach has been faulted forits stress on harmony to the neglect of the realconflicts of interest characteristic of work life,and for overemphasizing small group behaviorwhile failing to deal with the organization asa whole.

Variants of these two attitudes toward man-agementwhich reflect contrasting theories oforganizationcontinue to proliferate. The twoare based on fundamentally different notionsof what makes organizationswhether factory,store, or officefunction, and hence on meth-ods for improving their operation. At present,the human relations approach has becomeidentified with the popular view of Japanesemanagement, but both schools have Americanorigins. This is not, of course, to say thatAmericans cannot learn from foreign experi-ence. Organizations in other countries haveadapted management practices originating inthe United States to their own needs, and itmay be time for a reverse flow into Americancorporations.

Organizational Types andManagement Styles

The Manager as Professional

In the United States, management is a disci-pline with its own graduate schools and ad-vanced degrees; M.B.A. programs increased byan order of magnitude over the past two dec-

ades, and now graduate more than 50,000 menand women each year. In contrast to Japan andWestern Europe, where top managers tend tomove up from the ranksand a few individualsstill reach high levels having started on theshop floorAmerican firms, especially thelarger, publicly held corporations, have tendedto bring new employees directly into manage-ment-track jobs. Typically graduates of aca-demic programs in business administration,some of them fill staff positions, others movequickly into middle management. Thus themanagement professionwith its extensivenetwork of specialized academic programshas become a principal vehicle for transmittingand validating the techniques used in Ameri-can business.

Management training in this country pre-pares people for work in hierarchical organiza-tions. Distinctions between those who plan andthose who do the work are more sharply drawnin American corporations than elsewhere; thisdivisionand the equally sharp distinctions be-tween those responsible for production, or"operations," and the rest of managementhas increasingly come under scrutiny and criti-cism." In contrast, Japanese and Europeanbusiness practices are rooted in on-the-job ex-perience and company-run training programs.Management institutes exist, but are typicallyoriented toward the needs of midcareer ex-ecutives seeking fresh perspectives.

While the ideal types of "American" and"Japanese" management are exaggerations thatfail to capture the variety existing within thetwo countries, they nonetheless point to differing conceptions of the nature of modern orga-nizations. The Japanese model is based onauthority stemming from tradition and social-ization; the American approach is less personaland more legalistic. Central to the Japanesemodel are group decisionmaking, cooperationbetween labor and management, and long-termtenure in an organization viewed as analogous

"See ch. 6. The following pair of articles in the July-August1981 issue of the Harvard Business Review are typical examplesof this criticism: R. H. Hayes, "Why Japanese Factories Work,"p. 56, and S. C. Wheelwright, "JapanWhere Operations Real-ly Are Strategic," p. 67.

331


to a family. Ideally, these result in a well-integrated system, with human resources as thefirm's most important long-term asset."

Decisionmaking in Japan

Symbolic of the Japanese approach is theringi seido (approval system), through whichmiddle-level personnel obtain sanction and ap-proval from the top echelons by circulating adocument to which each person affixes his sealor signature." The process yields systeniaticbut slow "bottom-up" decisionmaking. A deci-sion is final once the company president addshis seal; since many individuals participate,there is considerable communicationif notalways true consensusand, once the outcomehas become apparent, little uncertainty. Con-trasts are frequently drawn between thetendency of "individualistic" Americans tocontinue pushing their own views, even aftercontrary decisions by upper management, andthe Japanese casewhere, as the saying goes,"when the train leaves the station, everyone ison board." The point is that whatever disagree-ments precede the ringi decision, they are sup-posedly buried afterwards, the policy fully sup-ported by all.

While authority in a Japanese company, isvested in 'the president, employees at manylevels participate in the consensus-buildingprocess. Not all of them have the precise andwell-defined responsibilities that characterize,job descriptions in an American corporation.Ambiguity attaches to organizational structurein Japan, rather than to people as in the UnitedStates. The Japanese system does not involve

--much bargaining among managers'anagers and subor-dinates, nor is it participative in the sense oftenused in the West. In contrast to U.S. practice,where management decisions and businessplanning get detailed attention, the ringi systemallows people throughout the firm to agree ongeneralities, with the specifics to be worked outlater.

"N. Hatvany and V. Pucik, "Japanese Management and Pro-ductivity," Organizational Dynamics, spring 1981, p. 8.

"For a detailed description of the ringi seido, see M. Y.Yoshino, lapLn's Managerial System: Tradition and Innovation(Cambridge, Mass.: MIT Press, 1968), pp. 254ff

Group decisionmaking as embodied in thetraditional Japanese approach is a good fit withcorporate organizations that .offer individualemployees considerable security and involvethem with the company outside their im-mediate duties and working hours. Companyhousing and recreational facilities, groupoutings and even vacations, along with inter-nal training programs, can all be viewed as in-centives for building loyalty among a fairly im-mobile labor force. In, best light, the system is"wholistic" in orientation; in worst light, it is,a sophisticated brand of industrial paternal-ism." The widespread acceptance of companyrather than craft or tide unions and the 'com-paratively few days lost to strikes in Japan(table 72) indicate that this labor-managementsystemoriented toward consultatiod'and con-formityhas worked to the benefit of the cor-.porations that have designed and implementedit. As table 72 illustrates, large numbers ofworkers participate in strikes even in Japan,but little time is lost because work stoppagesare short, often serving functions that are atleast partially symbolic.

Contrasts With the United StatesExtensive involvement with the company

outside normal working hours and groUp deci-sionrriaking diverge markedly from patterns inthe United States, whererather than spread-ing responsibility for decisions through theorganizationtop management is expected to

"As late as 1976, more than one-quarter of Hitachi's male em-ployaes still lived in company housing; the figure had been nearly40 percent in 1967. See Improvements in the Quality of Work-ing Life in Three Japanese. Industries, op. cit., p. 69.

,Table 72.Work Stoppages Due to Labor Disputes

"49 Several Countries (1978)/ Total number of Total number ofparticipants in employee work-days

work stoppages lostUnited States 1,600,000 39,000.000Japan 660,000 1,360,000West Germany 490,000. 4,280,000France 1,920,000 2,200,000United Kingdom . 1,040,000 9,400,000SOURCE: "Japan, An International Comparison 1990," Ketzai Koho Center, p.49.

332


provide leadership. Corporate cultures in theUnited States give pride of place to strong-willed executives who leave their mark on an

1.organization. Power, 'status, and privilege also-attach to Japanese executives, but with real dif-ferences. The ability of managers to show im-mediate resultsprofitability over the nextquarter being the current target of criticsiscentral to the American model, group effort tothe Japanese. Well-defined and often narrowresponsibilities, centralized authority, rigidlyhierarchical organization chartsplus the pos-sibility of swift promotion and high rewardscharacterize the "results-oriented" manage-ment styles of American firms. Ambiguity isviewed as undesirable; expertise cultivated;men and women enter the firm as specialistsin accounting or finance, marketing or stra-tegic planning. Individualism is tolerated,but within well-defined boundswitness the"white-shirt syndrome" still hanging over com-panies like IBM." The comparatively highlevels of personnel mobility in the UnitedStates, and the tradition of adversarial relationsbetween unions and management, are part ofthis picture.

A further difference between Japanese andAmerican management practicesdiscussedin more detail in chapter 6is the emphasiscompanies in Japa, place on manufacturingand its integration \-ith the rest of the firm.Toyota's much-noted system of just-in-time(kanban) production and inventory control isonly one example. Since Japanese managerstend to rise relatively slowly through the ranks,with periodic lateral moves, stress on manufac-turing is perhaps natural. In contrast to thesituation in the United States, where produc-tionmore especially quality controlhas lit-tle prestige, is even viewed as a dead-end job,a number of Japan's top corporate executivesbegan their careers as quality control ormanufacturing engineers.

Both Japanese end American approaches tomanagement have their strengths dnd weak-nesses. Few corporations exhibit management

" "Life at IBMRules and Discipline, Goals and Praise ShapeIBMers' Taut World." Wall Street Journal.Apr. 8. 1982. p. 1.

99 -111 83 - 22

styles as clear-cut as the stereotypes suggest;in both countries, firms have identities thatmay vary from division to division as well aschanging over time. The wholistic orientationof the Japanese style carries strong paternalis-tic overtones, with discrimination againstwomen and minority groups a fundamentalpart of the system." And, although Japanesemanagement is sometimes viewed as people-oriented, personal interactions are marked bypervasive if subtle status distinctions. Paternal-ism does lead to job security for some fractionof the labor force in Japan, security which isless common in the United Sta't'es. The Ameri-can approach, while often assumed to maxi-mize opportunity, does so in part by encourag-ing competitionsome would say to excessamong individuals seeking advancement andpersonal gain.

Comparisons of American and Japanesemanagement often focus on particular tech-niquese.g., quality control circles in Japan,management by objectives in the UnitedStatesrather than the schools of thought, suchas scientific management or human relations,.from which these techniques derive. Burim-provements in management seldom result fromthe isolated adoption of some technique. Thisquotation from a Japanese engineer points outthe difference between technique and under-lying attitude:"

One difference I find hard to explain to myWestern colleagues is that we do exactly thesame things that the industrial engineer doesin Detroit or Pittsburgh; but it means some-thing different. The American industrial engi-neer lays out the work for the worker. Our in-dustrial engineers are teachers rather thanmasters. We try to teach how one improvesone's own productivity and the process. Whatwe set up is the foundation; the edifice theworker builds. Scientific management, timeand motion studies, materials flowwe do all

7°Even the more bemused commentators on the Japanese mod-el. such as Ouchl, note its racism and sexism. See W. G. Ouchi,Theory Z: How American Business Can Meet the Japanese Chal-lenge (Reading, Mass.: Addison-Wesley, 1981). p. 91.

77Quoted in P. F. Drucker, "What We Can Learn From .JapaneseManagement," Harvard Business Review. March-April 1971. p.117.

333


that, and no differently from the way you doit in the States. But you in the States think thatthis is the end of the job; we here in Japanbelieve it is the beginning. The worker's jobbegins when we have finished engineering thejob itself.

It is too easy to write off such statements asempty philosophizing.

/Worker Participation

The past decade has seen continuing intorestin industrial democracy, more so in WesternEurope than in the United States," Stemmingat least in part from persistent economic prob-lems, some companies and some governmentshave experimented with methods for increas-ing the involvement of the labor forcepar-ticularly blue-collar employeesin decision-making and work design. One aim has been tomoderate wage demands. This section outlinesseveral of the modes of employee participationthat have evolved in Europe, as well as the

_quality control circles originating in Japan. Thepurpose is to capture some of the variety offoreign approaches, and ask how such mech-anisms might help the productivity and com-petitiveness of American industry. A largenumber of specialized techniques for redesigndf the working environment and employee in-volvement have been developed, both here andoverseas; no attempt has been made to describeany except quality circles, which are coveredbecause they have attracted so much atten-tion."

"Much of. the material.on European countries in this sectionis based on A. L. Ahmuty, "Worker Participation in Manage-ment Decision-Making in Western Europe: Implications for theUnited States." Congressional Research Service Report No.79-136E, Apr. 23, 1979. See also P, C. Roberts, H. Okamoto, andG. C. Lodge, "Collective Bargaining and Employee Participa-tion in Western Europe, North America and Japan," Tim Tri-lateral Commission. 1979.

"For an overview of a number of these, sce R. M. Kanter,"Dilemmas of Participation: Issues in Implementing Par-ticipatory Quality-of-Work-Life Programs," National Forum,spring 1982, p. 16. Several case studies can be-found in J. A.Fadem. "Automation and Work Design in the United States,Working Paper Series No. 43. Center for Quality of Working Life,Institute of Industrial Relations, University of California, LosAngeles. 1982.

Participative Mechanisms

In the United States, industrial democracyhas been associated with collective bargainingby labor unionsan interpretation of workerparticipation neither so encompassing as inWestern Europe nor quite so narrow as inJapan. While American unions have continuedto bargain over wage-benefit packages, Euro-pean workers have succeeded in extendingtheir influence over workplace and organiza-tional decisions. In some contrast, quality con-trol (QC) circles were developed by managersin Japan as tools for improving labor produc-tivity and product quality. Most of the interestin QC circles among Americans has alsonrig-inated with management. If American work-ers, particularly in companies with strongunions, have sometimes been reluctant to em-brace QC circles, quality-of-work-life programshave found, a better reception with labor.

The worker participation movement in West-ern Europe is based on two presumptions: first,that labor is just as important to production ascapital; second, that blue-collar employees havethe right to be represented in corporate deci-sionmaking. Participatory mechanisms includework groups at the shopfloor level, work coun-cils at the plant or enterprise level, collectivebargaining, labor representation on boards ofdirectors, employee-owned enterprises, andworker representation on socioeconomic ad-visory bodies to governments. Beyond thesedirect involvements, publicly owned compa-nies are a longstanding fixture on the Europeanscene, with governments paying more or lessattention to their management depending onpolitical pressures and economic conditions.

At the shopfloor level,: work-life programs.--give employees a voice in determining how in-dividual tasks should be performed, with the_aim of increasing job satisfaction as well as im-proving productivity. Employee involvementin work methods can be viewed as a reactionagainst the scientific management tradition, inwhich an experttypically an industrial engi-neerhas full responsibility for task design.Sometimes, work-life programs reduce produc-

33,1

Ch. 8Human Resources: Education, Training) Management 331

tivity (as traditionally measured), a sacrificethat firms like the automobile manufacturerVolvo appear to have accepted in the interestsof employee satisfaction. (Volvo replaced anumber of assembly lines with batch assemblyoperations, giving workers more variety.)

At the enterprise or plant level in Europe,work councilsindependent of unionsgiveemployees a voice in codetermining a firm'sfuture. Labor representatives\on these councilsparticipate in financial and other business deci-sions, although at the head of the agenda tendto be matters like personnel policy, health andsafety, and shopfloor organizational practices.American-owned companies in West Germanyhave seldom been comfortable with codeter-mination; in the United States, the few labor-management committees that have been estab-lished tend to have a much narrower fous, andto be viewed primarily as vehicles for enhancedcommunication. One of the best knowi is theNational Committee to Improve the Qiility ofWOrk Life, established by the United \ AutoWorkers and General Motors in 1973. Currenteconomic conditions may motivate more uchexperiments.

._

One of the most far-reaching experimen s inemployee participation has been instituted inWest Germany. In the early 1970's, the Min-istry of Research and Technology, in coopera-tion with the Ministry of Labour and Social f-

rrr.\fairs, began a program aimed at the "huaization of work." Based on the Work Counci sAct passed by the German parliament in 197the premise is that government should not onlysafeguard employee health and safety, but undertake to improve opportunities for individualdevelopment and participation in decisionmak-ing.8° In general, the response of workers tothese initiatives seems to have been less posi-tive than for earlier programs of codetermina-tion, particularly in industries like electron-ics where the workplace is already relatively

°°"Research on the Humanization of Work," Action Pro-gramme of the Federal Minister for Labour and Social Affairsand the Federal Minister for Research and Technology," Doc.No. 2181/74e. See also Programm Forschung zur Humanisierungdes Arbeitslebens, Der Bund,.sminister fur Forschung andTechnologie, 1979.

benign. West German workers have remainedmore interested in power over matters such ashiring and firing practices.

Blue-collar employees in the United \Stateshave restricted their attempts to influencecom-pany policies and decisions to the traditionalconcerns of labor-management relations.\Un-ion officials have been ambivalent about mov-ing beyond questions of wages, benefits, andworking conditionsprobably fot fear of losingsome of the bargaining power that comes withan adversarial stance.81,In contrast to WesternEurope, participation by'American workers onboards of directors has been ralemostlybrought on by circumstances such as Chrys-ler's recent financial plight. Although the manyplant closings in industries like steel have ledto proposals that employees purchase facilitiesscheduled to be shut down, few such planshave gone forward.

While collective bargaining is virtually uni-versal in advanced market economies, there aremany differences of form and substance. InJapan, about 95 percent of all unions are orga-nized on an enterprise basis.82 In addition tocollective bargaining between unions and man-agement, negotiations take place each springbetween groups of, firms and unions. The"spring Offensive" is most visible in the steel;electrical machinery, shipbuilding, heavy ma-chinery, and automobile industries, as well aspublic corporatio (where a special mediationcommittee decidGs on the settlement). Wagedecisions during the spring offensive help setpatterns for smaller firms. Still, compared withthe. United States or many European nations,labor in Japan has little real power.

Quality Control Circles,QC circles have been heavily publicized as

mechanisms for worker participation. Quali-ty circles are relatively autonomous, composed

"For an evaluation of labor-management committees in theUnited States, see K. Frieden, "Workplace Democracy and Pro-ductivity," National Center for Economic Alternatives, Wash-ington, D.C., 1980, p. 31.

"' "Labor Unions and Labor-Management Relations," Japan In-stitute of Labour, Japanese Industrial Relations Series, Tokyo,1979.

335

332 Intornational Competitiveness in Electronics

of a small group of workersperhaps a dozentypically led by a foreman or senior employee."In Japan, financial incentives play a relativelyminor role, without the emphasis on prizes forsuggestions or improved performance thatsome American firms have adopted. QC meet-ings in Japanese companies are often held out-side normal working hours, and workers maynot be paid for their time. Although the circlesnow work on job-related problems beyondquality control per see.g., production meth-ods, worker trainingthey grew out of thepostwar stress on quality inspired by Ameri-cans such as Deming (ch. 6). The contributionmade by Japan's business leaders was the ex-pansion of quality control to involve participa-tion by virtually everyone in the firm. Employ-ee training via circles, for example, is intendedto reduce the need for specialists in quality as-surance and production engineering. As dis-cussed in chapter 6, the quality and reliabilityof electronic products depends on factors rang-ing from engineering design to relationshipswith suppliers; while the quality of many Jap-anese goods is now excellent, iL would be a mis-take to attribute this to any one technique suchas the QC circle.

Cole notes that even in Japan enthusiasmwithin a QC group tends to wane, and circlesneed to be periodically revitalized. It would beno surprise to find a Hawthorne effect at workin many of the success stories involving QCcircles (i.e., a situation in which any of a widevariety of changes in the workplace environ-ment would improve employee motivation andproductivity, at least temporarily). The effec-tiveness of QC circles also depends on thetent of employee identification with the com-pany; members participate more fully if theyfeel that. their work is recognized and ap-preciated within the organization. A group-oriented Japanese corporation is more likely tofoster such attitudes than many of the Amer-ican firms now experimenting with QC circles.

"Cole has carried out the most systematic studies on qualitycircles. The discussion below is based largely on Work Mobili-ty, and Participation, op. cit.. pp. 135ff. Also see R. E. Cole, "WillQC Circles Work in the U.S.?" Quality Progress, July 1980, p.30: Improvements in the Quality of Working Life in ThreeJapanese Industries, op. cit., pp. 76ff; and Inagami, op. cit., pp.31-34.

But even in Japan, QC circles are sometimesperceived as a coercive management tool.Overenthusiastic accounts of quality controlcircles in Japan sometimes give the impressionof a panacea; in reality, Japanese firms varywidely in the extent to which they utilize QCcirclesregardless of commitment to circle ac-tivities, they are only one manage ent tech-nique among many.

Over a hundred American firms includingGeneral Motors, Ford, and Genera Electrichave experimented with QC circ es, but thequestion of whether or not they v ill work aswell in the United States as in Japan has notbeen answered. Certainly there are obstacleshere that do not exist in the typical Japaneseorganization. In the U.S. context; for example,monetary incentives may, be essential; theLockheed program is typical in that employeesare not expected to meet after hours, or withoutextra pay." Experience also shows that Ameri-can middle managers must be persuaded to ac-cept and support the QC approach, else theymay perceive the circles as challenges or as im-plicit criticisms of past performance.

Unionized firms add another dimension.Where QC circles have been introduced intoAmerican companies without the consultationand support of union leaders, they have notbeen successful. Organized labor remains am-bivalent; AFL-CIO spokesmen have felt that QCcircles could be a tool for breaking up unions,and the evolving attitude appears negative.85

Japanese firms with plants in the United.States have generally introduced circles grad-ua'Icr avid with considerable care, if at all.Quasar, owned by Matsushita since 1974, didnot install its first circles until 1982; the com-pany plans to have 25 in operation by the endof 1983.88 QC circles in Japan function in a con-

64"Quality Cc-ntrol Circles Save Lockheed Nearly $3 Millionin Two Years," Quality, May 1977, p. 14.

°sR. S. Greenberger, "Quality Circles Grow, Stirring UnionWorries," Wall Street Journal, Sept. 22, 1981, p. 29.

"Information from Quasar. Thus far, the company views itsexperience in the United States with QC circles as successful,but perhaps. not so successful as in Japan. For examples of otherexperiences in electronics, see J. D. Couger, "Circular Solutions,"Datamation, January 1983, p. 135.


text that includes enterprise unions, a relativelyimmobile work force, and seniority-based wageincreases. Not all the elements in the Japaneseapproach or in QC circles themselves are like-ly to prove attractive to workers and managersin the United States.

Japanese and American ManagementStyles: How Much Difference?

Do Japanese firms operating in the UnitedStates exhibit a distinctive management style?Or in adapting to the new setting do they actmore like American firms? Keeping in mindthe structural differences that have been out-lined,, how different are management .styleseven within japan from those in the United.States? By comparing a foreign subsidiary bothto its parent and to local competitors, variablesof ownership and geography can be separated.This section presents the conclusions of a studyof managerial differences among U.S.' andJapanese firms. The survey sample includedupper and middle managers ',Pm: 1) Japanesesubsidiaries of American companies, 2) Japa-nese-owned subsidiaries in the United States,and 3) both American and Japanese firms intheir home country. Appendix 8A, at the endof this chapter, explores the data on nationaldifferences in management style more system-atically.87

The survey results show that American- andJapanese-owned electronics firms do not di-verge greatly in management style. In manyrespects, managerial practice§ were moreclosely associated with geographical locationthan with ownership; i.e., Japanese-ownedfirms in the United States acted more likeAmerican firms, U.S. subsidiaries in Japanmore like Japanese companies. In itself, thisshould be no surprise, given that foreign sub-sidiaries everywhere are mostly staffed by localpeople. Even if upper managers come from the

e'App. 8A. together with the summary here, is based on a reportprepared for OTA by M. A. Maguire. It includes an independ-ent analysis of data from a project directed by R. T. Pascale.The subset dealing with electronics has been of primary interestto OTA. For a discussion based on dl the data, including otherindustries, see R. T. Pascale and A. G. Athos, The Art of JapaneseManagement (New York: Simon & Shuster, 1981).

parent, there is only so much they can importand implement.

The one respect in which Japanese-ownedfirms in both the United States and Japan standout is their emphasis on employee motivationand participation, and on diffusion of respon-sibility through the ranks. The survey resultsindicate that the anecdotal evidence on Japa-nese concern for employee motivation reflectsa genuine distinction: in terms of the models Ofmanagement style outlined earlier, the Japa-nese approach is closer to the human relationspole. At the same time, the range in behavioracross both Japanese and American firms iswide.

Japanese-owned firms stress communicationand personal interaction both horizontally andvertically. At least some, aspects of consensualdecisionrnaking have been transported to theUnited States. One can question the extent towhich Japanese managers accept and act onthe information received through these com-munication channels, as opposed to using themto manipulate opinion and impose top manage-ment decisions. Nonetheless, in employee sur-veys, managers in Japanese-owned firms bothhere and in Japan were more often describedas sensitive to others and accessible to subor-dinates than managers of American.ownedcompanies. This in itself contributes to employ-ee motivation and satisfaction.

Such behavior patterns can be associatedwith the human relations school of manage-ment. The principal contrast with American-owned firms is along the in formal dimensionsof organizational behavior; there was little dif-ference between the U.S. and Japanese firmssurveyed in terms of organizational hierarchyor formal lines of communication. The distin-guishing features of Japanese management ap-pear to be rather intangible, matters of attitudemore than method.

U.S. subsidiaries of Japanese companies havegenerally found this emphasis on human reia-tions and employee participation to work well:Typically, the firms surveyed have modifiedmanagement techniques imported from Japanto fit the American context without abandon-

337


in3 the human relations thrust. Furthermore,some of 'the best performing American firmsdisplay a similar concern for employee par-ticipation, with the, implicit goal of giving in-dividual workers a stake in the suctess of theenterprise. While it is impossible to cIL;iermine

(he precise degree to which human relations-priented management contributes to the per-formance of particular companies, it does ap-peal to be a common trait in well-managed andcompetitive organizations in both countries.

Summary and ConclusionsCommitment to the development and utiliza-

tion of human resources is closely associatedwith corporate success, and, through this, withindustrial competitiveness. In electronics, U.S.manufacturers have had difficulty filling crit-ical positions in engineering; a concurrentshortage of skilled technicians, while not sowell publicized, could prove as serious a bot-tleneck. At present, the United States seems indanger of falling behind other countries attraining people in the skills needed for high-technology industries like electronics; deficien-cies exist in both public and private sectors.Education, provided first and foremost by thepublic schools, determines the skills andcapabilities that people bring with them to thpwork force. The abilit' to continue learningon the job as well as offalso depends on thequality of that formal edocation. While someAmerican firms provide or encourage contin-uing education and training for their employ-ees, others do little or nothing.

Inadequacies in the education and trainingof the American labor force are growing moreserious. Beginning at secondary levels, thepreparation of Americans in science andmathematics is simply not on a par with otherindustrialized nationse.g., Japan. A smallerfraction of U.S. college students major intechnical fields. While many American univer-sities are, at the moment, limited in their abilityto handle greater numbers of engineering stu-dents, a more fundamental problem is the rel-atively small fraction of the college-age popula-tion qualified to enter such programs. Thetypical U.S. high school graduate is not onlypoorly prepared in mathematics and science,but uninformed concerning technology. Deli-

ciencies in mathematics are most serious; thesedisqualify people at an early age from a broadrange of career opportunities, depriving theNation of a vast potential resource.

For those qualified for admittance, programsin engineering, athematics, and the sciencesoffered by Ame ican colleges and universitiesboth underg aduate and graduateremainunsurpassed. Nonetheless, they have slippedrelatively; di,A2',.aering schools, in particular,are suffering :k1:-om a lack of qualified facultyand from inadequate and obsolete equipment,The press -es of expanding undergraduate en-

qr°ilLiTtillit; Jducation, Continued low enroll-ments

to a deterioration in the

ments of Ph. D. students mean that the short-age of engineering teachers will continue; whatmight have been a transient problem is rapid-ly turning into a serious long-term concern,

Moreover, the average American worker isless prepared than his or her counterpart in anumber of other countries for productive em-ployment in industries like electronics. As aresult, the United States is heading towardmore shortages c` skilled blue- and grey-collarworkerstechnicians, designers and drafts-men, engineering aides, field service person-nel. Lik9wise, many white-collar jobs are filledby people with little understanding of mathe-matics, science, or technologyand with lit-

prepakation for comprehending technicalsubjects even on a lay basis. Meanwhile, un-employment in the United States has been ris-ingin part the result of a mismatch betweenwhat people are able to do and what needs tobe done.

33 0,

Ch. 3 Human Resources: Education, Training, Management 335

One way private firms can compensate fordeficiencies in formal education is to establishin-house training and retraining programs; inaddition to such efforts, many Americar, firmssupport continuing education outside the com-pany. The incentives for such efforts, however,are lower here than 'Western Fir., .pe or

because of .liouldly of the U. laborr ire. The freqc nncy with which Americ nstake new jobs heightens the risk that the com-pany will lose its investment. Nonetheless, anumber of U.S. electronics companies have de-veloped ambitious employee training efforts,and the semiconductor industry is developingprograms in conjunction with universities thatwill help to educate new people, as well as sup-porting the R&D base. Despite their promise,such initiativta will l by thein.,,01,,,,5' he suf-l-icient to nic,ut. the skill requirements of thet.lectronics industry in the years ahead, muchless the broader moods of the U.S. economy.

Coved, ;nit :. UnitMate, and tocalhas traditionally carried the

major responsibility for education-and train-ing; expanded public sector programs for train-ing and retraining appear necessary for build-ing the competitiveness of American industry.As demographic forces tilt the labor forcetoward greater proportions of older workers,retraining will be essential if the talents of mid-career employees are to be effectively utilized.As U.S. industry continues to advance techno-logically, workers who find themselves...dis-placed by structural.change will'be dependentOn retraining to find productiVe employmentelsewhere. As job opportunities shrink forthose with limited skills, men and women withpoorer educations, and without the developed.ability to learn on the job, will more and morefind themselves unemployable. Given the com:petition and mobility characteristic of theAmerican economy, the private sector cannotreasonably be expected to provide the neededtraining and retraining; only government bod-iesat all levelscan take on this responsi-bility. ,

The efforts of private industry begin with thepeople available in the labor pool. In largemeasure, the art of management lies in max-imizing the contributions of existing and pro-

spective employeesto which end a numberof the more successful electronics companies,in the United States as well as Japan, havedeveloped management systems that empha-size employee participation. Giving individualsa voice in decisions that affect them increasesmotivation and commitment to the organiza-tion.

Despite the vogue for Japanese managementtechniques, the human relations approach isin no way unique to Japan or to Japanese cor-porations; the similarities among competitivefirms in Japan and th '?tilted States are morestriking than the differences. Specific mech-aniSms, such as quality control circles or labor-management committees, appear of secondaryimportance compared to less tangible signs ofattentiveness by management to the attitudesand talents of employees.

While many U.S. corporations have devel-oped their own brands of human resources-ori-ented management, others could profit by moreattention to worker participation; Americdnmanagers seem to be gradually realizing thatthey may be underutilizing their employees.Table 73 shows that executives of U.S. firmsrank employee participation as the most impor-tants'ingle influence on productivity. Whetherthey act on such beliefs is another matter; but,of the forces that affect competitiveness, man-agement is the most immediately amenable tochange by individual companies. A renewedcommitment by American companies to the de-velopment and utilization of human resourcescould pay large dividends in international com-petition.

Table 73.Rankings by American Managers ofFactors Contributing to Productivity

Factor Average ran kaEmployee participatton programs 3.61Better communications 4.11Better labor-management relations 4.45Increased training 4.46Quality improvement , 4.81Increased automation 5.02Productivity incentive programs 5.13Cost reduction programs 6.01Increased R&D 6.28a8ased on a scale of .1 to 10, with 1 being the most effective and 10 being the least.

SOURCE: Mechanical Engineering, September 1981.

336 lntemational Competitiveness in Electronics

Appendix 8A.Japanese and AmericanManagement Styles: A Comparison'

Survey ResultsA survey covering managers and oth .1 employees

in four electronics companies provides the basis forthis comparison:

company Al, an American consumer elec-tronics firm operating only in the UnitedStates;company J, a Japanese consumer electronicsfirm with operations both in the United States(J-A) and in Japan (J-J); andcompanies /12-J and A31, the Japanese sub-sidiaries of two American firms, one a manu-facturer of computers, the other of semicon-ductors (not necessarily in that order).

All the firms were high performers in their respec-tive portions of the electronics industry.

The data can be grouped in several ways. For in-stance, a geographE; grouping gives: first, the twoorganizations in the United Statesone American-otvned (A1), and one Japanese-owned (J-A); and,second, the three operations in Japanone Japa-nese-owned (J-J) and two American-owned (A2-J,A3-J). Alternatively, grouping the sample by own-ership yields a set of three American-owned firms(Al, A2-J, A3-J) which can be compared with theJapanese-owned organizations (J-A and J-L. Formost purposes, the ownership distinction is moreilluminating, probably because top managers whoset the tone of an organization generally came fromthe parent firm. In contrast, most of the middle-level managers had been recruited locally; thus inorganization J-A they were largely Americans.

The survey covered both middle and upper man-agers, utilizing interviews as well as writtenresponses. Nonmanagerial employees were alsosampled via questionnaires to gather data on jobsatisfaction. The data must be interpreted with cau-tion because of the small number of organizations.At the same time, the survey results for electronicscome from a much larger body of data covering 10industries; differences across industries were few.

'This appendix is b Ised on "Personnel in the Electronics In-dustry: United States and Japan," prepared for OTA by M. A.Maguire under contract No. 033-1360. The report includes anindependent analysie of data collected for a project directed byR. T. Pascale. Pascale's own treatment. including discussion ofCompanies in other industries. can be found in R. T. Pascaleand A. G. Athos. The Art of Japanese Management (New York:Simon & Shuster, 1981).

A primary objective was to gather information oncommunications and decisionmaking styles. Sur-vey questions were designed to indicate whetherAmerican firms differed from Japanese in the ex-tent to which decisionmaking and communicationscould be described as hierarchical and formal (thehypothetical U.S. model) rather than informal andcooperative (the hypothesis for Japan).

The results show that all the American-ownedfirmsA1, A2-J, and A3-Jrelied more heavily onwritten communications, both here and abroad.More surprisingly, firm J-Athe U.S. subsidiary ofa Japanese companywas in many respects more"Japanese" in decisionmaking and communica-tions than the parent organization (J-J); the datashow a greater proportion of upward communica-tion and a lower proportion downward in theUnited States than in the same firm's home offices.Overall, however, the survey resultstable 8A -1showed much less variation in patterns of commtl-nication among these firms than the pure Japaneseand American models would predict. Additionalsurvey questions indicated that the subsidiary A2-Jis more "dependent" on its American parent, asmeasured by written communications with head-quarters, than the subsidiary J-A was on its Japa-nese parent. -

The survey results also shed light on hierarchyand formalization in the organizational structureof each company in terms of the size /level ratio: thetotal number of employees in the organization di-vided by the number of hierarchical levels. Thelower the size/level ratio, the more formal and hier-archical is the firm's structure. Again, the resultsmay seem somewhat surprising: the Japanese com-pany was the most hierarchical, with its domesticand U.S. operations scoring the same-133 (J-J) and134 (J-A). One of the American electronics firmsmeasured 150 in its Japanese organization (A3-J),little different from the Japanese-owned company.The other two American-owned organizations hadratios of 284 (A2-J) and 533 (A1). In other words,none of the American firms are particularly for-malistic or hierarchical on this measure (which canbe rather sensitive to differences in the overall sizeof the companies compared). Another indicator, theextent to which they make use of written job de-scriptions, found the American-owned companiesranked higher in formalization.


Table 8A1.Responses of Middle and Upper Managers to Questions Dealing WithCommunications and Declsionmaking Styles

Companywide averagesAl JA JJ A2.1 A3J

Questions dealing with manager's own behavior:Number of telephone calls and facetoface contacts per day 81 69 72 24 55Number of written communications per day 10 4 3 8 7Hours in meetings per day 2 2.5 3 2.3 3Percentage of calls to those higher in the organization 21% 25% 23% 14% 36%Percentage of calls to those lower in the organization 40% 31% 31% 56% 37%Percentage of meetings with those higher in the organization 13 %' 16% 41/4 8% 100/aPercentage of meetings with those lower in the organization 64% ;i6% 84% 88% 80%Questions dealing with manager's evaluations of their supervisors' declsionmaking styles:Percentage of decisions Supervisor makes alone 36% 21 % 29% 23% 25%Percentage of decisions supervisor makes after factual Input from subord'nates 25% 30% 20% 40% 25%Percentage of decisions supervisor makes with participation by subor tes 43% 49% 51% 37% 50%SOURCE: M. A. Maguire, "Personnel In the Electronics Industry: United States and. Japan," prepared for OTA under contract No. 033.1360, p. 8.

Responses to questions about characteristicsessential to managerial success revealed a greateremphasis in the Japanese-owned firms on commu-nication within the organization both vertically andhorizontally; this was true both in domestic (J-J) andAmerican (J-A) operations. Managers in the Amer-ican company Al would tend to "make as many de-cisions as possible at his/her level without bother-ing senior management," and "respect the chainof command, discuss ideas with immediate supe-rior before discussing them with members of otherdepartments." In contrast, managers in the Amer-ican subsidiary of the Japanese company J-Athought it important to "communicate extensivelywith managers in other departments;" managers inthe parent firm (J-J) also stressed communication.Within one of the American-owned subsidiaries inJapan, A2-J, the responses indicated a feeling thateach manager should make as many decisions aspossible at his/her own level. Here the surveyresults do cor.--m a difference in management at-titudes between Japanese- and. American-ownedcompanies, with the American-owned electronicsfirms exhibiting a greater degree of independentdecisionmaking even within their overseas subsid-iaries.

Questions calling for a composite picture of themanager immediately above the respondent elicitedseveral distinctions among the five organizations.On eight dimensions, those questioned were askedto describe the actual characteristics of theirsuperiors (not the attributes they would like to see).The managers in the U.S. subsidiary J-A were de-scribed as: "readily accessible to subordinatesseveral echelons below," "permits broad latitudefor subordinates to work out solutions to problems

in the wr, way,' and "sensitive to others whowork for h 1." In the parent firm in Japan (J-J), thetypical manager "tries to achieve consensus" and"permits broad latitude for subordinates," but isalso described as aggressive.

While a reasonably uniform picture emerges forthe subsidiary J-A and its parent J-J, there was muchgreater diversity among the characteristics ofmanagers within the American-owned firms. Thiswas especially notable in company Al. Likewise incompany A2-J, respondents agreed on only onething: that their superiors were aggressive. Coupledwith the similar characterization in the Japaneseorganization J-J, this suggests that, while ag-gressiveness has not always been viewed as cen-tral to Japanese management, it may in fact becommon in high-performing firms in both coun-tries. The survey results do paint a more hetero-geneous picture of the managers in American-owned organizations. American companies oper-ating in japan exhibit some of the traits associatedwith Japanese management, but it is the Japanesecompany which, as expected, has managers whomost strongly emphasize consensual decisionmak-ing and human relations. On this dimension, thecomposite managerial portraits indicate a clear dif-ference in American and Japanese styles.

The human relations school stresses sensitivityto subordinates. Table 8A-2 compares responses ofnonmanagerial employees to questions related tojob satisfaction, together with data on rates of ab-senteeism and expenditures on employee pro-grams. The Japanese-owned firms might be ex7pbated to exhibit a greater degree of manager-ernployee jnteractionpresumably leading to

-greater satisfaction among the labor force. Theresults in tabb 8A-2show that very few workers


Table 8A.2.Data Related to Employee Satisfaction

Location of organizationUnited States Japan

Al J-A J.1 A2-J A3-J

Percentage of workers rating themselves "very satisfied" with their jobs 20% 28% 2% 0 0Percentage of workers rating themselves "satisfied" or "very satisfied" with their jobs 74% 88% 58% 63% 95 °/a

Daily absenteeism 3% 1% <1 % < 1 % 1 °/a

Social/recreational expenditures per worker $1.40 $15 $33 $50 $38SOURCE: M. A Maguire, "Personfml in the Electronics Industry: United States and Japan," prepared for OTA under contract No. 033-1360, pp. 19, 20.

in Japan are willing to describe themselves ashighly satisfied with their jobs, but the picturechanges considerablywith firms in Japan compar-ing more favorablyif "satisfied" responses areincluded.z Japanese firms, known for their com-pany-as-family approach, might also be expected tospend more on social and recreational opportuni-ties for employees. As table 8A-2 indicates, this isindeed true for organizations within Japan, regard-less of ownership. In any case, the results in table8A-2 on job satisfaction should be interpreted withcaution, as such questions typically yield high pro-portions of positive responses. Moreover, clear-cuttelationships between expressions of job satisfac-tion and measured productivity levels are rarelyfou nd .3

The differences observed between subsidiarieshere and parent firms in Japan may result from con-scious decisions to downplay Japanese manage-ment practices. The style that emerges is likely tobe a hybrid of American and Japanese practices.In any event, this conclusion follows from thesurvey data as a whole: there is no sharp contrastbetween the management approaches of American-and Japanese-owned companies. Many of the pat-terns observed are more closely associated with thegeographical location of the organization than withownership. Upper managers from the parent firmtend to adopt many practices of the host country.On some dimensionse.g., accessibility of mana-gers to lower level employeesthe Japanese-owned/firms do stand out. But in other cases, there are noclear distinctions; only on measures of sensitivityto employee attitudes and participation are theseconsistent.

2Japanese workers also express relatively low rates of satisfac-tion with activities such as quality circles. Sec S. Takezawa. etal., Improvements in the Quality of Working Life in ThreeJapanese Industries. (Geneva: International Labour Offi-e, 1982),

98.pp. 77. 98S. E. Weed, T. R. Mitchell. and W. Moffitt, "Leadership Style,

Subordinates' Personality and Task Type as Predictors of Per-formance and Satisfaction in Supervision." Journal of AppliedPsychology. vol. 61, 1976.

Matsushita's Purchase of Quasar

What happens when a Japanese corporation takesover an American firm? Changes in managementpractices might offer insight into the Japanese ap-proach. The purchase in 1974 by Matsushita Elec-tric of Motorola's Quasar divisionwhich pro-duced televisionsprovides a case in point, (Un-fortunately, conspicuous examples of a U.S. firmtaking over a Japanese enterprise are lacking.)

After Matsushita tc-ok control of Quasar, the newowners reorganized the factory operations, locatednear Chicago, invested in new equipment, andbegan redesigning the product line, At the centerof these efforts was the goal of improved productquality. In contrast to the old production system,which relied on as many as seven quality controlinspectors per assembly line, Matsushita adopteda more integrated approach with responsibility forquality spread broadly. By 1980, the defect rate onQuasar's assembly lines was about 2 defects per 100sets, compared to lh defect per 100 sets for Mat-sushita's factories in Japan.4 These quality im-provements were the result of systemwide changes.While resulting from a series of decisions made byMatsushita's management, they comprised farmore than just matters of style or technique.. Forexample, the company's extensive modernization.of the capital plant entailed expenditures of about

0 million for automated equipment, as well as aneritkrely new chassis factory in Mexico.5 Motorolaoffiaials stress that they knew just as well what hadto belone to make the Quasar facility more effi-cient, but had decided to allocate available re-sources to other parts of the company's business.

If Quasar's gains in product quality and plant ef-ficiency came at considerable cost in new equip-

4J. Mihalasky and A. B. Mundell. "Quality and Reliability ofSemiconductors and CTVs: United States v. Japan," Report No.C972 prepared for OTA by Consultant Services Institute, Inc..under contract No. 033-1170.0, p. 77.

ST. C. Hayes. ''The Japanese Way at Quasar," New York Times,Oct. 16. 1981, p. Dl.

342.


ment investments, the emphasis on worker par-ticipation and responsibility for quality is alsosignificant. Quality control circles were not in-troduced until recently, but Quasar employees havebeen encouraged to set their own production tar-gets and to meet in informal weekly discussionsabout plant operations with foremen. Such prac-tices are hardly unique or exotic, but the atten-tiveness to all aspects of the manufacturing processstands out. Still, none of this has helped Quasar ex-pand its market share L,ubstantially.

Quasar, like other Japanese subsidiaries in theUnited States, shows a flexible and adaptive man-agemen` 3tyle, with manufacturing operations andquality control having a central place. Nonetheless,if and when Japanese companies hire still larger1 oportions of American managers, and adapt fur-ther to the U.S. environment, they may becomemore like wholly American organizations.6

A recent study by the Japan External Trade OrganizationtIETROJ on Japaneseowned manufacturing operations in theUnited States indicates that the number of Japanese nationalstransferred to subsidiaries tends to decrease over time. See"Japanese Manufacturing Operations in the U.S.." Japan Exter-

ConclusionBoth the survey results and the Matsushita exam-

ple indicate that well-run organizations tend to beopen to new ideas and methods, including thosecoming from the lower levels of the organization.Distinctions between Japanese- and American -owned firms are fewer and less clear-cut thansometimes claimed. While American employeesmight resist some of the techniques associated withJapanese management, worker participationevenloyalty to the firmcan be fostered in a variety ofways. Some of these methods smack of paternalism,but not all. As a number of American firms haveamply demonstrated, worker participation and at-tention to human relations can be a big help inbuilding competitive organizations.

nal Trade Organization. September 1981. Data on managerialstyles collected for the JETRO study confirm the trends describedhere: Japanese subsidiaries evolve styles that mix features com-mon to the Japanese model with other practices more charac-teristically American.

CHAPTER 9

Employment Effects

Contents

PageOverview 343

Impacts of Technical Change in Electronics on Employment 344The Automation Debate of the 1950's and Since 344Factors Affecting Employment Levels. -, 347

Employment Trends in the U.S. ElectrOnics ndustry 349Consumer Electronics '353Semiconductors ,. 356Computers \ 357

Effects of Import Penetration and OffShore Assembly 359Consumer.Electronics 1\

-,Semiconductors360362

Foreign Investment in the United States \ 363

Projections of Employment Within the Electronics In1dustry 364

Future Employment Patterns in Other Industries 366Manufacturing 367Services \ 370


List of Tables

Tobit Nu. Page74. Employment in Selected Portions of tile U.S. Electronics Industry 35075. Predicted Growth Rates by Occupational Category Over the 1980's 36576. Occupational Distributions in Electronics as of 1980 366

List of Figures

Figure No. Page55. Unemployment in Industrial Nations. 34556. Labor Productivity and Employment by Sector of the U.S. Electronics Industry

. (a) Consumer Electronics 351(b) Semiconductors 352(c) Computers 353

57. U.S. Employment in Television Manufacturing 35458. U.S. Employment in Consumer Electronics 35559. U.S. Employment in Semiconductors ar1:1 Related Devices 35660. Effort Levels Associated With ProdUct and Process

Design for Integrated Circuits 35761. U.S. Ethployment in Vacuum Tube Manufacturing 35862. U.S.' 'Employment in Cqmputer Manufacturing '359

Employment Effects

OverviewShifts in the competitiveness of an industry

like electronicsor for that matter technicalchange alonehave both direct and indirectconsequences for employment. In addition tochanges in the labor force requirements bothof firms within the industry and firms that sup-ply it, the effects can spread broadly across theeconomy. jobt,pportunities within the UnitedStates appear or disappear with changes in de-mand for electronics products, with shifts ininternational competitive position, and with in-creases in productivity. These forces interactin complex fashion.

Will continuing developments in electronicscomputers, office and factory automation, in-formation servicescause employment to in-crease or decrease? Such questions have beendebated for years, in the context of this andother industries. The conventional response isthat technical change creates, in the aggregate,more jobs than it destroys. While the kinds ofjobs available will changeas terminals appearon more desks, opportunities for systems ana-lysts (who plan and help operate data process-ing installations) replace those for keypunchoperators, for one instancenew technologycreates new demand fast enough that total em-ployment goes up. The conventional responseassumes that such patterns will continue. But,just because in the past technical change cre-ated more jobs than it destroyed does not meanthat this will be true it 'ile_tvture. Such ques-tions are broader than can be addressed here.Too many forces affect levels of employment,'not to mention skill requirements. Analysis ona detailed, disaggregated basis sufficient to iso-late the influences of electronics (and upon it)would be extraordinarily difficult. This chapterhas more limited aims: to summarize what ispresently known about employment in elec-tronics, both past trends and future prospects.

Within the industry, changes in competitive-ness have immediate consequences for employ-ment. If the U.S. electronics industry declinesin competitiveness, and sales fall in domesticand/or foreign markets, employment will fol-low. If rates of increase of sales drop, employ-ment may also declinedepending on increasesin labor productivity. Similarly, if. U.S. elec-tronics firms expand their overseas productionactivitiesfor re-import or for sales in foreignmarketschanges in domestic employmentnormally follow. As competitive advantagesshift internationally, labor market dislocationscan occur even if the total number of jobs re-mains the same. Such dislocations can includegeographical shifts in demand for workers,along with changes in educational and skill re-quirements; as computers and other electronicsystems have become more sophisticated,white- and grey-collar jobs have expandedmuch more rapidly than openings for unskilledor semiskilled workers.

Shifts in the international competitiveness ofAmerican electronics firms also affect otherparts of the economy. Moreover, structural un-employment can be created by changes in elec-tronics technology that alter the ways goodsare designed and manufactured. Electronictypesetting has reduced the need for skilledworkers in newspaper publishing. Technologi-cal change may create new jobs for supervisoryand maintenance workers, but it is hard to im-agine that as many people will be employed indesigning, manufacturing, and maintaining in-dustrial robots as are displaced by them. Still,net effectsparticularly over extended periodsof timecan seldom be disentangled from theother factors on which employment depends.If aggregate economic growth is slow, and pro-ductivity risese.g., because of in-vestments inlabor-saving equipment like robots, or com-


puter-integrated manufacturing more generallyjobs will be lost unless other sectors of theeconomy, such as services, compensate.

The preceding chapter explored the educa-tion and training of American workers, as wellas management practices which determinehow effectively the talents of the labor forceare utilized and the possibility of shortages ofthose with specialized skills. Chapter 8 in-cluded extensive comparisons between theUnited States and Japan. Here, the focus is pri-marily on the United States, beginning with areview of the automation debates of earlier

years. Next, data on employment trends in elec-tronics are examined in the context of importpenetration, as well as offshore manufacturingby American firms. The chapter surveys em-ployment forecasts for electronics, along withcase studies of impacts on other manufactur-ing and service industries. While there is noway of knowing how aggregate employmentwill fare, technological changetogether withshifts in the competitive positions of Americanelectronics firmswill clearly have major im-pacts on some industries and some jobcategories.

Impacts of Technical Change inElectronics on Employment

The Automation Debate of the1950's and Since

People have worried over technologicalchange because of its impacts on employmentand sometimes actively resisted new technol-ogiesat least since the beginnings of the in-dustrial revolution. The automation scare ofthe 1950's focused on computers taking overthe workplacea fear that has resurfaced,more so in Europe than the United States.Twenty-five years ago, some commentatorspredicted steadily rising unemployment due toautomation; others were skeptical that comput-ers alone would have such grave consequences.Throughout the 1960's,' a number of interna-tional groups, including the Organization forEconomic Cooperation and Development(OECD) and the International Labour Office,continued to study the effects of computers andautomation on employment. As it happened,the industrial nations experienced an upswingin economic growth during the 1960's that putthe automation debate temporarily to rest. Fall-ing levels of unemployment were sufficient in-dication to many that overall demand was thekey to jobs, with structural aspects decidedly

cr; lnnti ae aaorPaatP ripmarirl prim/

I'

The 1970's brought renewed concern; eco-nomic growth slowed and unemployment rose.The trend was sometimes masked by the upsand downs of the business cycle, but by the endof the decade, as figure 55 shows, it was clearthat unemployment had been steadily rising inmost of the industrialized West. Now the ques-tion has become: Will this trend persist?

Rather than mainframe computers as in the1950's, people now point to microprocessorsand microcomputers as the new technologieswith the greatest potential job-displacingeffects.' As was the case 25 years ago, optimistsand pessimists view the consequences of suchdevelopments quite differently. To the opti-mists, labor-saving technology is nothing new.Many more jobs will be created than lost, theysay. Moreover,,in the short term the impactsof microelectronics will not be that dramaticbecause most investments in automated equip-ment come during periods of economicgrowth, when capital is available. As a result,workers may be redeployed but only rarely willlose their jobs. The optimists view structural

'See, for example, C. Norman, Microelectronics at Work: Pro-ductivity and Jobs in the World Economy, Worldwatch I'apor39 (Washington. D.C.: Worldwatch Institute, October 1980); Ad-__ _

13

12

11

10

9

8

Ch. 9Employment Effects 345

Figure 55.Unemployment in Industrial Nations''

E 60E 5

4

1,'7'

United ,Kingdom /

I//I1 Ii,United

States i // .'e // --4'.'.r..' r

"--r ...v.,/ .... , , ,

/ , , ,, 1/ %/ I "c..."' France/ / ^ i , Italy ---

,i

./* -"*.--....of / ."\ \... '''''. I'. -...._

.---.... t, ...../11.-=.......%.%. I\ i i 1"- "11`... .''s,...Z..... / \*.c..//'' .,/ \ / / .........

... a .,,, v. .... 1 j%

.......,, ". ......., IWest

Germany2 Japan

000 NO

0

1.yr

1960 1965 1970

Year

1975

o:,f,,,,,ahale,; the U S. concept Data for the United Kingdom exclude Northern Irelandb1982 data tor an c,nrnthes but ee, Unded Stoles and Italy based on fast 9 months.

SOURCE. Economic Report of the Pres,dent (Washington, D.C.: U.S. Government Printing Office, February 1983), p. 237.

transformation as a process that creates jobsin newer sectors v.-1 the economy: employmentin manufacturing may shrink but opportunitieswill increase in services; as the proportion ofmanual workers declines, the number of white-collar employees grows. Automation, further-more, will free people from some of the worstjobs: dirty, boring, dangerous factory work;sorting and filing; processing checks; perhapseven delivering the mail. To the pessimists, ofcourse. some of these lobs are not so hadand

1980 1982b

todial work, fast foods, selling insurance). Still,from the optimist's viewpoint., the expansionof high-technology industries means more op-portunities for an educated labor force. Com-petition from low-wage, newly industrializingcountries (NICs) need not cause great concern;so long as the world economy continues togrow, industrialized iiations can concentrateon advanced products made by better paid andbetter trained workers, leaving the lowertorilnnInov Qprtr, #1, Mir' 17r c . c .


Others are less sanguine, their skepticismrooted in the belief that the world economy isnow fundamentally different than in the 1950'sand 1960's. "Structuralists" argue that perma-nent shifts spelling chronic unemployment andunderemployment have taken place. Funda-mental to this view is the slow economicgrowth of the 1970's; to the pessimists, suddenrises in energy prices and other shocks to theinternational economy are not enough to ex-plain the slowing pace of growth. They arguethat, at least in manufacturing, the expansionin output needed to maintain cm-rent employ-ment levels has been increasingi.e., that out-put must grow more rapidly than in the pastin orthir to maintain a constant number of jobs.If true, and if this trend persists, it will become'more and more difficult to expand employmentby stimulating demand.2 Since labor produc-tivity in the manufacturing sectors of industrialnations has risen consistently faster than grossnational productiGNPjahe pessimists empha-size that compensating expansion in employ-ment must come from sectors other than manu-facturing. Many also argue that structural un-employment in advanced industrial nationsresults from a permanent shift of labor-inten-sive production to lesser developed countriesand NICs, where wages are low. In the longerterm, this might be a positive force; if interna-tional specialization takes place, the more ad-vanced nations should be able to concentrateon capital- and knowledge-intensive industries,and expand their employment in these sectors.But in the short run it leads to severe disloca-tions, already evident, for example, in consum-er electronics or steel.

The same causes and effectstechnologicalchange, productivity growth, shifts in ;nterna-tional comparative advantage, technology gapsare thus viewed differently by the optimistsand the pessimists. The latter see them as sig-nals of persisting unemployment. Unlike theoptimists, they emphasize obstacles to adjust-ment such as mismatches between the skills

and capabilities of workers and the require-ments of industry (ch. 8). They argue that em-ployment statistics for the United States al-ready underplay the extent of real unemploy-ment, not to mention underemployment.3

The debate between the optimists and pessi-mists ranges far beyond the electronics indus-try. But electronics technology has been a nat-ural locus of concern because it so clearly em-bodies labor-saving advances by which ma-chines perform tasks that people did in thepast. No wonder labor unions it?. the UnitedStates but particularly in Western Eu5opehave continued to raise questions aboutautomation and electronics, 'and sometimes ac-tively resisted new production methods.

The question: "How will electronics technol-ogy affect employment?" is unanswerable. Pos-ing the question more narrowly helps a little:Will continued developments in electronicsdrastically reduce the number of workersneeded in the manufacturing sectors of ad-vanced economies? Will the effects be benefi-cial through elimination of burdensome taskswhile creating new and more interesting jobs?These phrasings still cannot be treated withany precision, but at least are more suggestive.The problem is that no methods exist for deter-mining employment shifts caused exclusivelyby technical change. Too many other forces areat work. A second analytical probiem relatesto the type of employment impact. Advancesin electronics may eliminate a job in oneplantbut a similar job may open in a nearbyfirm or in a distant city. Alternatively, a dis-placed worker might be able, to find employ-ment only after retraining, or even reeducation.

'One study has claimed that 80 percent of American workersare "misemployed"i.e., are doing jobs for which they are ill-suited. See W. W. Harman, "Chronic Unemployment: An Emerg-ing Problem of Postindustrial Society," The Futurist, August1978, p. 213.

Leontief paints a grim picture 'of the effects of technologicalchange, mismatch, and misemployment:

To argue that workers displaced by machines show\d necessarilybe able to find employment in building these machipes does notmake more sense than to expect that horses displaced b§ mechanical


From the perspective of the individual, geo-graphical moves or retraining can aggravatewhat is already a severe blow on psychologicalas well as more tangible grounds.

Factors Affecting Employment LevelsDirectly or indirectly, the ability of American

firms to compete internationally links Many ofthe forces that affect employment. Increasingsales here and abroad provide the foundationfor a growing labor market, with aggregate ex-pansion creating new job opportunities unlesslabor productivity goes up even faster. Conven-tional methods of forecasting labor market de-mand begin with output projections. In a givensector, output and employment will depend incomplex fashion on aggregate demand; in a pe-riod of economic downturn, job opportunitiescan still increase in some industries. While thishas often been true in electronics, recessionarypressures during 1981 and 1982, as in 1974 and1975, show that the semiconductor industry isfar from immune from sales slumps and lay-offs.

For years, the interrelation between employ-ment and inflation was pictured in terms of thewell-known Phillips curve, which showed thathigh rates of inflation tended to correspond tolow rates of unemployment, and vice versa. Butby the end of the 1970's, the American econ-omy 'seemed prone to simultaneous inflationand unemploymentanother gloomy portentto those on the pessimistic side of the struc-tural unemployment question. One reason iswage and price rigidity. When demand falls,companies are reluctant to cut prices as ameans of expanding output, workers reluctantto accept pay cuts to reduce costs. Rather thangreater output and employment at lower wageand price levels, prices stay highaggravatinginflationsales drop, output must be cut, andworkers are laid off. Nonetheless, recent wage.concessions in the steel and automobile indus-tries show that adjustment is possible if theslump is serious enough.

Photo credit: PC

Final adjustments during color TV assembly

Employment is closely linked to labor.-produ `ivitycommonly measured in terns ofoutput per man-hour. If firms can pr ducemore with the same amount of lab r, theeconomy as a whole expands and so oes in-dividual purchasing power. Growl in pur-chasing power can create new dem,and whichwill in turn create new jobs; thus increases inproductivity do not of themselves/result in em-ployment losses. But if the overall economy isstagnant or growing only slowly, productivitygrowth in a given industry can well lead, notonly to decreasing job opportunities in that in-dustry, but to net job losses within the econo-my.

As this implies, sectoral shifts must be con-sidered. A worker rlignlared by ricino

.-


tivity and foreign competition in consumerelectronics finds little solace in growth else-where in-the economy. Similar patterns appearat higher levels of aggregation. As pointed outin chapter 5 (see fig. 32), employment in bothmanufacturing and agriculture has shrunk rela-tive to services in the OECD nations. The serv-ice sector makes an ever-growing contribz:tionto U.S. GNP, and the rate of job expansionthere has been high. What of productivity inservices? Since productivity has grown lessrapidly in services than in manufacturing (al-though productivity in many service sector cat-egories is notoriously difficult to measure).cverall employment levels have been main-tained in part by transfers of labor from manu-facturing to lower productivity service sectorjobs. Of course, factory workers cannot alwaysquickly move to service jobs, nor may theywant toparticularly if the jobs available arelow-paying or menial. The point is that sectoralshifts always irop;y some degree of dislocation.

The impacts of technological change takeseveral forms. Automation, interpreted broadlyas extending to jobs outside the traditionalmanufacturing sector, cuts into the need forlabor. Computers eliminate jobs for file clerks;banking machines displace tellers; instead ofthree people in the cockpit, new commercialaircraft need two. Great Britain's telephonesysterri provides a quantitative example: whenelectromechanical equipment was phased outin, favor of electronic switching during the1970's, employment dropped from over 90,000to 65,000.4

The effects of new technology depend inlarge measure on the motives for its introduc-tion. Investments aimed at rationalizing theproduction procer- by cutting costs, improv-ing efficiency, or adjusting to new conditionstend to cause net declines in job opportunities.The British telephone system is a case in point.On the other hand, technical change may ex-pand output or create new markets, resultingin many more jobs. Henry Ford's moving as-

, rrmin

labor productivity increased and costs were cutto the point that vastly greater numbers of peo-ple could afford to buy cars. Likewise, the in-troduction of color television cut into sales ofblack-and-white sets but expanded overall de-mand for TVs. Many examples could be drawnfrom the computer industry.

The export competitiveness of domesticfirms, as well as market penetration by imports,directly affect employment. Greater sales in ex-port markets mean more jobs at home. On theother hand, an influx of foreign goods may putAmericans out of work. In recent years, con-siderable attention has focused on jobs lost toforeign low-wage industries making productssuch as TVs or textiles and apparel. Never-theless, competition with advanced nations canbe equally importantevident in productsranging from automobiles and machine toolsto integrated circuits and aircraft. As industrieslike electronics become more thoroughly inter-.-national in character, it is seldom easy to disen-tangle the costs and benefits flowing fromshifts in competitive strength. Overseas pro-duction by American firms can be viewed asa loss in domestic job opportunities; it can alsobe seen, in at least some cases, as an entree intonew and expanding foreign markets (see app.B on offshore manufacturing for an outline ofthe complexities of such judgments).

Finally, employment levels always depend tosome extent on the fit between the demand formanpower and the skills and capabilities of thework force. Structural shifts affect not only theemployment levels in various economic sec-tors, but the kinds of people needed. In theUnited States, the unemployed Youngstownsteelworker may neither be qualified nor desireto move into a Silicon Valley electronics com-pany, especially since the pay is unlikely to bevery high. In advanced economies, growth inservices has led to a variety of changes in labormarkets. In Sweden, for example, as the econ-omy has grown and the service sector ex-

inhr, frrrn nnrtirinatinn among older


lieve that none is available.5 At the same time,women have joined the labor forces of the in-dustrialized nations in greater numbers, tak-ing many of the service sector jobs.

The match between supply and demand inthe labor marketnever perfectis thus an in-trinsic part of the employment question. Tosome extent, problems of skills and training arethose of response time; people's choices maylag new opportunities, as may programs ofstudy in educational institutions (ch. 8). Short-ages of entry-level electrical engineers in theUnited States have reflected, not only r,..pidgrowth in demand for the products of the elec-tronics industry, but slow response within theeducational system to new labor market de-mand. This is one way in which employmentis affected by public policies, at least to the ex-tent that schools and universities depend ongovernments (including State and local) for re-sources. Government programs can also helpmen and women who find themselves unem-ployed or underemployed develop new skillsand find new jobs. Adjustment is but one ofseveral avenues; during the 1930's, the FederalGovernment instituted many programs to ex-pand employment. These massive public worksefforts drew support from Keynesian theory,which held that demand stimulation could helpensure full employment.

Despite the experiences of the Depression,and the many job programs since, the UnitedStates does not have a comprehensive man-power policy at the national level. Althoughsome States have set up worker training pro-

°I1. Berg lind. "Unemployment and Redundancy in a %lost-Industrial' Labor Market," Work and Technology, M. R. Haugand J. Dofny (eds.), Sage Studies in International Sociology(Beverly Hills: Sage Publications, 1977), p. 201.

grams to help attract industry, retraining hasnever been approached systematically, in strik-ing contrast to nations such as West Germany;in addition to the vocational programs men-tioned in the preceding chapter, the GermanLabor Market Office matches unemployedworkers with openings through a nationwidecomputer survey' There are no parallels in theUnited States.

This brief review illustrates the difficulty ofassessing the consequences of changes in tech-nology or competitive position even in a singleindustry like electronics. First, many of the fac-tors are interrelated. How can shifts in compet-itiveness be isolated from the effects of aggre-gate economic growth, which determines de-mand for the industry's products? How directlymust gains or losses of jobs elsewhere in theeconomy be linked to changes within the elec-tronics industry (e.g., new technologies) tojustify an attribution to electronics? Should vir-tual employment and unemploymentjobs thatwould or would not exist in the absence ofchanges in electronics technologybe in-cluded? Finally, which impacts are most sig-nificant? Those on individuals? On companies?On entire industries? Or are all three of com-parable importance? What of regional disloca-tions? There can be no easy answers to thegeneral question of whether continuing devel-opments in electronics will have positive ornegative consequences for employment in theUnited States.

The following sections look in more detailfirst at changes within the electronics industry,then at effects on other sectors.

°L. Dobyna, "America Works When America Works," NBCWhite Paper, Juno 25, 1981.

Employment Trends in the U.S. Electronics IndustiyChanges in employment within any one in- declined from 26.0 to 21.8 percent of the

titoefr,, fnien 'Anne, I.. I 1. i i r_ P .1

-350 international Competitiveness in Electronics

sectoral basis; the number of jobs in the U.S.consumer electronics industry has declined,while in computers and semiconductors ex-panding output has brought rising employ-ment.

Analysis of such trends depends on how theindustry is defined and subdivided. For in-stance, data published by the Electronic Indus-tries Association (EIA) show 1.6 million work-ers in the entire industry in 1982.8 EIA, how-ever, bases its tabulation on very broad SIC(Standard Industrial Classification) categories.Among these is SIC 367, "electronic compo-nents and accessories," which has nine sub-divisions. Only one - 3674,. "semiconductorsand related devices"is among the portions ofthe electronics industry that OTA has focusedon, otherse.g., "electronic coils, resistors, andcapacitators"being less illuminating in termsof international competition. Therefore, discus-sion of employment in the rest of this chapteris limited to the following four SIC categories:9

3651Radio and Television ReceivingSets, Except Communication Types. (De-spite the title, this SIC group includes morethan just radios and TVs, extending tonearly all home entertainment or consum-er electronic products; consumer audio

',Electronic Market Data Book 1983 (Washington, D.C.: Elec-tronic Industries Association, 1983), p. 144. This is the total ofLabor Department employment figures for four Standard Indus-±: `al Classification categories: SIC 3651 (radio and TV receivers),366 (communications equipment), 367 (components), and 3573(computers). Communications, with more than 550.000 employ-ees in 1982, makes up one-third of the total.

°Defined in Standard Industrial Class.ification Manual 193'2(Washington. D.C.: Office of Management and Budget, 1972),pp. 190 (SIC 3651), 193 (SIC 3674). 192 (SIC 3671). and 180 (SIC3573).

equipment, public address systems, andamplifiers nor musical instruments all fallwithin SR, 365143674Semiconductors and Related De-vices. (This category includes virtually alltypes of microelectronic components,ranging to solar cells and bubble memo-lies, those manufactured by captive plantsas well as merchant firms.)3671Radio and Television ReceivingType Electron Tubes, Except Cathode Ray.(Virtually all vacuum tubes are includedexcept for TV picture tubes and other cath-ode ray tubes, and special purpose devicessuch as klystrons or X-ray tubes.)3573Electronic Computing Equipment.(Processors and peripherals of all types fallinto SIC 3573.)

In referring below to these S!C categories,more inclusive namese.g., consumer elec-tronics for SIC 3651have been adopted. Bothsemiconductors (3674) and the vacuum tubesthey have largely replaced (3671) are examined,so that growth in the first category can be com-pared to contracticm in the second,

During the 1970's, employment grew in twoof these four SIC categories, as table 74basedon data gathered by the Bureau of Labor Statis-tics (BLS)shows. In microelectronics, em-ployment has doubled, and in computers it hasgone up even faster, while the consumer elec-tronics category has shrunk. (Most of the con-traction in vacuum tube production predates1972.) In 1982, the nearly 800,000 workerscovered by the SIC codes in table 74 totaledslightly more than 4 percent of the 19 millionmen and women in tho U.S. manufacturing

74.Employment in Selected Portions of the U.S. Electronics Industry

Number of employees and percentage ofproduction workers (In parentheses)

SIC category 1972 1980 1982°

3651, consumer electronics 114,500 (74%) 85,900 (70%) 74,400 (67%)3674, microelectronics 115,200 (51%) 226,900 (44%) 230,000 (40%)3671, vacuum tubes 46,400 (70%) 42,600 (52%) 43,400 (61


work force. making even this portion of theelectronics industry larger than, say, steelmak-ingwhich employs half a million.*

Figure 56 compares trends in labor produc-tivity and employment (for production workersonly) over the past decade for each of the cate-gories except vacuum tubes. In all three charts,productivity is given as value-added perproduction-worker hour in real, Ltflation -ad-justed terms. Productivity growth in consumerelectronics, figure 56(a)where employmentdeclinedhas paralleled the all-manufacturingaverage, growing'slightly faster in earlier years.In contrast, computer manufacturefig. 56(c);;hows the most rapid rise in employment; the

"BLS figures for the first 10 months of 1982 show 18.9 millionworkers in manufacturing-11.2 million in durable goods, 7.7in nondurabtes.

20

number of jobs doubled, with productivity ris-ing almost as fast until the mid-1970's. Past thispoint productivity growth has slowedbut, aspointed out in chapter 5, productivity trendsin terms of value can be misleading when tech-nical change is as rapid as it has been in thedata processing industry. Even so, value-addedproductivity in computer manufacturing hasrisen much more rapidly than for U.S. manu-facturing as a whole. Many jobs have also beencreated in semiconductors, fig. 56(b), whereproductivity gains were again substantiallyabove the all-manufacturing average. The cy-clical nature of employment in the semicon-ductor industry distinguishes it from both con-sumer electronics and computers; the sensitivi-ty of semiconductor production to recessionis magnified by the tendency of purchasers toquickly cut back on orders when their own out-

Figure 56.Labor Productivity and Employment by Sector of the U.S. Electronics Industry

(a) Consumer Electronics (SIC 3651)

/14I

I

I

I Productivityall U.S.

I manufacturing

Productivity--consumerelectrcinics

110,000

100,000

90,000

O

80,000 (SO'C3

2a.70,000 .6

wE

60,000

50,000


2,0

20

(b) Semiconductors (SIC 3674)

90,000

80,000

Productivityall U.S.manufacturing

'

, - / iiI I

//r II /

1 /rI1 /

A /% , Employment

% i produclion

ti/ workers

---10 '''

01965

Productivitysemiconductors

1970

put drops, sometimes to double-order in up-swings for fear of shortages.

As the plots in figure 56 demonstrate, the por-tions of the electronics industry that showed thehighest rates of productivity growth also ex-perienced the highest rates of employmentgrowth. Increases in productivity were associ-ated with the creation of jobs, not their elimina-tion. The reason is simple: output in computersand semiconductors grew at very high rates,spurred by exports as well as domestic sales.The domestic market for radios and TVs grewmore slowly, exports were small, and importpenetration has been severe.

As the cases of computers and microelec-frnnirc when rates of chance in tech-

1975

Year

0

O70,000

00

2

60,000 6.-0

7

50,000

40 0001980

drop while productivity climbs, particularly ifcoupled with rapid technical change and highinvestment. But as the examples from electron-ics in figure 56 illustrate, there can be a greatdeal of variation across sectors: productivityrises at different rates; sometimes employmentgoes up, sometimes down. Still, over time, tech-nologically progressive U.S. industries havegenerally experiencednot only above-averageproductivity gains, decreasing real prices, andincreases in salesbut relative increases in em-ployment as well." While an increase in em-

"Denison and others have studied the contributions oftechnological change to economic expansionfor example. E.Denison. Accounting for United States Economic Growth(Washington, D.C.: Brookings Institution, 1074). For an analysisof trends in electronics, see W. Kendrick.--Impacts of RapidTechnological Change in the U.S. Business Economy and in the

40

35

30

25

Ch. 9Employment Effects d 353

(c) Computers (SIC 3572)

Productivitycomputers

140,000

it 130,.200

/11\ EmploymentA / production/ // / workers

/ \//

\ i// ,/ -----/ . '''s ooe. ..%

/ ----- ' '151 / V ...-''/ .-/ .....- Productivity-

1..all U.S.el............ .1.-----'°....."....

i

manufacturing

le1967 1970

Year1975

fal

120,000

u110.000

03

10D,000 2

090,000 a

0

80 000 6

70,000

60,000

50,000

1980

NOTE Census Bureau ',owes for employment may not agree exactly with BLS tipures cited elsewhere in this chapter. BLS figuresare collected at the plant level andmay include some workers from t the r SIC categories if a plant makes products that far under several categories Census figures are used Inthese charts forconsistency with productivity data from the Census Bureau Val-radded figures have been converted to constant dollars using the implicit price deflator forconsumer durables.

SOURCES 1965-71 1977 Census of Manufactures. 1978-80-1980 Annual Survey of Manufactures.

ployment is not inviolably associated with thedevelopment of new technologies and produc-tivity growth, the pattern is not an uncommonone. That employment goes up does not, ofcourse, mean that adjustment problems disap-pmrbut it can provide leeway to deal withthem. The next sections examine employmentby sector in more detail.

Consumer ElectronicsTrends in Employment

which was especially precipitous over the early1970's (as noted on the plot, the data cover TVsonly, not consumer electronics as a whole).Jobs for production workers dropped by halfbetween 1971 and 1981. Over these years, anumber of U.S. manufacturers either mergedwith Japanese or European producers or leftthe business. On the other hand, the industrynow includes more than 10 foreign companieSwith assembly operations in this country thatcontribute to the employment totals in thefigure.


60,000

50,000

40,000

30,000

20,000

10,000

01966

Figure 57.U.S. Employment in Television Manufacturing

All employeesProducWn workers

1970Year

1975 1980

aMonochrome production in the United States had dropped to low levels by 1975.

SOURCES: 1966-70 Television Receivers and Certain Parts Thereol (Washington, D.C.: U.S. Tariff Commission Publication 436, November 1971), p.1971-75 Television Receivers. Color and Monochrome. Assembled or Not Assembled. Finished or Not Finished. and Subassemblies Thereol(Washington, D.C.: U.S. International Trade Commission Publication 808, March 1977), p. A-117. 1976, 1977Color Television R..ceivers: U.S. Production,Shipments. Inventories. Imports. Employment. Man-Hours. and Prices. Fourth Calendar Quarter 1977 (Washington. D.C.; U.S. loternationa: Trade Commis-sion Publication 866, March 1978), table 5. 1978, 1979Color Telet,ision Receivers: U,S. Production, Shipments, inventories, Imports, Employment, Man.Hours, and Prices, Fourth Calendar Quarter 1979 (Washington. D.C.; U.3. International Trade Commission Publication 1036. February, 1980). p. A-7. 1980,1981Color Television Receivers: U.S Production, Shipments, Inventories, Exports, Employment, Man-Hours, and Price:, First Calendar Quarter 1982(Washington, D.C.: U.S. International Trade Commission Publication 1245, May 1982), table 5.

ago; the causes of the employment declines infigure 57 extend well beyond import competi-tion or offshore assembly, with technologicalchange a major force. Although the contribu-tions of the various factors cannot be quantifiedwith any precision, the spread of solid-statechassis designs and associated manufi,methods dramatically reduced employment re-quirements in the industry.

Figure 57 includes only those people in-vnlved in TV manufacturing. Television ac-

tronics market (ch. ,4, table 8), and rathbi lessTin terms of jobs. Total employment in SIC

3651which covers many other consumerelectronics productsis considerably greater,as shown in figure 58. Still, the number ofworkers here has been in decline since 1973,for similar reasons.

ProductivityAs domestic output of TVs grew over the

years covered by figure 57 (see ch. 4, table 9),

140,000

120,000

(al 100,000cu

O

80,000

Oiv

E7

60,000

Ct. 9 Employment Effects 355

Figure 58.U.S. Employment in Consumer Electronics (SIC 3651)

7" N.

40,000t All employees---- Production workers

20,000

0 /V11j___1960 1965 1972 1973 1974 1975 1976

YearSOURCES: 19601965-1977 Cerius of MOnufactures! 1972.82Bureau of Labor Statistics.

\ \ --- ....

output divided by the number of productionworkersjumped from 150 sets per worker in1971 to 560 in 1981. In terms of value-addedper production worker, productivity was up byabout .40 percent during the decade7-.a trendnot far different from that for the broader con-sumer electronics category seen in figure56(a).* During this period, the proportion ofdomestic value-added dropped as Americanmanufacturers shifted labor-intensive opera-tions to developing countries; whether madeby American- or foreign-owned companies,TVs produced in the United States now includemore imported components and subassem-blies. Because of these trends (table 13

'In terms of .:;onstrit 1972 dollars, annual value-added perproduction worker in TV manufacturing went from $22,200 in1971 to $31,600 in 1977, falling to $27.300 in 1981. See 1977 Cen-sus of Nfanufactures: Communication Equipment, IncludingRadio and TV. MC77 -I -36D (Washington, D.C.: Department ofCominerce, jutie'1980). p. 36D-5 and 1982 U.S. Industrial Outlook(Washington, D.C.: Department of Commerce. January 1982), p.343. C:onversions to 1972 dollars were made using the implicitprice deflator for consumer durablesEconomic Report of the

1977 1978 1979 1980 1981 1982

in ch. 4 illustrates the rise in imports of in-complete sets and subassemblies over the lat-ter part of the 1970's) simply dividing the totaloutput of TVs by the number of employees con-siderably overstates productivity gains.However, the value-added'productivity meas-ures adjust for this.

Thus, there is no question that ; ductivityincreased considerably during th '1970's, theresult of design changes and autos ation drivenby competitive pressures (ch. 6). As manufac-turers moved from monochrom' `o color pro-duction, they shifted to more highly automatedmanufacturing facilities. Somewhat later, re-designed solid-state chassis cut the number ofparts, hence the labor content; only 6 percentof the color TVs made in the United Stateswere solid-state models in 1970, but by 1976essentially all had been redesigned aroundtransistors." A good part of the productivitygrowth over the 1970's resulted from changes

356 International Corbpetitiveness in Electronics

in chassis design and associated manufactur-ing methods.

Although productivity gains in consumerelectronics have contributed to declining em-ployment; the composition of the work forcehas not changed greatly. As table 74 and figure58 both illustrate, the ratio of production work-ers to nonproduction workers has decreasedrelatively slowly. In TV manufacture ratherthan consumer electronics as a whole, the shifthas been greater, mostly taking place by themid-1970's (fig. 57). The semiconductor indus-try, for one 'example, has seen more rapidchanges in skill mix (table 74).

Imports' and Offshore 'Manufacture

Earlier chapters described the inroads madeby imported TVs, both monochrome and col-or. Few black-and-white sets are now manu-factured here. Orderly Marketing Agreements(OMAs) restricted imports of color sets duringthe period 1977 to mid-1982, but figure 57shows that the quotas did not arrest employ-ment declines. Still, jobs would have been losteven faster without OMAs.

American consumer electronics firms relo-cated many of their manufacturing operations

250.000

200,000a)a)

0

a)E 150.000

005

E 100.000z

50,000

to low-wage offshore locations during the1970's. While there are no precise figures onforeign workers employed in these plants, theDepartment of Labor believes that the numbermay be over 30,000more than employed indomestic TV operations.12 These people substi-tute quite directly for American workers.

Semiconductors

Since the mid-1950's, employment in semi-conductor manufacture has grown rapidly,from a few thousand when production of semi-conductor devices was just getting underway,to well over 200,000figure 59. These totals in-clude captive manufacturing. During twoperiods-1969-72 and 1974-76employmentdropped sharply as a result of recession.

As figure 59 also shows, the proportion ofproduction workers in the domestic industryhas declinedfrom 66 percent of the total workforce in 1963 to 40 percent in 1982. Majorcauses include the transfer-of production oper-ations offshore and advancing technology.More complex manufacturing methodsin-cluding automationhave increased the rela-tive need for technicians and other nonproduc-

ulnformation from Department of Labor.

Figure 59.U.S. Employment in Semiconductors and Related Devices (SIC 3674)

All employees----Production workers

0

1963 1970 1972

tt

1973 1974 1975 1976 1977 1978 1979 1980 1981 1982

Ch 9Employment Effects 357

tion workers. High levels of research and devel-opment have contributed to expansion in non-production ranks; the number of man-hoursdevoted to integrated circuit design has beenincreasing exponentiallyfigure 60. Techno-logical advance in microelectronics has thusbeen paralleled by a decrease in semiskilledand unskilled employees relative to skilledworkers and professionals in U.S.7based man-ufacturing. The result has been an "upskilling"of the domestic labor force. Employment op-portunities for technical personnelengineers,scientists, technicianshave grown rapidly. Asthese trends continue, the proportion of pro-duction workers in domestic semiconductoroperations will fall even more.

American semiconductor firms transferred"back-end" operations overseas at a rapid paceduring the 1960's, with more than 50 foreignmanufacturing plants established during thedecade." While point-of-sale plants have argu-

"A Report on the U.S. Semiconductor Industry (Washington.D.C.: Department of Commerce. September 1979), p. 84.

Figure 60.Effort Levels Associated With Productand Process Design for Integrated Circuits

300

200

E

ca.'

100

01960 1965 1970 1975

ably small impacts on domestic employment,offshore investments driven by lower wagesdirectly displace American workers, just as inconsumer electronics. Offshore manufacturingalso contributes to the declining proportion ofproduction employees in the United States. Un-skilled assembly labor accounts for most of thejobs' overseas; U.S. firms employ about three-quarters as many people in their foreign plantsas they do here: around 180,000, of which morethan 80 percentas many as 150,000are pro-duction workers.14 Among U.S. merchant semi-conductor firms, perhaps 90 percent of allassembly work is performed overseas.15

Many U.S. companies make semiconductorssolely for internal use, but no disaggregationof employment data is available for these cap-tive facilities. While most produce specializeddevices in relatively low volumes, with con-sidera'o)e variation in month-to-month levels,IBM is a large producer and large employer.Because some of the overhead and administra-tive tasks associated with captivcfnanufactur-ing may be performed elsewhere in the firm,the proportion of production workers is prob-ably higher than in merchant manufacturing.

As semiconductor production grew, the vac-uum tube industry (excluding cathode raytubes, hence TV picture tubes) declinedfigure61. While tubes still find specialty applications,by the early 1970's, substitution of semiconduc-tors had caused domestic employment to dropby one-third from the peak level of 1966.Although jobs in tube manufacturing have beenlost to technical change, far more people arenow employed in making semiconductors thanwere ever employed in making vacuum tubes.

ComputersComputer manufacturing, like microelec-

tronics, has seen rapid employment growthwith simultaneous productivity improvementalthough, as emphasized in chapter 5, pro-ductivity measures can be misleading where

"Summary of Trade and Tariff Information: Semiconductors1980 (U.S. International Trade Commission Publication 841, Control


80,000

70,000

60.000

50.000

40.000

30.000

Figure 61.U.S. Employment in Vacuum Tube Manufacturing (SIC 3671)

20,000

10.000

All employees----Production workers

1962 1964 1966 .., 1968 1970

SOURCE: Bureau of Labor Statistics.

the product changes so much. Regardless, advances in computer systems have created vastnumbers of jobsnot all in computer manufac-turing. Many of these new jobs have originatedin the user community, and in softWare pro-duction. Figure 62 illustrates job growth in theindustry itself, including peripherals. Evenmore so than in microelectronics, the trend hasbeen away from production employees andtoward skillcci workers and white-collar profes-sionals.

Unlike either semiconductors or consumerelectronics, employment in computers and pe-ripherals has not been greatly affected by im-port penetration or offshore production. ManyAmerican computer firms have invested over-

'

1972

Year1974

\ ;rowam=

1976 1978 1980 1982

generally served foreign markets. As in semi-conductors, some of this foreign productionmay substitute for exports from the UnitedStates, but overseas sales are often tied to localproduction, limiting the extent to which point-of-sale plants displace domestic jobs.

The summary above of employment trendsby sector in the domestic electronics industryshows that the number of jobs has inzreased,but not everywhere or uniformly. Increases insemiconductors and computers have morethan offsetin magnitudethe declines in con-sumer electronics and vacuum tubes. The com-position of the work force has changed; em-ployment gains have been greatest for nonpro-


Figure 62.U.S. Employment in Computer (and Peripheral Equipment) Manufacturing (SIC 3573)

450.000

400.000

350,000

300,0000

..._ 250,0000

200,000

150,000

100,000

50,000

0

All employees

Production workers

ea.

.

1955 1960 1965 1967 1968 1970 1972 1974

Yea,SOURCES: 1955-85-1977 Census of Manufactures. 1968.82 Bureau of Labor Statistics.

1976 1978 1980 1982

Effects of Import Penetration and Offshore AssemblyAn increase in imports or a transfer of man-

ufacturing operations offshore can cut intodomestic job opportunities. The United Statesis importing more manufactured goods of alltypes, not only consumer electronics and semi-conductors, making the import penetrationquestion especially timely. Moreover, to laborunions, offshore production amounts to the ex-port of jobs. For policymakers, both phenom-enabut especially imports--have been agrowing concern.

The employment consequences of importpenetration and offshore assembly are felt ina context of global shifts in market structure,implying long-term changes as well as imme-diate impacts on people, firms, and industries.The dynamics are important on both timescales. In expanding markets, firms that canrespond quickly to new opportunities any-where in the world may be able to increase ex-ports and consolidate their positions, aided byproducts that take advantage of new technol-

360 InternationalCompetitiveness in Electronics /

ogles. This happened during the 1970's,/whenAmerican semiconductor firms capitalized onthe shift toward metal oxide semiconductor in-tegrated circuits ahead of their overss rivals.Today, Japan's avowed goal of capturing morethan 30 percent of the world competer marketby 1990 (along with 18 percent of the U.S. mar-ket) reflects a belief that longstanding patternscan be disrupted when growth is rapid.

This section looks more closely at the effectsof imports and offshore production on employ-ment in consumer electronics and semiconduc-tors (neither is important at the moment incomputers). As pointed out in chapter 5, indus-tries 'do not rise or decline in competitivenesssimultaneously; looking at employment on asectoral basis gives only part of the picture, andthen an equivocal one: Still, the sectoral ap-proach is'a valid starting pOint, for reasons thatare discussed. in some detail in appendix B.

The first.question is: What are the causes ofimport penetration? Imports may rise becausedemand exceeds domestic capacity or consum-er preference shifts to foreign:made goods. Jap-anese penetration of U.S.. markets for dynamicrandom access memories (RAMs) is an exam-ple of the first case, TV imports at least in partthe second (imported automobiles are a moreobvious example). In the first case, jobs maynot be lost because of imports, but the rate ofincrease in domestic job opportunities mayslow. In the second case, immediate decreasesin employment are likely.

The full consequences of import penetrationdepend on the industry. Declining output insome induStriesa prominent recent instanceagain being automobilescan have major spill-over effects elsewhere in the economy. As salesof domestic cars lagged, jobs were lost in firmsmaking steel, tires, and components. Some-times companies can limit impacts on individ-uals by allowing employment to declinethrough attrition rather than layoffs; even so,the overall pool of job opportunities shrinks.

The effects of offshore production are nomorn strniehtforward. On the nne hand. all

gross domestic product. But what if firms canonly lower their costs and maintain or expandtheir markets by moving offshorewhether totake advantage of low-cost labor and be betterable to compete with imports, or simply tomanufacture their products nearer the ultimatemarket? Firms weigh a variety of such factorsin deciding whether to invest overseas, al-though ultimate decisions generally turn oncost savings. From the standpoint of the Na-tion as a whole, rather than a particular com-pany, the costs and benefits may be quite dif-ferent. Appendix B discusses the impacts ofoffshore manufacturing on the aggregate econ-omy and outlines the range of effects comparedwith alternatives available to the firm. This ap-pendix includes a case study drawn from theuxperience of an American company which invested in a subsidiary in Taiwan. Briefly, theconclusion of the case illustration is that theoffshore plantestablished to assemble auto-mobile radioshelped maintain competitive-ness vis a vis Japanese manufacturers andprevented even more U.S. :-:bs from eventual-ly being lost. As this suggests, in consumerelectronics the movement offshore by Ameri-can nroducers can be viewed as a defensivereaction to imports. In contrast, the motivationfor overseas manufacturing in the semiconduc-tor industry has been cost reduction and mar-ket expansion driven by domestic competition.The consequences for employment have beenmuch different.

Consumer Electronics

Almost half the consumer electronics marketin the United States has been taken by imports;in addition, many products assembled here de-pend heavily on imported components andsubassemblies. Penetration of consumer elec-tronics markets has coincided with employ-ment decline, as shown in figures 57 and 58.Imports of black-and-white TVs rosé from one-quarter to three-quarters of U.S. sales over theperiod 1967-77. Color TV imports peaked in1976 at a level nearly tenfold greater than in1967. then dropped because of OMAs. A thira


. Today, all U.S. TV manufacturers operateforeign production facilities. In addition to theattraction of low-wage labor, Items 806.30 and807.90 of the U.S. tariff schedules encourageoffshore assembly (ch. 11). During thelast halfof the 1970's, 30 to 45 percent of all color TVimports'entered under Item 807, although finalassembly remains concentrated herein partbecause of foreign investments to avoid the0 A-i mposed quotas'

Despite limits on imports, employment in TVmanufacturing did not recover. In testimonybefore the U.S. International Trade Commis-sion, the International Brotherhood of Elec-trical Workers reported that 20,000 workershad lost their jobs in the TV industry due toimports.lo To what extent are imports to blame,given that domestic productivity improvementsand offshore investments by U.S. firms havealso contributed to employment decline?

It is oversimple to argue that the total num-.ber of foreign workers engaged in.productionfor shipment to the United Stateswhether em-ployed by U.S. or foreign firmsrepresents do-mestic employment loss. In most cases, U.S.consumer electronics firms had little choiceconcerning offshore production. Movementabroad was a defensive reaction, not a strategyeimed at expanding markets and improvingprofitability. To assume that jobs overseassubstitute directly for U.S. employment is tan-tamount to assuming a stable competitive en-vironmentwhich was not the case. Rather,employment declines followed losses in com-

iitiveness; American firms had higher coststhan their rivals, and little scope for develop-ing strategies that would preserve domesticjobs. They pursued the. ,obvious route: in-creased .automation. to raise productivity athoMe. combined with transfers of labor-inten-__,.sive operations offshore. Only some companiessurvived; the other were purchased by moresuccessful manufacturt-,s or left the industry.In this senseas part of a more complex chain

mFestimony lietbre the U.S. International Trade Commissionr.11 P...... it,tre IT A:1(11-1C11.' Intornatinnal RrnthPrhnnd of Elpr-

of eventsimport competition must indeed becounted as the primary cause of job losses inconsumer electronics.'"

But this is not the whole story: Is it possiblethat the ready availability of Item 807.00 re-duced incentives for American managers to cutcosts and improve labor productivity at home?Might U.S. firms have avoided offshore pro-duction by adopting more capital-intensiveautomated manufacturing processes here? Jobsstill might have been lost, but the costs andbenefits would have shifted. The behavior ofAmerican executives is often contrasted withthat of their Japanese counterparts, who recent-ly have faced similar difficultiesi.e., competi-tion fro countries with much lower laborcosts. Some observers have claimed thatJapanese consumer electronics firms have in-vested more rapidly and more boldly in, mech-anized production technologies such as auto-matic component insertion (see ch. 6 for a fur-ther discussion of rates of adoption of automa-tion).

As with many such questions, the truth prob-ably lies somewhere between. The availabilityof Item 807 reduces the pressure to find cheap-er manufacturing methods at home. It is alsotrue that the Japanese, when themselves con-fronted with the rather sudden emergence ofcompetition from other Far Eastern countries,transferred some of their production to lowerwage sites. In part, such transfers were misocaused by the 1977 OMAwhich, bye limitingshipments from Japan, created incentives forJapanese firms to move to export platformsbut Japanese managers exhibit little rehictanceto take whatever steps seem necessary forpreserving hard-earned market positions.

-Along -with__ developing new __production_methodsan uncertain businessJapanesefirms would probably have shifted labor-intensive production abroad in any case, sim-ply for insurance. In this respect. Japanesemanagers have behaved much like Americans.

"For a generally contrary view, see. A. 0. Krueger. "Restruc-

' Competitiveness in Electronics

Photo credit: Control Automation

assembling electronic components

nployment would have dropped,,ithout OMAs, the lesson-7-re-. industriesis that controllingis to provide companies a res-ich they can take measures toompetitiveness is unlikely to im-lent prospects. Indeed, jus:: theto be true, as manufacturers

sts by improving labor produc-:tent\ that they succeed, employ -will even in situations;rowth in output takes place. Asnotection seldom functions asassistance to displaced workers.

etniconductorsof semiconductor devices havelily, exceeding imports by 1982-1 earlier years, the United States

more semiconductors than ituch trends portend job losses?oint, with Japanese manufac-haif or more of the burgeoning:et, will employment in this por-ronics industry suffer as in con-

rapidly. Furthermore, American semiconduc-tor manufacturers have exported much moleactively than consumer electronics firms.

Figure 59 showed the steadily growing em-ployment in the U.S. semiconductor industry.Domestic jobs more than doubled during the1970's; offshore employment probably ex-panded even faster. The question again is: Doimports, or foreign workers employed in theoverseas operations of U.S. firms, stand for jobopportunities lost to Americans? Imports andoffshore manufacturing are more closelycoupled for semiconductors than for consumerelectronics. Nearly 40 percent of U.S. semi-conductor sales are classed as imports, butmore than three-quarters of these are re-im-ports by American firms under Items 806 and807 of the tariff schedules. Offshore produc-tion is central' to the U.S. industry. Still,shipments from Japan have also risen swiftlyover the last 5 years.

The offshore facilities of U.S. semiconductormanufacturers concentrate on the labor-inten-sive steps in the production process primarilyassembly. In the mid-1970's, Finan estimatedthat manufacturing costs for integrated circuitscould be cutlin half through offshore assem-bly.18Cost/p9ce competition has thus been the -.primary motive for foreign °investments; Amer-ican semiconductor firms moved offshore toreduce costs rid expand markets. Moreover,the competiti9n has been largely among domes-tic firms; investments predate Japanese com-petition by a decade and more. If in the caseof consumer electronics, offshore manufactur-ing was a reaction to import competition; insemiconductors the primary motivations wereoffensive. Capital investment requirementshave.been_one_olthe forces_at_work. In order .

to keep up with demand, semiconductor firmshave been unde\r, continnal pressure to add newcapacity (ch. 7). Offshore assembly offered flex-

1taw. F. Finan, "The International Transfer of SemiconductorTechnology Through U.S.-Based Firms," Working Paper No. 118,National Bureau of Economic Research, December 1975, p. 60.The savings are greater for simpler integrated circuits and dis-

. . . .


ibility; firms could avoid the risks of invest-ments in automated equipment that might soonbe outdated, expanding capacity without tak-ing funds from capital-intensive wafer fabrica-tion and testing equipment.

What are the implications for job opportuni-ties? As the case study in appendix B illus-trates, these depend in part on the time hori-zons. Given rising foreign competitiveness inmicroelectronids, offshore production nowhelps meet international as well as intranation-al competition. If overseas manufacturinghelps U.S. firms maintain their Competitive-ness, the net impact on domestic employmentmight he positive over the longer term. Further-more, point-of-sale plants are sometimes ableto sell in markets to which the U.S. parentwould have difficulty in exporting because oftrade barriers. In some instances at least,American firms may thus be able to strengthentheir long-term competitive -.position by in-.vesting overseas, enlarging domestic as wellas foreign employment. Still, in the short term,offshore investments, cut the number of job op-portunities for Americans. In this respect,questions of the impact of Items 806 anc1,807of the tariff schedules are similar to the moregeneral problemisolating the consequencesof foreign direct investment of any type onemployment. Such matters have been investi-gated extensively over the years. The mostcommon conclusion is that direct investmentby American corporations has increased netemployment in the United States: nevertheless,the opposite result is sometimes reached, againdepending on the particulars. In the im-plicationsboth short and long termcan onlybe evaluated on a case-by-case basis, anc, de-

: Tend on assumptions concerning the future.competitive environment for American firms.

Foreign Investment in the United States

Foreign investments here bring yet anotherdimension to questions of domestic employ-

ment. Japanese investment in the United Stateshas grown rapidly, from a cumulative $152 mil-lion in 1973 to $4.2 billion by 1980.19 The desireto open new markets and to ease trade frictionsare among the forces behind this influx. Japa-nese-owned firms now assemble nearly 4 mil-lion color TVs here each year. North AmericanPhilips adds well over a million.

As this suggests, most of the past investmentsin electronics have been limited to consumerproducts. Japanese interests seem bound towiden, however, with plants for assembling in-tegrated circuits the next step. In typicalforeign-owned manufacturing plants, only afew ur--;er management slots are reserved forexec from headquarters. Viewed strict-ly froz:1 an employment perspective, therefore,onshore manufacturing has positive conse-quences for the United States. Viewed morebroadly, the picture becomes mixed: many ofthe skilled and professional jobs remainoverseas.

Generalizations about employment thatwould apply to all parts of the electronics in-dustry are impossible. In the case of consumerelectronics, import penetration is closely asso-ciated with job loss. In contrast, employmenthas grown steadily in both domestic and for-eign operations of U.S. semiconductor firms;overseas investments have helped cut costs, ex-pand markets, and increase competitiveness.Simply in terms of numbers of jobs, expansionin semiconductors and computers has morethan offset declines i consumer electronics.This does not mean, of course, that such trendswill persist indefinitBly. Nor is it any consola-tion to people who find themselves out of work.The rest of the chapter looks to the future.

_

iglapanese Manufacturing Operations in the United States."Japan External Trade Organization. September 1981.

364 International Competitiveness in Electronics_ .

Projections of Employment Withinthe Electronics Industry

Before examining impacts on other parts ofthe economy, this section treats the industryitself, in the context of Bureau of LaborStatistics (13LS) employment projections to1990. The perspective is much broader than thediscussion of possible shortages of engineers.and skilled workers in the previous chapter.

Ideally, projections of future trends wouldbe based, not only on a model for aggregateeconomic expansion, but also on sector-spe-c:ific variablesgrowth in particular productmarkets, demand for workers with certainkinds of skills, levels of imports and exports...Unfortunately, this much detail is seldom at-tempted. BLS projections, virtually the onlyanalyses available with industry-specific out-put, are based on an econometric modelalimited tool, although representative of the stateof the art.20

BLS began making econometrically basedemployment ()injections two decades ago, in-troducing a macroeconomic demand model in1975. Their current procedure includes fivebasic steps: 1) pr,7!;ctions for the economy inthe aggregate; .2) disaggregation of GNP by de-mand categories; 3) distribution of demand bycategories to producing industries; 4) outputprojections by industry sector based on aninput-output table; and 5) forecasts of labor pro-ductivity, total labor hours, and.number of peo-ple employed at the sectoral or industry level.A critical input in terms of employment is theestimated gross demand for the products of an

.gross.. output is divided.. by anestimated productivity level (output per em-ployee-hour) to yield the labor hour projectionfor the industry, and thus employment. Themodel as a whale is sensitive to a wide rangeof assumptions, most fundamentally those forGNP growth. BLS's recent projections havebeen based on GNP increases ranging from 2.4

percent annually (the "low trend") to 3.8 per-cent (the "high trend"). These assumptionscompare with a 197379 average of 2.8 percentper year. BLS has assumed growth in labor pro-ductivity to stabilize at,the rather low levels ofrecent years.21

On this basis, BLS predicts that aggregategrowth in U.S. employment will range from 1.6to 2.0 percent annually over the decade of the1980's, considerably below the 2.7 percentyearly rise for 1975-79. Women will get two-thirds of the new jobs. The durable goods por-tion of manufacturing is expected to growfaster than the all-industries average, non-

urables slower.

Output increases in computers and relatedequipment should lead all other manufactur-ing industries; employment in the computer in-dustry will grow from about 420,000 in 1982to perhaps 600,000 by the end of the decade.If these projections prove realistic, employ-ment in the computer and peripherals sectorwill comprise as much as 3.1 percent of thetotal manufacturing work force by 1990, com-pared to 1.6 percent at the end of the lastdecade. Employment in the electronic compo-nents sector (SIC 367) is expected to grow atabout 2.2 percent per year in both low- andhigh-growth scenarios, well above projectionsfor manufacturing as a whole. In the low-growth scenario, 33 of the 150 industries ex-amined show employment drops. One of theseis radio and TV manufacturing, with an an-ticipated decline_averaging 1.A_percent per yearover the period 1979-90. Thus, if BLS projec-tions prove realistic, past employment' trendsin electronics will persist: there will be con-tinuing decline in consumer electronics, rapidgrowth in computer manufacturing, and con-

"For more detail, see V. A. Personick. The Outlook for In-


siderable expansion in components. Together,these three portions of electronics might, underthe most favorable circumstances, account formore than 7 percent of U.S. manufacturing em-ployment in 1990. The projections are all con-ditional, needless to say, and BLS's approachshares the principal limitation of virtually allforecasting techniques: current trends are ex-pected to continue, breaks with the past seldomanticipated.

BLS also estimates employment by occu-pational category across industries; in allscenarios, white-collar jobs grow faster thantotal employment, blue-collar jobs slower.White-collar workers will make up slightlymore than half the 1990 labor forcethe frac-tion is slightly less nowwith notable increasesin the professional and technical category.22

Table 75 lists occupations in electronics forwhich BLS predicts the greatest percentage in-crease during the 1980's. All are grey- or white-cellar jobs. The nonelectronics categories areincluded for comparison; 5 of the 10 fastestgrowing occupations in the complete BLSlisting are electronics-related. Despite the hit.,hgrowth rates, categories starting from a modestbase will not account for large numbers of newjobs.

==M. L. Carey; "Occupational Employment Growth Through1990," Monthly Labor Review. August 1981, p. 45.

Table 75.Predicted Growth Rates byOccupational Category Over the 1980's

Occupationa

Predicted increasein employment

(1980-90)

Paralegal.Data processing machine mechanic

(9/3°/D

Computer operator 72Computer systems analyst 68Business machine se,iice technician 60Computer programer 49Employment interviewer 47Computer peripheral operator 44Psychiatric aide 40

allonInclusive .stest growing occupations in electronics are listed together withselectea occt.;:ations outside of electronics (in italics) for comparison.

Photo credit: Western Electric Co.

Semiconductor wafers being loaded into furnace0

As shown earlier, the electronics industry ex-perienced a more-or-less gradual shift towardfewer production workers and more white-collar workers during the 1970's, with the big-gest change in semiconductor manufaocturing(fig. 59). Table 76 gives occupational break-downs in consumer electronics,components,and computers according to BLS data for 1980.While BLS expects some further upskilling dur-ing the 1980's, the projections (not shown)which may or may not be well-foundedindi-cate these to be mostly matters .of a percent-age point or two. Note that the SIC categoriesin table 76 are broader than used earlier; theconsumer electronics data cover SIC 365, rath-er than the "home entertainment" subdivision,3651; electronic components, SIC 367, includesall types of components, not just microelectron-ics; and the computer category referred to earli-er, 3573. is a subdivision of SIC 357. Movingthe boundaries of these categories outwardprobably makes little difference for consumerelectronics and computers, but components asa whole are not nearly a, skill-intensive asmicroelectronics; thus the iportions of tech-nical professionals in table , are considerableunderrepresentations for semiconductor firms


Table 76.-Occupational Distributions in Electronics as of 1980

Consumerelectronics(SIC 365)

Electroniccomponents

(SIC 367)Computers(SIC 357)

White. and grepcoilar workers 27.9% 32.0% 59.0%Professional alid technical:

Engineers (and scientists) 3.6 6.2 11.3Engineering technicians 2.8 6.1 9.3Computer specialists 0.6 0.6 6.2Other 3.0 2.8 5.7

Managers 4.7 5.4 9.4Salesworkers 0.7 0.7 1.0Clerical 12.5 10.2 16.1Bluecollar workers 62.1% 63.6% 38.2%Craft 17.3 14.8 12.1Assemblers and machine operators 44.8 48.8 26.1Service workers and others 10.0% 4.3% 2.9%SOURCE Bureau of Labor Statistics.

in the ranks of white-collar workers (not allwith high levels of education or training) orskilled, grey-collar technicians-in contrast toconsumer electronics and components, wherethese jobs make up less than a third of the total.Employment expansion in computers will con-tinue to be most rapid in skilled categories(table 75); numbers of service and repair techni-cians and systems operators will increase,while jobs for keypunch operators-whoseskills are becoming obsolete-will dWindle, aswill work for those without special training.Likewise, in components, BLS estimates that

the number of professional and technicalworkers will grow from 87,700 in 1980 to over117,000 in 1990. In the more mature consumerelectronics industry, the absolute number ofblue-collar workers is likely to decline, as wellas the proportion. Taken together, the trendsindicate a continued shift toward more highlyskilled jobs in electronics. Computer manufac-turing, in particular, will be a leader in employ-ment growth and in demand for new.skills overthe next decade; the picture for this industryforeshadows trends expected elsewhere in theU.S. economy.

Future Employment Patterns in Other IndustriesIf analysis of past trends in electronics is

problematic, looking ahead to the impacts ofelectronics on other industries is a still moretenuous exercise. Yet it is a vital one, for futuredevelopments in electronics have far-reachingimplications for the entire economy. Usefulpolicy guidance could flow from an under-standing of how technological change affectsemployment patterns. Public and private train-ing and retraining programs would benefit ifvulnerable job categories, as well as those forwhich demand will rise, could be more reliably

fined occupational categories. This is expen-sive and time-consuming, demanding a sophis-ticated appreciation of how industry uses tech-nology; in consequence, such studies are sel-dom attempted.

Uncertainties abound. First. past trends-including examples of technical change in in-dustries other than electronics-can offer onlya general guide; there are no guarantees thatcurrent employment patterns-outcomes oflarge numbers of incremental and evolutionarychances -will cersist. Second. many imnats

lug to take one examplemay increase em-ployment in firms designing and building theequipment used, decrease employment in theapparel industry, but perhaps have positive im-pacts on employment at the retail level (onereason might be that custom design would be-come cheaper, with Smaller runs of styles andsizes sold in specialty shops). Attempting totrace such second and third level effects in-volves the interplay of business decisions, eco-nomic and product cycles, imports and ex-pOrtsnot to mention the unpredictable matureof consumer demand. The following sectionsdo not attempt to answer the!" question ofwhether electronicS technologies will have netpositive or negative impacts on U.S. job oppor-tunities, but simply illustrate some of the forcesat work.

European governments, sensitive to the po-tentially negative employment consequencesof electronics and automation, have commis-sioned numerous reports on the subject, withuniformly disappointing results. Micro-levelanalyses exploring impacts on a particular craftor industry are difficult to integrate withmacro-level studies and aggregate economicforecasts. Yet this couplingthe complex andevolving interplay among technical advance,utilization within various economic sectors,and the response of the labor marketis criticalon both supply and demand sides. For exam-ple, companies typically install labor-savingequipment in periods of economic expansibn,when workers can he transferred to other jobsrather ths,i laid off. Over the longer term, then,a given firm can often use normal attrition tohelp manage the size of its work force. Wherethis is the case, direct attribution of decreasesor increases in employment opportunities to

technology can be difficult to defend.23

While forecasting methods do a reasonablejob of predicting employment within either ag-gregate or disaggregate categories as long as

"The authors of a British study write: -Microelectronics tech-nnlnuy will affect manufacturing industry in so many ways that


change is slow and past trends supply prece-dents, the unexpected consequences of newtechnologies escape forecasting methodologiesvirtually by definition. Examples from the pastillustrate little beyond the seemingrandomnessof the impacts of technological change. Thisin itself is an important,lesson, but means thatthe state of the art is such that even well-documented historical case studies can seldomprovide direct policy guidance.

The basic problem is that, even if it werepossible to predict how technical change inelectronics would affect some other industry,there is no necessary relationship betweenthese findings and the consequences for theeconomy as a whole. Building up the pictureon a detailed, sector-by-sector basis would bea vast undertaking. MostsTof the past attemptswhether dealing with manufacturing or serv-ices or bothhave been more limited, fallinginto one of two categories: 1) elaborate butabstract analytical frameworks, typically_econometric; or 2) case studies outlining im-pacts on particular sectors. The first, ex-emplified by the BLS analyses. discussedearlier, have seldom been very illuminating interms of real-world experience. The secondoften yield insights that are useful but limitedto relatively narrow segments of the laborforcebank tellers, coal miners, postal work-ersas illustrated by the case examples thatfollow.

Manufacturing

Many of the studies addressing manufactur-ing begin by distinguishing between productand process applications. These overlap in thesense that computer-based process control sys,terns, to take one example, can be viewed ineithei light. As a "product," they are developedand sold by firms in the capital goods industry.In the alternate view, automated process con-trol is one aspect of an ongoing transformationof production in many industries. Employment

rtnnf re-siirsvA, ini brall xyiraIA,C nitbr mob hrrti-

366. International Competitiveness in Electronics

by market.24 In theory, the greatest gains comewhere new products are introduced into newmarkets. Pocket calculators and video gamesare c\amples. While they may replace othergoodselectromechanical calculators, for ex-ampleto the extent that new products expandmarkets or create new ones, employment willrise. Evi Sting products introduced into newmarkets have parallel effects. The "personal"computer is not a new technology or a newproduct so much as an ad 'ption of microproc-essor-based data processing systems to theneeds of individual households and small busi-nesses. Low-end minicomputers of the early19711'5, such as the PDP-8, were similar in manyrespects to current personal computers, but thePDP -8 was never marketed as such. In contrast,the introduction of new or replacement tech-nologies into old markets often cuts into jobopportunities. Recent and well-publicized illus-trations include electronic switching in tele7communicationsprincipally telephone sys-tems=and electronic typesetting'in the print-ing industry. In essence, these technologiescaused step changes in labor productivity, withsubsequent employment declines. In suchcases, output may expand, but not rapidlyenough, to compensate. In between the ex-tremes of the examples above fall many whichhave mOre moderate impacts on employment.

Several case studies are outlined below, in-cluding those of telecommunications and type-setting, to illustrate typical impacts of elec-tronics-related technologies on employmentpatterns.

The British Telecommunications IndustryThe introduction of electronic switching in

the British telephone system exemplifies thereplitcement case. Employment dropped from90,000 in 1973 to 65,000 by the end of thedecade, Jobs. were lost both in manufacturingand among those employed running the sys-tem. Declining export sales contributed to jobloss in the manufacture of telecommunications

. FelirnvIng NI.' ynti H. Rush, "The Impact of Micro-electronics on the U.K.: A Suggested Classification and Il-lustrative Case Studies." Occasional Papers Series, No. 7, SciencePolicy Research Unit, Universin, of Sussex, June 1978. .

equipment; within the system,fev,e astaers,service persOnn'el, and operators a needed.Further reductions may be in stcre, with fuilyelect -onic equipmentexpected around\1990cutting the work force to as littleas oneAnthits former size.25

Printing

Computerized typesetting provides a secondexample of the introduction of new productsinto old markets. High-speed photo-typesettingequipment, along with typesetting computers,have transformed the printing industry. Theequipment is much less labor-intensive than thehot-metal typesetters that have been replaced,and productivity has jumped. With electronictypesetting, an operator selects type size andstte, column width, spacing, and other layoutspecifications on a video screen, composing anentire page at a time; the older linotype ma-chines, stemming from the end of the 19th cen-tury, prrduced one line of type at a time. Afterelectronic photocomposition had been intro-duced at the New York Times, the Sundaycla-,.sified section could be completed in 20minutes rather than 3 days. Over the mid-1970's, the staff in the .s.nmposing room de-clined from 830 to 685 employees, and wouldhave dropped much further except for the abili-ty of the printer's union to maintain many jobsthat were in fact redundant.28 (Of course, if onelooks at media as a whole, electronics hascreated vast numbers of jobs.)

Unfortunately, while productivity is nowmuch higher, demand for books and .news-papers has not changed much. Between themid-1960'swhen only about 2 percent of alltypesetting in the United States was performed

"M. Wilkinson, ''System X: The Need to Shake-up the :hone-makers.' " Financial Times, Oct. 18, 1078.

"The union a2 the Times was more successful than most atholding on to jobs for its members. For a detailed treatment ofthis case, see "The Impacts of Robotics on the Workplace andWorkforce," Carnegie-Mellon University, Schou] of Urban andPublic Affairs, June 14, 1981, pp. 35fl, Othar examples of applica-tions of electronics technologies in printing' den be found in JR. Werner, "The Role of Electronics in The Modern News-.paper," and J. L. Boyd, R. E. Robey, and J. S. Richards, "Auto-mating Newspaper Production," sess. 21, The Role of Electronicsin the Graphic Arts, 1979 Electro Professional Program, New'York, Apr. 24.26, 1979.

C.


by the new machinesand the end of the1970's, penetration rose until about 90 percentof all newspapers were composed using com-puterized equipment..The impacts on printersas craft workers have been severe. Not only arefewer people needed for photocompoSition, butthey must have different skills. Few printershave found jobs as computer programers orservice personnel. Unions have been less con-cerned with the total number of job opportun-ities than with protecting individuals. Workforces have been reduced through attrition; thepension system created incentives for earlyretirement. Printers, proud of their traditionalcraft skills, were not very receptive to retrain-ing, although this had always been a centralpart of the uni,o's philosophy. While the strat-egies adopted h'y organized labor when con=fronted with such t c oblems have varied, theexample of the printing industry is not untyp-ical of instances where replacement technol-ogies have been introduced into existing mar-kets; labor-management relations tend to becritical factors in coping with job-displi,, rnteffects.

Electronic WatchesIn the watchmaking industry, an example

from consumer goods manufacturing, elec-tronically based products took more than halfthe total market within the space of a decade.In Switzerland alone, 20 to 30 percent of ex-isting assembly labor was displaced.27 Skill re-quirements for assembling electrOnic watchsare negligible. Along with deskilling of the pro-duction work force, international shifts oc-curred as firms in the Far East took over mar-kets for lower priced watches; most of the rel-atively simple integrated circuits neededare also made in Asia. Managements of Swisswatchmakers reluctant to switch to the newtechnology found their firms rapidly losingground, with effects on employment that wereeven more devastating than among manufac-turers choosing to embrace electronics.

,'Technical Change and Frnignyment. ,op. cit., p. 136.

Computer-Aided Manufacturing and Design

Continuing integration of computer technol-ogy into manufacturing operationscomputer-aided manufacturing (CAM)will eventuallyhave major consequencesfer erciployrnent (ch.6). Nonetheless, such developrneitsincludingrobots and software-programmable a 1..fiornatedequipment of many kindsshould generally beviewed as evolutionary steps in the .automationof the workplace, continuing down paths orig-inating many years ago. Much the same is trueof computer-aided engineering design (CAD),which consists in p-art of automating tasks-='ranging from drafting to numerical analysisformerly done manually. In addition, bothCAM and CAD make possible work that couldnot be performed at all in earlier years. Ex-amples include machining parts without theaid of drawings,-continuous balancing of rotorswith material removed by lasers, or finite-ele-ment analyses of stresses and deflections.

As computers spread through manufactur-ing, impacts on employment will be, at leastat first, incremental and random-si eming. Inthe longer run, productivity will be greatly im-proved; labor-intensity will drop, and largenumbers of manufacturing jobs will disappear,particularly those with lower skill levels. In thissense, the long-term effects will in fact be revo-lutionary. The work force will face continuingstructural shifts, and labor-management rela-tions will be under strain as accommodationsare sought:Changes in employment patternsin a given industry will depend on thecharacteristic production processeshowsusceptible they are to automationas well asgrowth in markets and shifts in competitive.ness. Computers will have their greatest irn:pacts when accompanied by large -scale reor-ganization of the work place,, as happened incontinuous process industries with the in-troduction of computerized process control.

Numerically Controlled Machine. Tools

The diffusion of numerically controlled (NC)(ch. 6) machine tools illustrates the results ofincremental improvement in manufacturing

3 7e)


technology. A survey of 24 American firmsrevealed comparatively limited impacts onemployment.28 NC machines were generallypurchased when business was good and out-put expanding; the new equipment helpedfirms produce more without hiring extraworkers. Nor were many employees displaced;most moved on to other production jobs,although skilled craftsmen sometimes foundthe transition to NC machines difficult. Man-agement also had to learn to operate in a newenvironment. Overall employment remainedmore-or-less static, but the skill mix changedand some individuals were faced with entire-ly new jobs. If the impacts of NC machine toolshave been mild, it would be misleading togeneralize this to future developments inCAD/CAM. NC machining is a major step inmetal cutting, but a much more modest devel-opment from the viewpoint of manufacturingtechnology as a whole; the next two or threedecades of advances in CAD/CAM will bringmore radical change to the factory floor.

Pet Foods

In an example of automated process control,a British firm with a large share of the pet foodmarket invested in a computer- controlled pro-duction system.29 Instituted with the goal of ra-tionalizing the production process, the systemwas expe,cted.to cut employment by three-quar-teis over a 5-year period. The proportion of un-skilled production workers dropped precip-itously, while more management and engineer-ing personnel were needed. An absence ofunions, combined with an extensive impaignto convince workers that the new c!uipmentwoulrl. eliminate the least desirable jobs, appearto h,: been critical factors in the acceptanceof the new equipment. The small group ofworkers selected by Management to run this'equipment expressed considerable satisfactionwith their greater responsibilities. The rest ofthe production work force lost their jobs.

2 R. T. Lund, et al., "Numerically Controlled Machine Toolsand Group Technology: A Study of U.S. Experiences:" KeportCPA 78-2, Center for Policy Alternatives, Massachusetts Instituteof Tethnologk, Jan. 13, 1978.

K . Dickson, "Petfoods by Computer: A Case Study of Automa-tion," ni The Microelectronics Revolution. T. Forester (ed.) (Cam-bridge, Mass.: MIT Press, 1981).

Are the Case Studies Typical?

None of these examples can be taken as rep-resentative. They are anecdotal accounts ofevents that have followed the introduction ofelectronics-related technologies. Technicalchange generally proceeds in piecemealfashion, with pace and impact that vary fromcase to case; given hindsight, of course, suchseemingly random and incremental events mayshow patterns invisible at the time.

In the examples recounted, the jobs createdgenerally called for different skills. Typical newopenings were for computer operators, or serv-ice and repair personnel trained to work on thelatest generation of equipment. While patternsof job loss and job creation vary across indus-tries, production jobsunskilled, semiskilled,or skilleddisappeared in all cases except NCmachining. Future employment impacts willbe influenced, not only by the technology itself,but by the general state of the economy at thetime new technologies are introduced, by theattitudes of workers and unions to automation,and by the choices of corporate managers. Insome cases, job losses will be mitigated by ex-panding markets, particularly if workers areretrained. Overall, however, a shrinking workforce in manufacturing points to continuingdisplacement and adjustment problems.

Services

The service sector has been growing morerapidly than manufacturing. Can the U.S.eamoniy continue to generate new jobs in serveices ai a high rate? Office work has been a ma-jor source of past expansion. With the elec-tronic office on the horizon, will this sourcedry up? If office automation begins to cut deep-ly into employment opportunities, the abilityof the service sector to compensate for lossesin manufacturing will be seriously impaired.

Office AutomationFortunately, this seems unlikelyat least in

the near term. Office work, breeding groundor Parkinson's Laws, will probably continue

to expand. At least some white-collar jobs seem


relatively impervious to automation in thesense that people can find other things to oc-cupy their time. This is partly a consequenceof the lack of c: :.-put indicators or other meas-ures of white - collar productivity. Nevertheless,beyond office wort:, electronics may v: du:.:e jobopportunities (or the rate of job creation) in sec-tors like !nsportation, retaling, banking, andthe postal system.

Less is known about the effects of automa-tion in services than in manufacturing. Overthe past two decades, the xerographic copierhas probably had greater impacts on officework than any other piece of technology, yetthese seem hardly to have been studied. Hasthe office copier created jobs? What have beenthe effects on organizational efficiency? No one

seems to know. It has saved so much drudgery,however, that few are likely to care.

More concretely, studies of the applicationof electronics to services generally findnotrevolutionary change, but gradual evolutionbest viewed as an extension of computer ap-plications already in place. Such studies em-phasize the extent to which workers such astypists or clerks whose job skills may becomeobSolete can be redeployed, seeing, for exam-ple, word processing as a straightforward ex-tension of typing.

The central features of the electronic officeexpanded applications of data processingequipment, including communications andword processinghave thus far been intro-duced into existing or conventional office en-vironments. In this respect, the analogy withNC machines and industrial robots is close.While office automation promises to reducestaffing needs in conventional jobs, new tasksare at the same time created in operating andmaintaining the systems, as well as using them.Since office work is seldom very efficient orwell-organized, computerization is likely tohave its first effects at the margins of these1,e ople-centered activities, rather than leadingto sudden and major shiftsWholesale reorga-nizations of the workplace will be slower thanin manufacturing.

Photo crsclit: Wang Laboratories

Word processing: one of the early stepsin office automation

Examining occupational categories makes itclear that the mode of utilization of the newtechnology is just as important as the speed ofadoption, again as in manufacturing. If newtechnology is instituted primarily as a substi-tute for narrowly defined functions such as in-ventory recording, bank telling, or filing,employment is likely to drop over the longerterm unless jobs expand in other areas (suchas sales). Where computers facilitate more ef-fective and extensive information processing,new jobs may be generated. Where demand fornew types of services is created, employmentwill rise.

Consider the proliferation of word process-ing equipment, which affects the tasks nowperformed by a well-defined group of employ-ees. Based on the results of work-measurement,tests conducted in organizations that haveswitched from typewriters to word processors,productivity often more than doubles. Whilethis might suggest that half the typing workforce faces unemployment, in practice nothinglike this has happened. Indeed, some firmshave invested in word processors in response

a shortage of typists. In other. instar..-es,. -here typists have been made redundant they.have moved to other parts of the organization.In many cases, people just write more words,

In fact, since word processors make it easierto produce multiple revisions of the same docu-


tricot, productivity cannot be measured simplyin terms of words or drafts typed. The charac-teristics of the technology lead directly to anincrease lie number of slightly modified ver-sions prk.., to a final copy, whether this is ashort letter or a report hundreds of pages inlength. While the benefits of this may be ques-tioned: the point is that oversimple estimatesof productivity gainsachieved or potFmtialoverstate the probable employment conse-quences.

Eventually, offices will be structured in sub-stantially different ways. Some jobs will beeliminated, others modified. Interactionsamong people, individually and in groups, willchange. Matters of timing and approach to theinf!allatiOn of new office equipment will, as inthe case of factory work, affect employee sup-port or resistance, thus the effectiveness withwhich the equipment is utilized, and people'ssatisfaction with their work.

Other ServicesService sector jobs outside the office include

health tare, retailing and selling of all types,banking, transportation, and postal services(more broadly, communications). In principle,electronics could alter many of these, butwhere and when -or whether is another mat-ter.

In banking, computer processing of mag-netically encoded checks has made it possiblefor the sanie number of employees to handlean ever-growing volume of transactions.tronic funds transfer remains costly, and thusfar has seen only limited use in retail banking;applications to interbank transactions havebeen much more prominent. Suchmenu have not led to work force redu0'.; as;during the 1970's, the number of people em-

ployed in banking in the United States grewby half, confounding predictions of employ-ment losses.° Two interpretations are possi-ble: the first is that growth in job opportunitiesslows under these circumstances; the second,that electronics allows banks to expand theirfunctions in ways that would otherwise be pre-cluded. These interpretations are not mutual-ly exclusive; both have some validity. Clearly,electronics technology has modified and ex-tended banking functionsan obvious exam-ple is the automated 24-hcur teller. Nonethe-less, in Europe, employment growth in bank-ing and insurance has already begun to slow;a well-known report to the French Governmentpredicts that one-third of all jobs in bankingad insurance might he eliminated over the

uecade ahead.31 While perhaps overly dra-matic, such predictions point to the concernthese issues have 'aroused, particularly inWestern Europe.

Like electronic funds transfer, electronicmail has been viewed with some apprehension.The U.S. Postal Service has made notablestrides in productivity over the past decade,even reduced its labor forcebut seldom as aresult of electronics. In the future, electronicmail may cut deeply into job opportunities forpostal workers; employment could drop by 20to 25 percent over the next 20 years.32

3°See j. Ilenize, -Evaluating the Employment Impact of in-formation Tachnology," Technological Forecasting and SocialChange, vol. 20, August 1981, p. 41.

,'Microelectronics at Work: Productivity and lobs in the WorldEconomy, cp. ci'., pp. 36-37. The French report is S. Nora andA. Minc, The Computerization of Society (Cambridge, Mass:MIT Press, 1980).

'2Implications of Electronic Mad and Message Systems for.U.S. Postal .Airvice (Washington, D.C.: U.S. Congress, Off ice ofTechnology Assessment, OTACIT-183, August 1982), ch. 6. Thupostal service employed nearly 700,000 people in 1980.

Suriimary and ConclusionsExamples from both manufacturing and

services can be interpreted either optimisticallyor pessimistically; in the absence of more sys-tematic studies, the question cf employment

impacts resulting from technical change inelectronics on the economy as a whole cannotbe answered. But regardless of the view onetakes, unprecedented adjustments lie ahead for

both individuals and firms. Should aggregateemployment increase. the introduction of newtechnology will alter the jobs that people doand change their interactions with one anotherShould total employment increase only slow-ly compared to growth in the labor pool, ordecrease, the adjUstment problems will be ex-traordinarily severe, more so in a country likethe United. States which has little experiencewith manpower policies, and where many peo-ple have come to view adjustment assistanceas a fallure.

In recent years, the number of new job op-portunities generated by the U.S. economy hasslo-ved. A good deal of the future expansionwill he in computer-related fields, ,nly thosewith appropriate training and skills be in

position to ke advantage of these oppor-;unities. Upskilling in the {:omputer industryhat; been going on for as indicated bythe increasing' proport!A.;.q of white-collaremployees, compared to production workers.In fact, the white collar-blue collar distinctionno longer carries much meaning; the laborforce is becoming increasingly stratified. Dis-tinguishing those with specialized skills fromthose without is only a starting place for ex-amining the many new gradations.

-A common notion, for example, is that com-puters will bring "user friendliness" to manyjobs so that unskilled workers can performthem. This is potentially misleading. User-

endliness permits people with good skills toork with complex and sophisticated systems

that otherwise would demand highly special-expertise. User-friendliness also tends to

:hangi. the abil required in the labor force.Efficient utilization of a word prOcessor de-pends en different skills than manual typing.Mistake .free entry is not So imp,,rtant, but tak-ing advantage of the full range of capabilitiesof the system requires a certain grasp of itslogiclind capabilitiesmental skills, not man-ual (and different from the spelling and gram-

,rrna now learned in school). Proanctivity inmany types of jobs increasingly depend onsuch abilities; it would he doubly unfortunate.if the i.S. electronics indu:;:.ry were to.suffershortages of trairied at the same time


that large numbers of Americans find them-selves without work liecause they lack thecapabilities that this and other industries de-pend on.

Like all technical change, advances in elec-tronics will bring a mix of positive and negativeeffects; at present, there is little factual basisfor either an optimistic or a pessimistic viewof the longer run impacts. Firms manufactur-ing:electronics products will, for some years,continue to create substantial numbers of newjobs. In U.S. manufacturing as a whole, the rateof growth of job opportunities has alreadyslowed, and jobs may gc down in absoluteterms. A major source of declines will be com-puter-assisted automation. Will job growth else-where compensate? Anticipating events in theservice sector, where productivity growth hasbeen low, is more problematic than in manu-facturing. While there may be only a few casesof employment impacts as severe as in news-.paper printing, there will be a multitude of ad-justment problems for individuals; these arelikely to accelerate as electronics technologycontinues to permeate both manufacturing andsere es. In the end, much t. ill depend on over-all rates of economic growth.

job opportunities also depend on comp.r.ti-tiveness. Employment, typically falls when in-dustries lose ground in either domestic or L.-ternational markets. Even if aggregate eco-nornie growth brings greater demand, only themore efficient companies can take full advan-tage. Generally speaking, firms and industriesthat make effective u-se of new technologieswill generate new jobs, or if jobs are lost, thiswill come more slowly; indeed, companies.seldom have any choice but to adopt new tech-nologies if they wish to remain competitive.Those that move quickly (but not too quickly)can often gain an edge over their competitorsvia new products or productivity improve-ments in existing lines of business, Ultimate-ly, the greatest numbers of jobs inay disappearwhere firms, industries, or nations do not keeppace with technological advance.

Firms or industries whose competitive pusi-tion is already in decline may he forced toautomate or purgl.fie other routes..to lower costs

373

374 = International Competitiveness in Electronics

and higher productivity simply to survive. Insome cases, then, declining employment isassociated with attempts to revive competitiveadvantage. particularly when an industry orfirm is threatened with competition from low-wage countries. But it would be a mistake toattribute the accompanying job losses solely toimports. In consumer electronics, U.S. cor-porations have automated their production fa-cilities and mover' oft,hot e: this costs v. 5. jobsin the short term, but may expi,. d or helpmaintain the total market for American prod-ucts over the longer term. Moreover, as theelectronics industry becomes more and. moreinternational-with American firms procL,-:inggoods overseas for foreign markets as well asre-importation, and foreign firms setting upassembly plants hereit becomes increasing-ly difficult to evaluate impacts on the Americanlabor force in isolation.

Most fundamentally, only by using labor effi-cientlywhich often means investments inautomationcan U.S. firms ma' 'ain thci_r in-ternational competitiveness. Improvements inproductivitya pathway to increased compet-itivenesscan have serious employment im-pacts on particular groups of workers, geg-graphical regions, and industrial sectors. 'I lie1-3sential question is: How can the negative im

.ts rnemployment be minimized while cap-italizing on the potentials of new technology?

Only where the market is expanding rapid-ly can employment growth parallel,productivi-ty advances. This has been tile case in thesemiconductor and computer industries, butnot in consumer electronics. To the extent thatthe American economy continues to grow onlyslowly, many of the productivity gains flowingfrom applications of electronics and computerswill have negative first-order effects on employ-ment. Still, few practicable alternatives exist;once robots or other automated technologiesbecome cost effective, the pressures to usethem become virtually irresistible. More jobscould ultimately be lost through failure to adoptsuch technologies than by pursuing them.

The implication is straightforward: somepeople, companies, industries, and regions willlose competitiveness and lose jobs. Tile rela-tionships between technical change, employ-ment, and international competition may becomplex, but from the standpoint of publicpolicy, the negatives are wholly predictable.They cannot be avoided, but the country couldprepare for them, both to ease the inevitableadjustments and to help maintain U.S. com-petitiveness. Because changes in industrialstructure bring new job requirements, policymeasures aimed at encouraging both publicand privately funded education and trainingare central to effective adjustment policy.

CHAPTER 10

ustrial Policies

Contents

377

383

4

37(,).., .

;mu]

, ..- ...... -120

CHAPTER 10

Na.:onal Industrial Policies

OverviewGovernment policies directly and indirectly

affect the international competitiveness of in-dustrial sectors. The impacts can be positiveor negative, tangible or intangible; they mayfall on domestic firms or foreign enterprises.The American electronics industry has claimedthat the policies of the U.S. Government some-times damage its competitiveness, while for-eign industrial policiesparticularly those ofJapanalso place it at a disadvantage. This isa familiar argument: many U.S. business lead-ers assert, on the one hand, that U.S. policiesare -counterproductive and That they would bebetter off without Government interference,and on the opposite hand, that in other coun-tries government polices, far from being coonterproductive, give their competitors powerfuladvantages in international trade. Such ques-tions turn on the general tenor of relationsamong government; business, and other in-terest groups (consumers, organized labor) aswell as the details of policy.

As the importance of electronics became Ob-vious anti competition intensified, foreign gov-ernments sought policies that would promotethe growth and development of their own in-dustries. These trends seem bound to continue,not only in industrialized nations like Japan butin developing economies. Questions of centralconcern for American policymakers include:How do, industrial policies differ among na-tions? To what extent can the effectiveness ofthese policies be evaluated? Do actions takenby forer_m governments give the electronicsin-dustries of these countries significant com-petitive advantages? Can industrial policies"create" comparative advantage? These arehard questions. The monetary value of :;ubsi-dies can seldom be approximated accurately.Even where this is possible, it does not tellwhether the money was well spent or wasted.More important, the industrial policies of coun-tries like japan work in large part through in-

99-11 1 O 53 - 25

tangibles. When counting the yen does not suf-fice, how does the United States countervailsubsidies?

This ch,pter treats industrial policy in com-parative fashion, with special attention to in-stitutional context and the evolution of in-dustrial policymaking, as well as the place ofelectronics in strategies for economic develop-ment. Policies in the United States are coveredonly briefly; the next chapter treats U.S. tradepolicies in greater detail, while chapter 12 ex-amines policy alternatives for this country.

Industrial policy means different things todifferent people. To some, the term brings tomind government programs for supporting andpromoting targeted industries, typified by theFrench "plans" or Japan's government-fundedresearch and development projectssector-specific attempts to assist industry. Beyond sec-toral measures, a vast array of public poli-ciesdealing with taxation. trade, human re-sources, science and tecisnology, antitrust,labor markets and economic adjustment, gov-ernment procurementalso influence the de-velopment and viability of industries like elec-tronics. OTA prefers to view industrial policybroadly, as encompassing both sectoral target-ing and the many policy measures with ag-gregate rather than sector-specific aims thatoften have less direct effect's on private firms)

International competitiveness, at root, de-pends on the efforts of private firmsthis isas true in countries like Japan with relativelycomprehensive and well-developed industrialpolicies as in the United Statesbut public.policies help shape the environment withinwhich corporations .)perate and managers

. _U.S. Industrial Competitiveness: A Comparison of Steel, AMP-

anti Electronics (Washington 1).C.: U.S. Congress. 01Ike of Technology As56,17,ment, 07.1-15C-135, July 11P31), p. 151.

377

378 Internarional CommThriveness in Electronics

inhke decisions. The decisions of envernmentofficials are important t...)o. Natioici have ap-proached industrial policy differently; govern-ment i:Ltervention is more common and viewedmore positively in France or Tapan than inWest Germany or the United Sometimespublic policies are clearly defined end con-sciously developed, sometimes they evolve frad hoc fashion-, the traditional pattern here.

Still, such gei.: :alities can mislead: in allmajor industrial nations, policies toward theelectronics industry have changed over time;in several, debates over new approaches areunderway. Furthermore, policies often differacross an industry. More than 20 years ago.Japan began a series of programs intended tofoster the 1.,;r,)N.vth of an indigenous computeriminstrybut the government did little by com-parison to directly promote consumer elec-tronics. In the United States, public policiestoward the automoli: inclush'y have cer' firedon regulations, while trade issues have beenstre,sed in the context of steel. U.S. agricul-tural policy has been much more highly de-veloped than policies toward manufacturing.

Why then have some nationsFrance, Japan.Taiwan, for examplesmoved toward well-defined and rather comprehensive policiesdirected at electronics, while countries like theUnited States have not'? There is no simple

answer, but historical and institutional factorsas wcii as stages of economic development andthe exigencies of day-to-day politics play a part.

Where government has for years promotedindustrial development -France rather thanBritainpublic sector involvement in the econ-emy is more widely accepted as legitimate. Insuch countries, policies directed at a single in-dustry such as electronics have usually re-flected overall economic objectives. Institu-tional mechanisms that facilitate coordinatedpolicymakingcentral banks or developmentbanks, respected planning councils, centraliza-tion of responsibility within one or a fewbureaucratic ministries enhance the ability ofgovernment officials to implement industrialpolicies. These features are lacking in theUnited States. During the greater part of thepostwar pet iodwhen American industriessuch :3 electronics and aircraft were clearleaders world competitionpublic policieshere were directed, not at economic develop-,ment, but at regulation.

As th's chapter demonstrates,. industrialpolicies will be a prominent feature of the in-ternational competitive environment for theforeseeable future. While other countries arebusy developing them, the United States is stillgroping for a response.

The Context for Industrial PolicyPublic policies directed at electronics should

be viewed in light of a nation's overall eco-nomic developmentstrategy. Table 77 gives insummary form a number of indicators of eco-nomic position and induct; )o!'hy for fivecountries. Electronics and other h_gh-technol-ogy inde tries grow more important as manu-facturing and services displace agriculture. Injapan, agriculture accounted for more than 20no. ht of the gross domestic product (GDP)in 1955, when the electronics industry was in-significant by intern standards; by theend of the 1970's. ihe agricultural sector had

receded to less than 5 percent of GDP. From1976 to 1980 alone, Lie share of Japan's exportsaccounted for by elect; tics went from 9 to 14perccmt. 2

Such shifts, the results of complex-economic---currents, form part of the policy context. Majorchanges have also been occurring within elec-tronics. Continuing the example of Japan, con-

,Trends in the Electronics Industry in 1980 (Tokyo: ElectronicIndustries A.Isoci,lion of Japan, 1981), p "Industrial Reviewof Japan-1981.- Japan Economic lourn,11, p. 33.

Table 77,Economic and industrial Policy Indicators

United Stales Japan West Germany United Kingdom Taiwan

1. seNices as a proportion of gross domestic product (GDP)....2. P&D.erenditures as a percentage of GDP

3 Government as a source of R&D funds

(rather than mii.tary) R&D as ,3 ,..'.,rcentage of total R&D .

5 Electronics R&D as a percentage of toia: R&L. (1975),

6A. Government R&D spending on electronics as a percentage of

total government R&D ;pending (1975)

6B industry R8,D. spending on electronics as a percentage of

toe! industry R&D spending (1975'

7. Percentage 1978 government P.D funds going to:

63% (1980

2.39% (1981)

47.2% (1982)

70,0% (19S1)

22°A

30%

21%

55% (1979)

1.97% (1979)

27,4% (1980)

97.4% (1977)

28.4% (1979)

32%

26n'o

49% (1979)

2.32% (1980)

49% (19791

92.4% (1979)

30%

31%

30'

53% (1980)

211% (1978)

55% (1979)

69.7% (1978)

26%

34%

21%

38% (1979)

0.65% (1981)

43%a (1980)

NA

NA

NA

NA

Economic development 22% 13% NADefense

12% 52% NA8. Organizatm of policymaking syFiem Fragmented Centralized Decentralized Fragmented Centralized9. Gov rnment.business relations Adversarial Cooperative Structured

representation of

business and

labor views

Semiadversarial Cooperative

10 Patents granted 119811 65,770 50,904 13;429 22,924 NABalance of trade in electronics with the U'tited States (1981\

(millions of dollars) S4,235' $5,8780 $1,592 ±$1,696 -$1,635'12, Overall policy and sirdtegy Ad hoc Leapilog:

ind.genous

technology

development

Adaptive:

stresses

technology

development

Adaptive:

stresses

commercial

appocations

Catchup

NA r.01

qxcludes expenditures '01 mil,jary

Un[ted States yvith all nations

CNeptIve 57 lento exports to tee United Stales excerio,i,yj imports from the Unwed Stales.

Ten rear Economic Development Plan for TaiNan, Republic, of China," TaYoan Council for Economic Planning and Development, March 1181

Technical Change and FJ:'onomic Policy (Paris Organization 'or Economic Cooperation and Development, 19801, p. 31

Denser San gyo no KeiuSdiN3 no Hoho to sono Elcyo ni Kansuru Crime Hokoku (SurveyReport, on Trends in the Internationall:.alion of the Electronics Industry and Their Influence, Part II ol East and

Southeast Asial (Tokyo: Nihon DenSht Hikai Kogyckal (Electronic Industries Association of Japan), March 1681, p 121.

K. Scnalt, Industrial Innovation in the United Kingdom, Canada and the United States (London Contemprinl, July 19811, p. 9

J Bionson and H. B. Malmigren,"Technology and Trade Policy' Issues and An Agenda for Action,"report prepared for Department of Labor and OYce of the US Trade Represenlafroe, October 1981, p 158

S,,;7',ey of R&D Activities in the Yea 1980, Pepublic of China," National Science Council, Republic of China, 1981

ante 1980 Ohl ashingr.'1, D.C.: National Science Board, National Science Foundation, 1981), pp. 214214Ee,-,stronics. Jan. 13, 1982,

C J Mcsbacher, "Will R&D Funds Be More Than SP Billion in '82," Industrial Research& Development, January 1982 p 106Nat,onal Ri!..e.!15 of Science and Technology Resources-1982 iWashmgi311. DC ' National Science FOundaoon. 19821, p 33.

Devc;.:irent Report 1931 (Nevi York, Oxford University Press, 19821, p 114.

DOLT), 'or Science and TechnoIly, Nation' Research Council (San Francisco W H Freeman,. 19132i, p 519.

Elovforec Marne' Data 1982 (Weleplon, DC.. Electronics Industries k,.ociation, 19821. 7

from LIS Patent CPice, Embassy of Japan, Science Divinn, Coovalion Counci American Affairs. Republic of China

3 3

380 Inte. at;cnal Competitiveness in Electronics

surer products declined from two-thir is of_at country's excort.. of electronics in 1971 to

about half by the end of the decade, while si mi-conductor and computer exports increased.Many of Japan's consumer electronics ship-ments to the United States have been displacedby prc'lucts from other Asian nations, partlya resut of rapid industriaLzation in countrieslike Hong Kong and Koree.3

These changes in the composition of japansexports reflect shifty in the international divi-sion of labordeveloping economies are nowproducing more consumer electronic goods,while advanced nations concentrate on high-er technology products. Industrial policies canbe viewed as responses to such structuralchanges: they may attempt to modify or resistthem, to smooth adaptation to change, to com-plement or even induce it. "Success" is mostlikely when policies work to accommodate orreinforce rather than impede changes in in-dustrial structureprovided the policies arebased on sound judgments concerning thestrengths and weaknesses of a country's in-dustries, both domestically and in the interna-tional marketplace. This is no easy task: still,policies toward electronics in othk countries,if not the United States, should be v' wedin these termsas components of natio._in eco-nomic strategies based at least in part on per-ceptions and projections of structural shifts inthe world economy.

International economic conditions now favorAmerican corporations less than in the earlierpostwar years; this is one reason for the grow-ing interest in industrial policy for the UnitedStates. This country, along with the rest of theindustrialized 'est, has experienced low ratesof econom:c g riwt1L rising inflation, and highunemployment over the past decade. Competi-tion has intensified among firms here andabroad, all seeking to maintain or enhance theirpositions in markets that may be growing only

Under the-se conditions, governmentshave turned to industrial policy as a way out

?White Paper en Infernal lona! Trade-1980 (Tokyn: MinistryInternationat *I'rede and Industry, September 1080), p. 32. As

discussed elsewhere. Orderly Mari- -Hug Agreements have, alsocontributed to this shift.

of persistent economic problems. Moreover,aggressive industrial policies in one countrybreed responses elsewhere. The turn toweedindustrial policies. particularey in nations lack-ing a tradition of government involvement ineconomic affairs, is partly a reaction to thesenew circumstances; in other countries, in-dustrial policy is nothing new, just a continuation of past practices under a different ;lame

Policy OrientationsAs part of a nation's overall development

strategy, industrial policies can be directed atcatching up, leapfrogging, or staying ahead inworldwide competition (table 77). Absence ofa clearly defined industrial policy may indicategeneral satisfaction with the situation, the casein the United States until recently; lack of awell-defined industrial policy could also reflecta belief that it is improper for government toconcern itself with such issuesa widespreadattitude here. In contrast, during the 1960's theFrench and JapaneSe began supporting and de-fending their computer industries against whatthey viewed as an American challenge.

In many countries and at many times, defen-sive industrial pclicies have been devised--

intended to preserve existing economic structui s, maintain employment, and protect be-leagured firms and industries.4 Often defendedas temporary (ch. 1.1), protective measures fre-quenny turn out to be persistent if not perma-nent.

Adaptive inrlus: jai policies seek to en-courage structural change by facilitating shiftsof resources to growing and productive indus-triesthose in the process of becoming morecompetitive. Tn contrast to the defensive ap-proach, adaptive industrial policies begin withthe assumption that some sectors will eventual-ly decline in size and importance. in practicethe boundaries 1;etv.reen various sorts of in-dustrial policties are vague; for instance, sub-;sidles or protectien for a given sector may berationalized as a means of encouraging adap-

See W. Diebold, Jr.. Indesirial Policy a5 an Intervalimal IFS1.10(New York: McG:pw-11 1880), op. 7-8. for an outline of typosof industrial policies,

Ch. 10National Industrial Policies 381

tation, while in practice they function asdefenses against decline.

More ambitious than adaptive policies arethose that attempt to induce change. This im-plies moving beyond a response to economicforceshere government takes the lead in ini-tiating industrial change, with the object of im-proving the competitiveness of some sectorsof the economy. Both this approach and theadaptive strategy tend to tre associated with no-tions of dynamic comparative advantage andthe belief that governments can anticipate andplan for shifts in the structure of advantage.

As pointed out in chapter 5, the competitive-ness of all sectors of an economy cannot im-prove at once. To pursue a positive develop-ment strategy, .a nation must begin with at leastthe implicit acknowledgment that some of itsindustries will likely decline. Common groundconcerning the prospects for industry is easierto find in economies with simple structures.Nations that are still attempting to catch uphave an easier time in formulating policy; theyface fewer choices, fewer possibilities.

The Tools of Industrial PolicyIn market economies, governments bring a

more or >,eT,!,-; standard set of policies to bear oninductr,ai development measures used forpurposes ranging from improving competitive-ness to encouraging regional development orstrengthening the national defense. Regardlessof whether a country is attempting to pursuean integrated policy, a wide variety of govern-ment actions will inevitably affect the in-dustrial portion of its economy.

In the case of electronics, many countrieshave instituted policies affecting costs and sup-plies of capitalfor R&D as well as for invest-ment in plant and equipment. R&D supportscan take the form of low interest loans, directsubsidies, or government contracts. In West-Germany, government funding supports basicresearch as well as projects aimed at commer-cialization carried out by the laboratories of theFraunhofer "Gesellschaft; the German Ministryof Science and Technology also subsidizes con-),

tract research undertaken by smaller enter-prises, along with cooperative R&D in in-dustrial research associations. The Very High-Speed Integrated Circuit program of the U.S.Department of Defense is aimed at integratedcircuits (ICs) for military applications, but willhave commercial spinoffs. The EconomicRecovery Tax Act of 1981 included a tax creditfor R&D spending, as well as accelerated de-preciation of equipment used in research. Ja-pan also offers tax credits to firms that increasetheir spending for R&D over past levels. Be-yond this, the Japanese Government directlysupports projects aimed at commercial micro-electronics and computer technologies.

Many countries assist regions, small busi-nesses, perhaps entire industries through in-vestment grants and subsidies. The UnitedKingdom's National Enterprise Board provided50 million pounds to capitalize the semicon-ductor firm Inmos. In the United States, theSmall Business Administration loans money atfavorable interest rates and with lengthy repay-ment periods. Regional development loanshave stimulated investment by American andJapanese semiconductor firms in Ireland andScotland. National banks, particularly in-dustrial development banks, have been impor-tant vehi-Aes in many countries for channel-ing funds to particular sectors.

Government procurement is widely used tosupport national firms. Military procurementhas been much more important in the UnitedStates, France, and Great Britain than in coun-tries like West Germany. The "Buy Japanese"policies of public corporations such as NipponTelegraph and Telephone (NTT) were for yearsan integral part of Japan's policies in elec-tronics. In 1980, NTTwhich purchases siz-able amounts of communications and elec-tronics productsagreed, after lengthy negotia-tions, to open some procurements to foreignbidders. American firms have made only lim-ited progress in selling to NTT, but the atten-tion given the case indicates that governmentprocurement is becoming more subject to in-ternational negotiation, perhaps less usable asa tool for the promotion of domestic industries

385


(nontariff barriers to trade, of which this is anexample, are discussed more extensively in thenext chapter).

Still another category of policy measure in-cludes those bearing on the regulation of in-dustrial structure. Nations can influence thestructure of their industries by encouraging ordiscouraging mergers, not to mention national-izing firms or industries as the Mitterrand gov-ernment in France has done. American com-petition policy has emphasized the regulatorysidei.e., antitrust enforcementwhile inFrance and the United Kingdom, governmentshave steered companies into mergers (e.g., thecomputer manufacturers CII-Honeywell Bullin France and ICL in Britain) intended to create"national champions." Encouraging mergers,often through financial incentivessometimesreferred to in Europe as structural policyhasbeen a common feature of policies toward elec-tronics in most developed nation.

Some countries use foreign investment con-trols-to restrict inward flows of capital, andthus preserve domestic markets for local firms.In years past, such regulations, as well as re-strictions on imports and technology fromabroad, played a central role in the industrialpolicies of Japan; several examples in elec-trouics were outlined in chapter 5.5

Finally, tariffs and other varieties of tradepolicy are an ever-present force in internationalcompetition. Countries erect tariff walls to pro-tect new Or old industries; the European Eco-nomic Community, for instance, maintains atariff of 17 percent on ICs to discourage im-ports and stimulate domestic production. TheUnited States negotiated import quotas on col-or televisions with Japan, Taiwan, and SouthKorea during the 1970's in an attempt to dealwith the problems of this industry (as discussedin ch. 11). Trading nations all maintain exportpromotion measures intended to help localfirms sell in the world market. In the UnitedStates; the Export Trading Company Act(Public Law 97-290) passed in the fall of 1982is one of the most recent examples; modifica-

'See also R. S. Ozaki. The Control of Imports and Foreign Cap-- ital in Japan (New York: Praeger. 1972).

tions to the Foreign. Corrupt Practices Actlike,ise intended to support U.S. firms inforeign markets passed the Senate but not theHouse of Representatives in 1981.

Policy measures of the types outlined abovehave been deployed by governments every-where in their attempts to influence thedevelopment of industry and improve competi-tiveness. Generally speaking, tariff barriers,controls on foreign investment, and competi-tion policies were the tools of firit choice dur-ing earlier postwar years; since the late 1960's,as trade liberalization gained momentum anddirect trade barriers were dismantled, R&Dpolicies and investment stimuli have come tothe fore. In the wake of intensified competitionin a wide range of industries, trade negotia-tionsboth bilateral and nv.4.11ateralhave in-creasingly centered on szt..,5.tkiies and indirectbarriers.

While the typical rmx_ hdustrial policymeasures has shifted over titn,- .the group ofpolicy tools from which they ft re chosen hasnot changed veryl--...tch. The policiesof various nations draw on isrie sa:'.4'8 basic in-gredientsR&D supports, invez::-,:ent grantsand subsidies, public sector pocuzement,merger policy, controls on foreign investment,tariffs and other trade policies. Naiions com-bine these depending on their assessments ofthe strengths and weaknesses of their own in-dustries and the objectives of their economicdevelopment programs.

The key to effective national policies has lain,not in the individual policy tools but in theircombinationin the extent to which the poli-cies chosen complement one another and worktoward a more or less consistent set of objec-tives. The timing of policy initiatives and thereceptivity of private firms to government pro-grams are also important, but the success- orfailure of industrial policies is determined toa large extent by the ability of policymakers todevelop and implement a consistent frame-work and approach, one appropriate to that na-tion's position in the international economy.

The remainder of the chapter reviews indus-trial policy in a number of countries, with par-ticular attention to electronics.

(11 10National Industrial Policies 383

Industrial P6liciets 'ObrrtparedThe failures of in po,licy are. rnuali

more evident than th- does one.Weigh the contribuii .1f ricy,errinnentto economic develer,;;it--eiitlrre.1 on a generci!-or a sectoral basis v yQ 11.:,:g.kw.i1ry has bee ain the -take-off- stave,... .forcesMg more or less in co.re-.:0-1 spir?2,:lindustrid,',-ization? This was the :: japanessteel, shipbuilding, indi5tries in earlier years. a *Labor fore::and rapidly expandin :nart:,at5.',4(ere aided bythe government's push, at is iK)1,1; the case inother nations that ha-se booxi,l. experiene.erapid economic groo:th-.

Develop;Ag Covro.!TfAls

The past rise inthe electronics inc v,-,v:i.c.is c:1 el 1:4:amber of new-ly industrializing {NICs), mostthem in Asia. Mar.:.; ,.'.:.a-:ie7ilationsTaiwan,South Korea, Brazilhave chAen paths of goir-.ernment-guided economic davelopn'iont, albeitwith many gradations in the (rdent of goVerA-ment involvement. With the exceptionChina, which has emphasized "self-suffi-ciency," the Asian nations have relied heavilyon imported technology while capitalizing oncheap labor. In countries like Singapore, HongKong, and Taiwan, economic development pol-icies have relied more heavily on encouragingdiversified exports of manufactured goods thanprotecting local industries against import com-petition. The typical attitude toward foreignelectronics firms has been pragmatic, withAmerican and Japanese involvements toleratedor encouraged because of benefits in technol-ogy transfer and infrastructural development.

In years past, the electronics industries inmost NICs centered pn relatively simple 'con-sumer productsradios and black-and-whiteTVs, pocket calculators, electronic watches,toys and games. Now, policy pronouncementsfrom these countries are calling for shiftstoward more sophisticated goods. In Taiwan,which has perhaps the most ambitious govern-ment programs, the stated aim is a more knowl-

,,dgr .ir- ,Insive'aldustrial structure, Much as inplanners, also reconsidering

their l'-.adil.i;:nal.approach, have become moreopen teehnology exchanges and businessventures in.../,':;) 7ing foreign firms. While the in-dustries i7 -cou'atries like Taiwan and SouthKorea 13..avz. 2:1mady become major producersof ni:xlcileq.17-,.? products like color TVs,simpler rnic. .Az.,:,-.tctronic devices, and computer2c.,:ripherals, i!s far from certain that such na-tiVOS can suf:,c:ed in advanced electronics tech-riotogies. Wi%.7.7power limitations are the most

Ire re c st-.-a FA.

1.zioitith Koref.

11/1 .g.0.1'1;.'0?.. Government has consistentlyvapiriiclustrialization; the public sec -

to has perhaps been more pervasiveany of the other NICs. Policy in-

Vtr-1:21r.tints hale ranged from money to gui-reliatz-1..- it '2,,adirect taxes, raw materials

;.:A11,5 to exporters, target figuresfor R&D. Korea's export fi-

n..:13e.j,ni:. have also been unusuallyto other NICs.6

For ma- the Korean economy- ex-panded at rate, with annual increasesin gross national product (GNP) averaging 10percent over the period from the early 1960'Sinto the mid-1970's. Labor-intensive manufac-tured goods provided the foundation for thisgrowth;-exports have become much more im-portant to South Korea's economy over the past-decade, growing from 12 percent of GNP to 35percent.' Electronics has been an export leader,the most rapidly growing sector. Korea's elec-tronics industry is still small compared toJapan's, but it accounts for more than 10 per-

- cent of Korean exports.

More recently, South Korea's economic mira-.cle has fallen on the same hard times that have

'"Korea's Eximbank Provides Incentives To Diversify ExportMix. Destination,- IMF Survey. Nov. 26, 1979. p. 366.

7P. Hasan and D. C. Rao, Korea: Policy Issues for Long-TermDevelopment (Baltimore: johns I lopk;.ns University Press, 1979),p. 20.

381

3134 International Competitiveness in Electionics

afflicted the rest of the world. The slump wassudden: whereas Korea's output of electronicproducts grew at the astounding rate of 40 per-cent per year during the 1970's, production ac-tually fell in 1980, although rebounding strong-ly in 1981.8 South Korea's Government con-siders continued-growth in electronics neces-sary for recovery, and the industry remains afocal pr -t of development strategy. Korea'sfourth economic plan (1977-81) concluded thatlong-term export viability would depend onstructural changes in manufacturing. The plancalled for rapid increases in exports of elec-tronic products." Korea's Government assumesthat other developing economies will providestiff competition in sectors like textiles and ap-parel, where Korean industry has in the pastbeen strong; thus, the country needs to con-tinue moving into durable manufactures for ex-port. The government also intends to deem-phasize petrochemicals and heavy industrieslike steelsectors that helped lead Korean eco-nomic growth in past years. The fifth and latestplan released by South Korea's Economic Plan-

. ning Board proposes dramatic cuts in in-vestments in these portions of the economy,with expenditures on electronics boostedsubstantially." Table 78 summarizes proiec-tions by the Korean Government; electronicsexports are expected to climb to $14.5 billionin 1991. The most rapid growth is projected inindustrial electro- , cs products, including com-puters and communications equipment, witha heavy emphasis on microelectronics. Theshare of total electronics output accounted forby consumer products is expected fo beginshrinking by the latter part of the decade, witha pronounced move away from the less sophis-ticated components that are currently a staple

°Denshi Sangyo no Kokusaika no Hoko to sono Eikyo ni Kan-suru Chose Hokoku (Survey Report on Trends in the Interna-tionalization of the Electron,cs Industry and Their Influence,Part II on East and Southeast Asia) (Tokyo: Nihon Denshi KikaiKogyokai (Electronic Industries Association of Japan). March1981). p. 103; A. Spaeth, "Korea's Electronics Industry MakingRapid Gains in Shift to High-Technology Products," Asian WallStreet Journal Weekly, Dec. 20, 1982, p. 1.

Denshi Sangyo no Kokusaika no Hoko to sono Eikyo ni Kan-suru Chose Hokoku, op. cit., p. 56.

"N. Thorpe, "South Korea's Economic Program Reduces Ex-pansion of Several Major Industries," Wall Street Journal, July24, 1981, p. 24.

Table 78.Korean Etectron:cs Production

Output (millions of dollars)1981 19a6a

Consumer $1,600 $5,800Industrialb 490 2,700Components 1,710 4,800

$3,800 $13,300

Total electronics exports $2,200 $7,000

aProjected.corricutels and le!ecommunications equipment.

SOUF.a Spaeth, "Korea's Electronics Industry Making Rapi;.: Gains Shit!to High.Technology Products:' Asian Wall Street Journal Week, f, Dec.20. 19132. p. 1. The projections came from South Korea's Ministry ofCommerce and Industry.

of the Korean industry. Such a reorientationwill entail shifts in R&D emphasis, with in-,creases in funding for both product and proc-ess technologies. To this end, the Ministry ofCommerce and Industry has begun channel-ing funds to Korean electronics firms for de-velopments in semiconductors and comput-ers."

To help focus research efforts, the Korean In-stitute of Electronics Technologyestablishedwith government support in Gumi, the coun-try's Silicon Valleyis to be built into a center-piece for research in electroi-21'cs. The institutehas been installing production lines for verylarge-scale ICs; the equipment will be used forcommercial production as well as engineeringdevelopment." While the staff of the $62million institute remains small, planners hopethat it will eventually house more than a thou-saud research workers."

In addition to R&D assistance, the South Ko-rean Government has provided investmentfunds to electronics firms and supported themthrough procurements. For instance, Gold StarSemiconductora joint Korean-U.S. venturewill receive a loan of more than $40 millionfrom both foreign and domestic sources, in-cluding the Korea Development Bank, to man-ufacture telephone switching equipment which

"One report states that $800 million has already been investedby the government"Fourth Five-Year Plan," Electronics 14'etik-ly. Apr. 25, 1979, p. 19.

12-Korea's Electronics Industry Making Rapid Gains in Shiftto High-Technology Products," op. cit. Eventually, the instituteexpects to sell the production facility to a private firm.

""South KoPea Seeks Electronics Rebound," New York Times,Mar. 24, 1981, p. D5.

38°

Ch. 10Natio:la, Industrial Policies 385

will be purchased by the Ministry of Commu-nications.t4 A second major Korean electronicsfirmSamsung, also partly U.S.-ownedis in-volved in the project as well. When the govern-ment decided to begin color TV broadcastingin 1980, Samsung won loans to aid in the pro-duction of color receivers. Foreign firms havealso benefited from investment incentives, al,though South Korea's electronics industry hasbeen less dependent on overseas capital thanmost others in Asia. Foreign-owned companiesare exempt from Korean income, property, andcorporate taxes during the first 5 years ofoperation.15

Government generosity has not preventedbottlenecks such as rising labor costs and skillshortages among the 180,000 employees ofAsia's third largest electronics industry. Therecent push toward indigenous technologicalcapability implies heavy R&D commitments,but most South Korean firms have only limitedhuman. and financial resources to devote tothese ends. Furthermore, other countries arelikely to be cautious in transferring electronicstechnology to Korea now that the country'scompetitiveness is apparent. Japanese firJishave refused repeated requests for licensescovering video cassette recorder (VCR) tech-nology. 18 Korean producers have already dem-

""Cold Star Semiconductor Raising Loan for Move Into Ad-vanced Electronics." Asian Wall Street journal Weekly, Apr. 13.1981. p. 8. The company is owned 44 percent by Western Elec-tric and 56 percent by the Korean Lucky Croup.

15C. Webb, "South Korea," Electronics We^kly. Apr. 25.1979.p. 19.

"M. Inaba, **Koreans Press Japan To Share Video CassetteProfits." Elgctrorric News, Nov. 30,1980. p. F. Nonetheless, sev-eral Korean firms already produce VCRs of their own design.

onstrated their ability to compete in the colorTV market, but if they cannot get foreign tech-nology in other areas their progress in elec-tronics will be slowed.

In view of these obstacles, does South Ko-rea's development strategy seem feasible?There is no question that 'Korean firms are wellplaced to expand their shipments of productslike color TVs, passive components, discretetransistors, and small-scale ICs to more ad-vanced countries. Korea is already the world'sbigge=:3 producer of black-and-white TVs, andKoet.:n firm:: have been among the leaders asAsian nations have taken over mLch of theworld's i'.,roduction of consumer electronicsproductstable 79. But developing the capabil-ity for designing and developing new productsbased or domestic technology and resourcesis a more ambitious and less certain undertak-ing thin manufacturing commodity-like prod-ucts using standardized, well-understood tech-niques.

Taiwan

The Taiwanes electronics industry runs aclose second in sales to Korea (ch. 4), andemploys more people. Both governments havefollowed the Japanese pattern in emphasizingelectronics. At the center of Taiwan's current10-year econorni,-. plan (1980-89) is the develop-ment of the machinery, electronics, and 'nfor-mation industriesfavored because of highvalue-added, modest demands for energy, andcomparatively high technology content. Tai-wan has the best trained corps of engineers andscientists in the Far East outside of japan, mak-

Table 79.Market Shares in Consumer Electronics for Japan and Other Asian Nations

Share of total world market, 1979Japan

Videocassette recordersColor TVsMonochrome TVsRadiosAudio tape recordersAuto radios and tape playersOther home audio equipment, stereos

All other Asian nations Total Asian share0 93.2%

3.8 31.549.6 65.971.8 77.052.8 91.018.7 67.312.1 52.2

SOURCE: Denshi Sangyo no Kokusalks no Hoko to sono Eikyo nl Kansuru Chosa Hoko .0 (Survey Report on Trends In the internathmallzation of the ElectronttAIndustryand Their Influence, Part II on Efailt and Southeast Asia) (Tokyo: Nihc7 Kikal Kogyokai (Electronic Industries Association of Japanj, Mn ch 1981), p. 2.

3

.386 4 Interrztiorral Competitiveness in Electronics

ing the more technology-intensive sectors nat-ural targets. The country's development plansencompass ICs, computers and peripherals,and high-end consumer products such asVCRs. The Taiwanese, like the South Koreans,are not satisfied with their image as manufac-turers and assemblers of components, pro-ducers of cheap TVs and consumer goods. Ac-cording to the current government plan, elec-tronics output is to double over the decade."

For some time, the Government of Taiwanhas been encouraging shifts from labor- toknowledge-intensive industries. One vehiclehas been the Electronics Research and ServiceOrganization (ERSO), which gets about 40 per-cent of funding from public s:_.urces. ERSO,established in 1974, is one of four divisions ofthe Industry Technology Research Institute(ITRI); projects have included computerized in-dustrial control systerris, Chinese languagecomputers, and semiconductor development.The organization also negotiated the technol-ogy transfer agreenient with RCA that helpedTaiwanese firms produce c-MOS '-Cs for thecountry's watch industry.18 ERSO is engagedin manufacturing as well as R&D, and hashelped introduce improvadquality control pro-cedures in Taiwan's electronics industry.

Wage increases have rendered Taiwan'slabor-intensive industries increasingly vulner-able to competition from other developingcOuntries,.an important motive for the govern-ment's stress on knowledge-intensive sectorsand another parallel with Korea. Policy pro-nouncements call for greater use of computer-based autorn lion tu increase productivity andexport competitiveness. ITRI leaders hope thatTaiwan will be able to independently developsmall computers and the associated softwarefcr both domestic and export markets. Govern-ment planners believe ti.at Taiwan will havethe best chance of success if, instead ofattempting to challenge IBM or the Japanese,

Economic Development Plan for Taiwan. Repub-iii Tai wen Council for Econdmic Planning aid DiVfA(.!;' ' March 19812, p. 39.

"K. 11. laiwan Pushes High Technology." ElectronMay 3. 1980, p. 1GO.

the country's efforts are concentrated onspecial-purpose machines compatible with theChinese language, along with minicomputers,peripherals, and software.19 Examples of theinitiatives being discussed include join'. ven-tures with Western firms in which government-sponsored training efforts would nrovideskil!.;,1 workers f.. r` ,.are development.20

Along with other 2-:Isian electronics indus-tries, Taiwan depends heavily on exports (table19, ch. 4; Taiwan exported 80 percent of itselectronics production in 1979, South Korea 70percent), with the bulk of these shipmentsgoing to U.S. markets. Taiwanese firms suchas Tatung and Sampo have already set up colorTV production facilities in the United States.With en economy that has Teen growing at anannual rate of about 8 percent, unemploymentat less than 2 percent, and a persistent tradesurplus with the United States, Taiwan's elec-tronics industry is well positioned for furtherexpansion. But Taiwan faces many of the sameproblems policymakers in Korea are grapplingwith. The country will need grever numbersof well-trained technicians and engineers,higher levels of spending on R&D, and contin-ued improvements in labor productivitythelatter of growing significance as wages rise.

As for South Korea, Taiwan may not havethe financial and human resources needed forrapid development in electronics based on in-digenous technology. And again, foreign pat-ent holders fearful of new competitors appearreluctant to negotiate agreements with Tai-wanese companies, particularly in more ad-vanced products such r mr leaderswithin the Japanese e haveurged "accornmodai, pug Asianeconomiesmeaning that Japan should con-centrate on leading-edge technologies whileimporting less sophisticated goods from else-where in Asia; but if Taiwan's government is

Ving. "Tai I is Counting on Its Computer ineustry toly 'he Finnorny,- Asian Wan Street lour-

'y,grile or Perish: ElectronicsM. message." i attar Winds, October 1980. p. 11.

"Noig,%is Softwiire Center ir 1.'",'Iwart.- Asianjour- Mar. 15, 1982, p. 4-

3;i ti

Ch. 117NatiotJat Industrial Policies 387

serious about its commitment to high technol-ogy. such an accommo6ation would probablynot be acceptable' 21

China

More strongly committed to self-sufficiencythan other industrializing economies, Chinesprogress in electronics and other industries hasbeen unevenin part because of longstandingconflicts between the development of scienceand technology and the quest for revolutionarysocial change. China's desire for :self-sufficien -cy has also created obstacles to efiiment massproduction; as a case in point, components are.still soldered into circuit boards by hand, whilein the West wave soldering has been employedfor more than 20 years. This is not to say thatthe country's electronics industry is unrelieved-ly primitive: the People's Republic of China(PRC) builds mainframe computers as well asICs roughly comparable to mid-1970's U.S.products. Nonetheless. until recently most ofthe computers were one-of-a-kind machines,'asking even transportable software.22

The picture has changed in the last half-dozen years as a new consensus on the impor-tance of science and technologyone of the"four modernizations" advocated at the FourthNational People's Congress in 1975emergedamong China's leaders.23 in the National: Planfor Development of Science and Technology,announced at a nationwide science conferencein 1978, eight technical areas were singled outfor special emphasis, among them computers.In calling fOr the development of China's capa-Wlity in a wide range of electronics technolo-e,es, including large -scale ICs, microcomput-ers, peripherals, software, and computer net-werks, the plan termed computer science andtechnology "a conspicuous hallmark of thelevel of modernization of a country."' Thereestablishment of the Science and Technology

''See the summary of the Electronic incinstry Associatton ofJapan's report en Asian electronics in A-; in Wit11 Street four-

Weekl;% gene 8. 1981.72K. Berney, "Computer Sales to China." China Business

Reviow, September-October 1980, p. 25."R. P. Sulitmeier, Science, Technology nod China's Drive for

Modernization (Stanford, Calif.: Hoover Institution Press, 1980).

Commission, a central agency for policymak-ing and implementation, is a further indicationf othe government's new direction.

Ten factorises in China now produce comput-ers,: ranging from microcomputers to machinessimilar to PDP-1.15 and IBM 360s.24 The State.Administration of Computer Industry tSACI)has programs underway to utilize the nation'sexisting computing capability more efficient-ly, and intends to move tc ward smaller ma-chines rather than relying on large main-frames. As one route to such objectives, SACIis establishing joint ventures with foreign con-cerns. In 1981, the China Technical ServicesCorp. and the Japanese firm NEC (NipponElectric Co.) signed an agreement for a com-puter center in Beijing. NEC will provide amedium-sized machine free of charge, and anannual 4 -month training course for 30 to 40Chinese software specialists. The Chinese willsupply other facilitiesi. along ro,rith the center'sstaff, including interpreters. A similar agree-ment has been signed with Sperry Univac,while negotiations have taken place with otherU.S. firms, including Wang Laboratories and

.Honeywel1.25 Both Japanese and Western firmshope to establish themselves, in the potential-ly lucrative PRC market.

AS such ventures indicate China is puttinga good deal of effort into training computerspecialists as a basis for more effective utiliza-tion of information processing technologies.Electronics and computer technicians willstudy-at an Information Processing and Train-ing Center, established in '1979 wit.from the United Natti. As. Among the plans 'forthe center, to be equipped with as Burroughsmainframe, as well as five Hewlett-Packard3000 series minicomputers, are developmentof a world patent index, collection of inforrna-tion on food supplies, a data base on powergenerrtirm and distrfi.Johoti :Tor the ElectricPower studies of urban traffic flows,and "Taae YoecoP( .iii modeling.26 Soch endeav-

"D. Burstein, "Chinese' Fomerit Another Revolution," Elec-tronics, Jan. 13, 1983, p. 115.

"K. Berney, "China's Computer Revolution," China BusinessReview, November-December 1981. p. 14.

""U,N. Aid for China's Computer Modernization.- China lies;.ness Review, September-October 1980. pp. 33.34.

331

385 Jn.ternational Cornpet;tive-7ess'in Electronics

ors imply that China--at least at presentmaybe more interested in applications than inbuilding production capacity, not only in com-puter ea uipment but in electronics more gen-erally-.

Other NicsRapid growth in the Asian electronics in-

dustry extends well beyond Taiwan and SouthKorea. Hong Kong's companies, which havebeen basically assembly-oriented suppliers ofproducts like watches and calculators, ac-counted for 13 percent of the colony's exportsin 1980.27 In. Singapore, which has also beena major assembly site. the government has in-troduced policies intended to encourage semi-conductor manufacturing as well as productticn of computer hardware and software; thegovernment, for example, owns 25 percent ofTata Elxsi, a joint venture involving U.S. andIndian interests formed to make mainframeprocessors.28 Government policy in Hong Konghas been less intrusive than in Singapore, butthe electronics industry there has also beenmoving toward high technology.

Clearly the Asian NICs are all, in one wayor another, attempting to learn from andemulate Japan's approach to industrial policy.In the earlier postwar yearS, Japanese com-panies imported technology, while governmentdecisions favored heavy industries; newly in-dustrializing nations in Asia have already aban-enned this approach for one more like Japan'scurrent industrial policy. The question iswhether South Korea, Taiwan, and other NICshave, at this juncture, the resources to supper'technological self-sufficiency. Rot even if ;heirprogress _in developing how technol-ogies proves slow, these countries will be in-creasingly competitive in world markets forless sophisticated electronics products, wellable to challenge manufacturers anywhere thatfail to maintain a technological edge.

l'"Says Electronics Could Lead as Hong Kong Export Earner."Electronic News, Oct. 26, 1981. p. FF.

15"See CPU. Software Mfg. Leading Singapore's Ftiture." Elec-tronic News. Dec. 7, 1981, p. (2

The discussion above does no more thansketch in a few of the outlines of industrialpolicy toward electronics in developing Asianeconomies (Japan is treated in some detailbelow). Outside Asia, governments in countrieslike Brazil and Mexico have also nurturedrapidly expanding electronics industries.Brazil, for instance, has used access to itsrapidly growing market as the carrot for ac-quiring U.S. minicomputer technology.29 In allthese countries, foreign investments by Ameri-can and/or Japanese firms have been one of thestarting points for indigenous development.Today, these nations are aggressively attempt-ing to strengthen their own capability andreduce their dependence on more advancedcountries. None of the policies employedtheestablishment of government- supported R&Dfacilities, tax breaks and financial subsidies forlocal firms, preferential procurement, govern-ment encouragement 0: or participation injoint ventures with foreign firmsare uniqueor even unusual. Such measures are part of thestandard list. Still, government planners inNICs have often pursued them more consist-ently and forcefullySouth Korea is especial-ly striking in this regardthan have developedeconomies. This is partly because the paths arewell marked for NICs in comparison to ad-vanced nations with complex industrial Simi'tures. The explicit focus on stre ,gt*,01mestie tecl' nie '.1 how- -0, shift inemphasis has led to increaseJ demands fortransfers of technology as a condition for salesor investment by foreign firms. Countries mak-ing such demands rIT alternatively, offeringincentives to ".tract .;hnology inflowsseelift 144 as a pro , equil, for building their owncapabilities. Soffit.; multinational electronicsfirms have accepted these conditionswhichat times have been a prerequisite for marketentry, a tactic that Japan employed in yearspastmore readily than others. The draftUNCTAD (United Nations Council on Tradeand Development) code on technology transfer

Baranson u nd H. B. Malmgren, "Technology and TradePolicy:. 1Aues anti a Agenda for Action," report prepared forDepartment of Labwr arat Office of the U.S. Trade Representa-tive. October 1981, ; 125-126.

3 9

Ch. 10Naiional Industrial Policies 339

illustrates the strong desire among developingnations everywhere for technology acquisitionon more favorable terms.

A problem that the developing Asian econ-omies all sharesL,me more so than othersis expanding their pools of engineers and tech-nicians. Countries like India, Taiwan, andKorea have labor forces containing substantialnumbers of engineers and scientists, many ofthem educated in the West. Nevertheless, whilesome of these nations have managed to mobi-lize their human resources more effectivelythan others, none of the NICs have enoughskilled people t move rapidly into high-tech-nology electronics production: They do haveone advantage: their engineers are not paidnearly as well as in the advanced countries.With salaries perhaps one-quarter those in theUnited States, the industrializing Asian econ-omies are striving to capitalize on lower R&Dcosts as they earlier did with unskilled labor."While it is unlikely that any of these countrieswill quickly bridge the commercial and tech-nological gaps separating them from Japan andthe West, and while their nrroarbes to in-dustrial policy differ in ' ,uvernment in-tervention and relianc: 1 market mecha-nisms, all seem committee: to some variety ofcoordinated industrial policy as a means o5supporting local electronics manufacturers inboth domestic and world mark

United States

The U.S. Government has not developed aconsistent, systematic set of policies directedat industrya task that, even if judged desir-able, would be much more difficult for theworld's most complex economy than one thatwas still industrializing. It has become a com-monplace to note that, while numerous publicpolicies exert direct or ]tidirect effects on firmsand industries, the American approach is adhut_ In this sense, then, U.S. industrial policyalso differs from that in 1;:ipari or many of theEuropean nations. While the Federal Govern-

3°A. Spaeth. "Asian 'NICs' Rely on Cheap Brairraower To PlanOutput of More Advanced Gr)ds." Wall Street Journal. Jan. 5,1983. p. 25.

ment has paid more attention to some indus-tries than others, this has most often been aresult of political pressures, as in the case oftextiles, or national security considerations.And not even in the Department of Defensecould one find anything like an "electronics in-dustry policy." Following World War H, U.S.foreign economic policy centered on an ambi-tious recovery program in Western Europe andJapanthe Marshall Plan. But despite this em-brace of economic planning for other parts ofthe world, domestic economic policies have re-volved around macroeconomics and regula-tion. The United States has avoided promotion,planning, and targetingthe common tools inother countries.

In electronics, microlevel involvements, leav-ing aside national defense, have generally hadregulatory thrustswitness the lengthy anti-trust prosecution of IBM. One reason the U.S.Government has been willing to endorse eco-nomic planning overseas but not at home liesin the unrivaled strength of American corpora-tions _:wring most of the postwar period. Inlight of the success of American firms such asBoeing, IBM, or General Electric in world mar-kets, the focus of policymakers here on free-market competition is quite understandable;for the Federal Government to consider poli-cies that would promote "national champi-ons"as the French didwhen these champi-ons already existed, would have seemed super-fluous if not counterproductive.

Public policies have, nonetheless, exertedconsiderable influence on the American elec-tronics industry. Military procurements stim-ulated developments in computers and semi-conductors. Since the 1960's, trade policy hasbeen a persistent concern in consumer elec-tronics. Taxation, regulations of many kinds(particularly in the f-Aecommunications sector),patents, protection of computer softwareallhave been debated in various contexts. But intotal, the Federal Government's policies havebeen a patchwork, often based on objectivesquite different from those motivating the in-dustrial policies of other countries. Antitrustenforcement stands out especially.

393


AntitrustWhere competition policies in other coun-

tries have been vehicles for mergers, joint ven-tures, and consolidation, notably in the com-puter industrythe rationale being to createcompanies big enough to compete effectivelyantitrust enforcement in the United States hasaimed it breaking up large enterprises.31 De-spite the common association of bigness withbadness. American law does not prohibit oli-gopoly (industries dominated by a small num-ber of firms), but limits predatory or exclu-sionary tactics. Therefore, antitrust violationstend to be difficult to prove, cases lengthy andexpensive.32

How has antitrust enforcement influencedthe international competitiveness of Americanelectronics firms? As has been the case so oftenwith U.S. industrial policy, the side effects mayhave had the greatest impactin this instance,uncertainty over the intentions of the Depart-ment of Justice and the Federzl Trade Commis-sion. Business and industry in the UnitedStates claim perhaps, with justification thatantitrust enforcement is ambiguous and threat-ening, that Government officials, knowing theline to be vague, try to keep companies farback. Instances in which enforcement inten-tions have been known to actually stopmergers, joint ventures, or acquisitions in elec-

"On U.S. antitrust law and enforcement. see U.S. IndustrialCompetitiveness: A ComParison of Steel. Electronics. and Auto-mobiles. op. cit., pp. 184-185. Also ch. 12 of the present report;J. W. McKie, "Government Intervention in the Economy of theUnited States." Government Intervention in the Developed Econ-omy. P. Maunder (ed.) (London: Groom Helm, 1979), p. 75; andM. Keller, "Regulation of Large Enterprise: The United StatesExperience in Comparative Perspective," Managerial Hierar-chies: Comparative Perspectives on the Rise of the Modern In-dustrial Enterprise. A. D. Chandler, Jr., and H. Daems (eds.)(Cambridge, Mass.: Harvard University Press. 1980), p. 161.

"The Department of Justice initiated its suit against IBM in1969, with the trial beginning in 1975. A decisionwhich wouldcertainly have been appealed regartass of the verdictwas stillwell in the future when the case was dropped by the Govern-ment in January 1982. At the same time, the Justice Departmentresolved a 7-year antitrust suit asking that AT&T divest itselfof Western Electric, the communications company's manufactur-ing arm. On the settlements, see ''Statement of William F. Bax-ter, Assistant Attmt.,,y General, Antitrust Division, on RecentActions of the Department of Justice in U.S. v. AT&T and U.S.v. IBM, Before the Committee on the Judiciary, U.S. Senate,"Jan. 1982.

tronics are few. One case aro-ie at the end ofthe 1970's. when GE and Hitachi proposed ajoint venture to manufacture TVs in the UnitedStates. The two companies suspended their ne-gotiations after the Justice Department threat-ened to sue under provisions of the ClaytonAct.33 More frequently, the possibilityeven ifremoteof costly and protracted litigationseems to have caused American firms to steerc!ear of cooperation in R&D.34 While much ofthe complaining by the business communityover antitrust reflects no more than the usualantagonism toward Government regulation, itdoes appear that companies have been little in-clined to explore the bounds of the permissible,simply because the risks have been seen as fargreater than the rewards. Largely as a resultof repeated expressions of concern, the Depart-ment of Justice issued a set of written guide-lines covering joint R&D ventures, but a gooddeal of ambiguity nevertheless persists.35 Evenwhere no single project has great import, ageneral discouragement of joint R&D effortscould eventually have a large cumulative im-pact. Moreover, if joint international researchprojects proliferate, American antitrust lawinthe absence of more concrete guidancemaypresent an obstacle to participation by U.S.firms.36

Trade and Foreign Economic PoliciesIf antitrust has recently been at the forefront

of U.S. industrial policies as they have affected

"J. Crudele and J. Hataye, "Fear for TV Jobs as Justice BlocksGE-Hitachi Venture," Electronic News, Dec. 4, 1978, sec. I. p. 1.

34s..-ae, for example, D. H. Ginsburg, "Antitrust, Uncertainty,and Technological Innovation." Antitrust Bulletin. winter 1979.p. 635.

"Antitrust Guide Concerning Research Joint Ventures (Wash-ington, D.C.: Department of Justice. November 1980). At the endof 1982, the Justice Department announced thai it woidd not seekto bar the formation of Microelectronics & Computer Technol-ogy Corp., the joint venture involving a dozen U.S. firms in-tended to pursue R&D in advanced electronics technologies (ch.5).

"For a proposal that foreign enterprises be allowed to partic-ipate equally in the government-sponsored R&D efforts of all na-tions, see Report of the U.S.-Japan Economic Relations Group,January 1981, p. 80. Japan has recently agreed to open itsfifth-generation computer project, and others like it,,to Japanesesubsidiaries of U.S. companies. See U. C. Lehner. "U.S.. JapanPact Would Bolster Joint Research," Wall Street Journal, Nov.1, 1982, p. 35.

39,1

Ch. 10National li,dustrial Policies .= 391

the computer industryvia the IBM casetrade policies concerned with dumping andother unfair practices have been central in con-sumer electronics. Trade policies and their ef-fects ale treated in detail in chapter 11; thepoint here is simply to note their significanceas part of U.S. industrial policy. After years oflitigation, the competitive battle in color TV isstill proceeding in the courtroom as well as themarketplace. A legalistic thrust analogous tothat in computers has dominated public policyimpacts.

To take a somewhat broader perspective, asworld competition in electronics has increased,U.S. policymakers have renewed their attemptslo reduce overseas trade barriers. Nontariff andindirect barriers restricting the entry ofAmerican products into foreign markets havebeen particular targets. A new flurry of activitycame in 1982; the many bills introduced inCongress that could be loosely grouped as deal-ing with trade reciprocity illustrate the depthof concern. Progress on such questions will beslow; since most countries view subsidies andother tools of industrial policy as internal mat-ters, they are difficult to address via interna-tienal negotiations.

Procurement and R&D

In contrast to the antitrust and trade orien-tations visible in computers and consumer elec-tronics, American semiconductor firms haveseldom, since the 1960's, been directly affectediy public policies. Through the 1950's and1960's, the Federal Government stimulated de-velopments 'n microelectronics by purchasingsemiconductors for military and space pro-grams, as well as by supporting R&D (much thesame was true for the computer industry in itsearly years). During this period, the Govern-ment purchased a large fraction of U.S. semi-conductor outpute.g., for the Minuteman IImissile. In 1965 the Department of Defense ac-counted for about 70 percent of U.S. IC sales,while by the end of the 1970's, the figure haddropped to around 7 percent.37

"An Assessment of the Impact of the Department of DefenseVery-Nigh-Speed Integrated Circuit Program, National MaterialsAdvisory Hoard Report NMAB-382 (Washington. D.C.: NationalResearch Council. January 1982), p. 6. Also see ch. 4.

Because the military market is now so smallcompared to commercial sales, specializedcontractors do much of the work on devicesfor defense systems. Largely in response to theslow rate of introduction of advanced micro-electronics technologies into military hard-ware, the Department of Defense initiated anR&D program directed at very high-speed in-tegrated circuits (VHSIC) beginning in fiscal1979. With an initial 6-year budget of more than$200 millionsince expanded substantiallythe VHSIC program is intended to speed thedevelopment of ICs that meet military needs.Involving all three services, VHSIC has beenstructured around bidding by firms and groupsof firms for contracts covering a variety of well-defined R&D tasks. Although the ICs them-selves will be tailored to military applications,research results in areas such as processingtechnology, computer-aided circuit design, andsystem architectures will find their way intothe commercial efforts of U.S. merchant firms.While most of the VHSIC contracts are closerto development than basic reseai--.;h, the De-fense Department has also initiated a programentitled Ultrasmall Electronics Research in-tended to support R&D that will pay off 10 or20 years in the future.38

Even with the increases stemming from theVHSIC program, Federal support of R&D insemiconductor-related technologies remains amuch smaller fraction of total U.S. semicon-ductor R&D than in tit- 1960's. While the com-parisons are less thal:. -..:taightforward becauseallocations of spending to R&D categories tendto be rather arbitrary, and disaggregated dataseldom available, an idea of the current sig-nificance of Federal funding can be piecedtogether.

For 1980, the latest year for which data isavailable, total U.S. R&D spending by the "elec-tronic components" sectorwhich is con-siderably larger than microelectronics alonehas been put at $1.354 billion.3° For the same

°The 5-Year Outlook on Science arid Technology 1981 (Wash-ington. D.C.; National Science Foundation NSF 81-10. 1981). p.33.

"Electronic Markel Data Book 1982 (Washington, D.C.: Elec-tronic Industries AssoCiation. 1982). p. 121. The figure. from datacollected by the National Science Foundation, is for SIC category367. which has nine subdivisions. Of these, semiconductors (SIC3674', is 4-ertainly the largest performer of R&D.

395

392 International Competttivenesc tn Electronics

year, tabulations of R&D spending by U.S. mer-chant semiconductor firms from sources suchas annual reports give totals in the range of$800 million. It is more difficult to determinespending on microelectronics by captives,which seldom- re-port such data separately.Allocation of software development costs alsoleads to ambiguity; as microelectronic devicesbecome more complex and more like completesystems, soft are becomes a major part of theresearch, design, and development effort.

In any case, given that IBMlargest of thecaptive producersno doubt spends severalhundred million dollars annually on microelec-tronics, total U.S. R&D expenditures on semi-conductor-related technologies in 1980 musthave been well over $1 billion. How much ofthis did the Federal Government provide? Forfiscal year 1980, Government expenditures forR&D related to ICs have been reported as $61million, rising to $71 million in fiscal 1981.4°

'°An Assessment of the Impact of the Department of Defensel-ery-High-Speed Integrated Circuit Program, op. cit., pp. 20-22.Fur purposes of this rough comparison. R&D related to ICs canbe taken as equivalent to R&D related to microelectronics. TheFederal contribution includes work performed in Governmentlaboratories. but this accounts for less than 10 percent of the

*Is

Phob croon:Pr:el Corp.

iy 16-bit microprocessor

Evidently, then, the Federal Government con-tributes something between 5 and 10 percentof the total. This estimate illustrates the con-tinuing decline in the Federal presence; overthe period 1958-76, Government spending ac-counted for about 15 percent of all U.S. semi-conductor R&D."

Indeed, it appears that even in the earlyyears, Government purchases were a greaterspur to the industry than R&D contracts.42 Byproviding a guaranteed market, Governmentprocurementmostly for military purposesstimulated the growth of the industry at acritical stage in its development. At the time,semiconductor manufacturing was a far differ-ent business than today; it was part of thedefense sector of the economy, whereas salesto the Government are now dwarfed by salesto computer manufacturers and other nonde-fense customers.

TaxationThe Economic Recovery Tax Act of 1981

(ERTA) was supposed to speed economicgrowth and build U.S. competitiveness by in-creasing incentives for saving and investment.

totalthe rest being contracts and grants to industry. universi-ties, and independent research laboratories. About 10 percentof the Government money comes from_the National ScienceFoundatioh (NSF), most of the rcst from the Department of De-fense. NSF's share of basic research support is closer to 30 per-cent. In fiscal 1980. the VHSIC program accounted for 40 per-cent of the Government's total spending on microelectronicsR&D. As table 77 indicates, overall R&D spending by the.U.S.Government is heavily skewed toward military needs comparedto countries like Japan or Germany.

IBM spends well over $1 billion annually on R&D: the com-pany's R&D spending on very large-scale ICs has been reportedto total about $1 billion over the period 1977-80G. Gregory."The U.S. Wages Micro -War," Far Eastern Economic Review,Mar. 16, 1979, p. 124.

"A Report on the U.S. Sem4...mductor Industiy (Washington.D.C.: Department of Commerce. September 19N9). p. 8. The es-timates are those of the Semiconductor Industry Association.In 1038. Department of Defense contracts and grants accountedfor nearly a quarter of the industry's R&D spendingN. J. Asherand L. D. Strom, "The Role of the Department of Defense in theDevelopment of Integrated Circuits," Institute for Defense Anal-yses paper P-1271, May 1977, p. 3. The percentage has thus beenfalling more or less steadily for many yea!!;.

"Asher-and Strom, op. cit.; J. M. Utterback and A. E. Mur-ray, "The Influence of Defense Procurement and Sponsorshipof Research and Development on the Development of the CivilianElectronics Industry," report CPA-77.5, Center for Policy Alter-natives: Massachusetts Institute of Technology, June 30, 1977.


The wholesale changes in U.S. tax policy em-bodied in ERTA have affected all industries;perhaps most significant are the altered depre-ciation schedules discussed in chapters 7 and12. ERTA also extends a tax creditamountingto 25 percent of any increase in R&D spendingover a base figureas part of a package of in-centives for research. Although young high-technology companies may not have profits toset against the tax credit, at least some elec-tronics companies will benefit from the R&Dprovisions more than from accelerated depre-ciation.

As pointed out elsewhere, the ERTA packagehas thus far had little perceptible effect on in-vestment. Moreover, it appears that, in com-parison with other U.S. industries, the relativeattractions of investments In electronics mayhave been diminished. The telling point in thecontext of U.S. industrial policymaking is this:such outcomes have been neither intended noranticipated, instead resulting from the unex-amined give-and-take ofthe political process.

Industrial Policymaking

As many of the examples above indicatefrom antitrust through taxation (many otherscould be adduced)public policies influencingthe American electronics industry have lackeda framework and sense of direction. The verynotion of objectives or "goals" for policy, inany but the most immediate sense, has beenanathema for policymakers here. In contrast,other countries have pursued economic devel-opment quite consistently, making use of nu-merous policy tools. While in the United Statesthere has beey no one agency to serve as a focalpoint for Industrial policies, other nations havedeveloped policymaking approaches involvingmore or less permanent industrial advisorycouncils, ministries accountable for well-defined policy areas, mechanisms for coordina-tion. Here, many agencies participate in policydevelopmentsometimes on a regular basis,sometimes infrequently.

More often than not, policies affecting elec-tronics have been formulated with little con-sideration of possible impacts on internationalcompetitiveness. National security, antitrust,

99-111 0 - 83 - 26.

macroeconomic policy have taken prioritycompetitiveness and economic efficiency haveSeldom been at'the forefront, or even in view.Trade policy complaints in consumer elec-tronics have come from domestic firms andtheir employees, with Federal agenciesill-equipped to take an independent viewreact-ing to these pressures. Short-term response topolitical pressures has in fact been the com-mon denominator of U.S. industrial policy.

Yet as competition has intensifiedin com-puters and microelectronics, jet aircraft andtelecommunications systemsboth Congressand the executive branch have begun to debatethe question of a more explicit industrial policyfor the United States.'" In addition, the Depart-ment of Defense through VHSIC and other pro-grams, the National Science Foundation, andthe Department of Commerce have all stud-iedeven attempted to design, often underrubrics such as innovationpolicies that wouldstimulate basic as well as applied research, and

4,To give only a few examples, and leaving aside such relatedtopics as innovation or productivity, late in 1980 the Subcom-mittee on Trade of the House Ways and Means Committee issuedthe United States-Japan Trade Report (Sept: 5, 1980), calling foroverall improvement of the economy rather than trade protec-tion as a response to Japan's growing challenge in high-tech-nology industries. A report by J. GresserHigh Technology andJapanese Industrial Policy: A Strategy for U.S. Policymakers (Oct.1, 1980)recommending a more focused U.S. response was pub-lished soon thereafter under the auspices of the same commit-tee. The Subcommittee on Trade's Report on Trade Mission tothe Far East (Dec. 21. 1981) reiterates many of the same themes.More recently, the Joint Economic Committee his released astudy by M. Borrus, J. Millstein, and J. Zysman entitled interna-tional Competition in Advanced Industrial Sectors: Trade andDevelopment in the Semiconductor Industry (Feb. 18, 1982)

,which stresses the importance of electronics for overall econom-ic development.

Dozens of hearings in Congress over the past several yearshave covered such issues, two examples being Industrial Poli-cy, hearing, Joint Economic Committee, Congress of the UnitedStates, May 18, 1982; and U.S. Industrial Strategy, hearing, Sub-committee on Economic Stabilization, Committee on Banking,Finance and Urban Affairs, House of Representatives, Sept. 22,1982.

Trade and tax policy debates inside the Reagan adMinistra-lion have dealt at least peripherally with electronics, as has aninteragency study on high-technology tradesee An Assessmentof U.S. Competitiveness in High-Technology Industries (Wash-ington, D.C.: Department of Commerce, February 1983). TheCommerce Department has begun work on an inventory of in-dustrial policy measures employed by other countries, while heDepartment of Labor has a long-standing interest in industrialpolicies, particularly as they deal with adjustment.

3 9


encourage productive investments by the pri-vate sector. There is no dearth of concern overpolicies affecting industries like electronics,but little consensus as yet on the direction that'policy initiatives should take."

Chapter 12 addresses policy alternatives forthe United States in some detail; here the pointis thatin contrast to ongoing debates over in-dustrial policy in other countries, which tendto focus on review and redirection of measuresalready in placethere is still no consensus inthis country on the need for a more coherentindustrial policy, much less on the form itmight take. In a sense, the United States is start-ing off behind in the race to develop effectiveindustrial policies simply because U.S. in-dustries like electronics led the competitiverace for so many years.

FrancePerhaps more than any other advanced West-

ern nation, France has centralized and coor-dinated its industrial policymaking as one ofthe primary ingredients in an interventionistapproach to economic policy. While the toolsand tactics have shifted over time, the policiesadopted by the current Socialist Governmenttrace their origins to the planning processadopted by France in the aftermath of WorldWar II. The continuity of the French system,like that of Japan, is one of its salient charac-teristics.

The SettingThe French have accepted government in-

volvement in the economy as legitimate andnecessary. Industrial policies are part of a con-text that includes extensive public ownershipof both manufacturing organizations and finan-cial institutions; under Mitterrand, the elec-tronics firms CII-Honeywell Bull, Thomson,

440TA's comparison of U.S. competitiveness in three industriesled to the suggestion of a "macroindustrial" policy. The intentwould be to provide infrastructuTal support for American indus-tries, rather than moving toward, explicit Government decisionsfavoring some sectors over othes. Examples would be policiesdirected at labor markets, technological development, humanresources. taxation, and economic adjustment. See U.S. Indus-trial Competitiveness: A Comparison of Steel, Electronics, andAUtomobiles, op. cit., pp. 157-165.

and CGE (Compagnie Generale d'Electricite)have joined the roster of national enterprises,together with a number of banks. The goal hasbeen not only to increase the financial re-sources and market power of French corpora-tions, but to create prestigious flagships thatcan' ead the economy. Saving jobs in threat-ened industriese.g., steelhas also been animportant motive; furthermore, the govern-ment's plan for the electronics industry prom-ises to create 80,000 new jobs over the 5-yearperiod 1982-86.45 "National champion" firmswere a capstone of French industrial policyduring the 1960's, when France became thefirst nation to mount a direct challenge to IBM.Aircraft, nuclear power, and telecommunica-tions have been other government .favorites.The idea of national champions never reallydied, and has simply been revived in slightlydifferent form under Mitterrand; electronicscomputers, semiconductors, consumer prod-ucts, communications, office automationis tobe at the core of France's ftture industrialpolicy.

Policymaking mechanisms in Francecen-tered on the ministries of Industry and ofEconomy and Financediffer greatly fromthose in the United States, as might be expectedin a country where the idea that the state canand should play a role in industrial develop-ment has been widely affirmed. In the policy-making system that has evolved, the Ministryof Economy and Finance takes the lead inchanneling funds to favored sectors (ch. 7),while the Ministry of Industry is more heavily.involved in day-to-day matters, as well as tech-nology and microlevel planning. Within theMinistry of Industry, the Directorate of theElectronics Industries and Data Processing isresponsible for efforts such as the GovernmentProgram. for Development of Electronics, an-nounced late in 1982. Since the Socialists tookpower in 1980, the Ministry for Research andTechnology has taken a larger role in industrialpolicynot only the, design of policies forhigh-technology industries like electronics, but

" "Government Funding for Electronics Industry Discussed."West Europe Report. Science and Technology, No. 118, JointPublications Research Service JPRS 81678, Aug. 31, 1982, p. 3.

396


also restructuring elsewhere in the Frencheconomy. Broad 5-year economic plans con-tinue to be part of the policymaking process,although they have receded into the back-ground compared to 30 years ago.

The staffs of French ministries tend to sharesimilar educational backgrounds, typically theprestigious grandes ecoles. Not only the publicsector, but large industrial enterprises as well,are managed by a small and homogeneouselite; the Socialists placed at the head of na-tionalized firms by the Mitterrand Governmentare in most respects indistinguishable from themen they replaced. The closed nature of thissystem helps the French bureaucracy wieldauthority more like that granted public officialsin Japan than in the United States. Distinguish-ing features of French industrial policyanemphasis on sectoral measures, perhapsstronger even than in Japan, and the encour-agement of corporate. consolidationreflectnot only the power of the bureaucracy but thecommunity of interest binding industry and thestate.

Planning

Much has been written on indicative plan-ning in France, which can be traced to the im-mediate postwar period and the Monnet Planfor rc,.onstruction. As the term "indicative"suggests, the country's 5-year economic planshave not been imperative, but based on con-certed actions mutually agreed on. In earlieryears especially, officials in the Planning Com-mission played key roles in bringing togetherleaders in government and industry.

A major function of the planning process hassimply been to gather information about pastprogress and future prospects by sectors in theeconomy. The hard decisions have been madeelsewhere, with the role of the Planning Com-mission largely facilitory. While the rapidpostwar recovery of the French economy can-not be attributed solely to planning and in-dustrial policy, the planning exercise hashelped crystallize perceptions among thebureaucracy, as well as decisionmakers inprivate industry, creating a shared referent for

government and industry. Business has beenable to operate within a fairly predictablecontext.

Finance

The French financial system, like the plan-ning mechanism, enhances government influ-ence over economic development. Capital al-locationssee chapter 7are controlled to con-siderable extent by administrative fiat ratherthan market forces. A rather small number offinancial institutionsclosely tied to thebureaucracy whether or not actually national-izedlink government policymakers and com-panies seeking funds. The Treasury determinesinterest rates on bonds; through the Ministryof Economy and Finance, as'well as a varietyof semipublic lenders and the banks, thegovernment can exert considerable leverageover credit decisions. Specialized institutionssuch as the Institut de Developpement In-dustriel (IDI), funded from both public andprivate sources, provide risk capital to medium-.sized firms; IDI has also made equity invest-ments in the computer firm Compagnie Inter-.nationale pour l'Informatique (CII). Even inlight of the French Government's traditionaluse of finaritial channels, Mitterrand's invest-ment plans for electronics are extraordinarilyambitious. The industrywhich is now rough-ly half nationalizedis to invest $20 billionover the period 1982-86, with the governmentproviding about 40 percent of the total." It isnot clear where the money will come from.

Le Plan Calcul

French policies as they have affected elec-tronics have been shaped, as elsewhere inEurope, by historical circumstancei.e., therelative weakness of French industry com-pared to American corporations. The resultduring the 1960's was a concerted thrust incomputers known as Le Plan Calculnotunlike what the French are now undertaking_in electronics as a whole.

"Ibid. The 5-year investment plan calls for the governmentto provide 55 billion francs of the 140 billion total.

393


By the early 1960's, the enormous strengthof IBMcombined with the comparative weak-ness of European firmswas perceived as aserious threat to the viability of the French elec-tronics industry. It was the "American chal-lenge"viewed as a technological lead, but inreality just as much commercial superioritythat stimulated an ambitious effort by theFrench.'" The well-known Plan Calcul came in1966, on the heels of serious difficulties for theFrench computer industry. In the "AffaireBull" of 19Ril, the American firm General Elec-tric had purchased Machines Bull, a falteringFrench computer manufacturer. At about thesame time, the U.S. Department of State re-fused to grant export licenses for two of Con-trol Data's largest processors. These were tohave been used in the development of fusionweapons; the refusal helped convince Frenchpolicymakers of the, need for an independentcomputer industry. Since then, if not before,the French military, although taking some careto stay. in the background, has had a major sayin industrial policy .decisions affecting elec-tronics and telecommunications.

Le Plan Calcul was intended to build an in-dustry capable of challenging IBM; to do this,

, bureaucracy engineered the merger of twoexisting manufacturers, forming a new publiccorporation CII. The government providedcapital to the fledging champion, but as theproduct of a union between two firms whichtogether held no more than 7 percent of theFrench computer market, the new companyhad a long way to go.

CII's efforts were directed first and foremostat medium to large mainframesthe rise of theminicomputer was just beginning and had notbeen widely recognized. CII was to be an ex-port leader, as well as providing for France'sown needs, of which national security was atthe forefront. Although CII was attacking IBMat the latter's point of greatest strength, French

__ planners hopedby providing export and_othersubsidies, encouraging shipments to the Sovietbloc and developing countries, protecting CII

"This interpretation of the "challenge" is elaborated by N.Jequier, "Computers," Big Business and the State. R. Vernon(ed.) (Cambridge, Mass.: Harvard University Press, 1974), p. 193.

against foreign competitors, and guaranteeingdomestic procurements for the company'sproductsto enable the firm to challenge IBM.A related series of measures over the late 196e'sand early 1970's comprising Le Plan des Com-posants (Plan for Components) was to helpwith the development of semiconductor de-vices, primarily for computer applications.

Despite the support provided CII, the firmnever approached its targets. American com-panies continued to dominate sales in France,and CII's chief market turned out to be thegovernment. By 1975 Le Plan Calcul had effec-tively been abandoned, as a variety of factorscombined to defeat the best efforts of Frenchpolicymakers (who now insist that their effortsat least prevented further erosion of the na-tion's indigenous capabilities). Some critics em-phasize the contradictions inherent in a pro-tective strategy within a highly competitive in-dustry, and a policy designed and implementedby technocrats with little experience of com-mercial realities." Hindsight shows the effortto have been overambitious, an attempt to con-front American firms across a broad line ofproducts rather than in selected niches. In thissense, national goals took precedence oversound business strategy. Finally, the moneythat the French Government pumped into CIL,-perhaps $350 million between 1966 and 1976looks rather insignificant next to; say, IBM'sresources."

Into the 1970's, then, French policy towardthe electronics industry centered on one com-panyCII. By 1975, when the failure of Le Plan'Calcul was clear, the government encouragedCII to merge with Honeywell Bullthe de-scendant of Machines Bull that emerged fromthe sale of General Electric's computer busi-ness to Honeywell. The new company, CII-Honeywell Bull (CII-HB), was majority French-owned; it quickly received further government

44J Zysman, Political Strategies for Industrial Order (BerkoleIV--Calif.: University of California Press, 1977), p. 99.

4°The $350 million estimate is from Technical Change and In-dustrial Policy: The Electronic Industry (Paris: Organization forEconomic Cooperation and Development, 1980), p. 46. Jequier(op. cit., p. 217) gives a figure of $120 million between 1966 and1970.

,Ch. 10National Industrial Policies 397

assistance totaling perhaps $700 million.5° In1982, after prolonged discussions, the Mitter-rand government effectively nationalized CI l-

R, which will become a subsidiary of a gov-ernment-controlleafriolding company taking upthe old name Machiiies Bull." Machines Bullwill be the centerpiece of Mitterrand's com-puter thrust, discussed below, with CII-HBresponsible for mainframes.

Le Plan des Composants

Recognition that CII -HB, needed infusions ofsemiconductor technology led to a new 5-yearcomponents plan in 1977. Military require-ments were also a strong motive. Since the1950's, France had maintained a small buthigh-quality semiconductor research effort.However, this had never been translated intoa commercially viable merchant industry. Bythe early to mid-1970's, perhaps a hundred.French engineers and scientists were engagedin R&D on advanced microelectronic devices;the country did not have the capability formass-producing ICs. French engineers had lit-tle background in microprocessors, nor inMOS (metal oxide semiconductor) ICs, most oftheir expertise being in bipolar devices for com-munications and consumer products. Le Plandes Composantsdesigned by the ministriesof Industry and Defense, plus the PTT (respon-sible for postal services and telecommunica-tions)was intended to rectify these deficien-cies. In contrast to Le Plan Calcul, France didnot attempt to keep foreign participants out,but sought to build on American technology.In this way, French planners hoped to movetoward self-sufficiency in microelectronics,with the eventual goal of a major share of theEuropean market.

The vehiCles included three joint ventureslinking American semiconductor firms with

"Technical Change and Industrial Policy: The Electronic In-dustry, op, cit., p. 46. Other estimates have ranged as high as$1 billion. Prior to' the merger with Honeywell Bull, C11 had beena participant with Philips and Siemens in the European con-sortium. Unidata. The consortium did not prove workable.

51:C11-Honeywell Bull Announces Restructuring in Line WithFrench Plans for Computer Firms." Wall Street Journal, Dec.21, 1982. p. 30. Honeywell's interest has been reduced to about20 percent.

Frenth partners. France would get technology,the U.S. participants access to the FrenCh mar-ketparticularly the lucrative telecommunica-tions sector, well protected by the PTT, Thesejoint ventures, in which the French partnersheld controlling interests, tied Thomson toMotorOla, Saint-Gobain to National Semicon-ductor (in a firm named Eurotechnique), andMatra to Harris. In addition, the plan supported two more firms: Radiotechnique, aPhilips subsidiary in France, and EFCIS, orig-inally owned by the French atomic energy au-thority (Thomson purchased a majority interest,in EFCIS in 1977).

Le Plan des Composants was developed ata time when France's Government was redis-covering market forces. Attempting to learnfrom Le Plan Calcul, French planners decided,to support a number of firms. Rather than fun-nel the money set aside for the program to asingle champion, the five companies wouldcompete with one other. Although an elementof competition was thus built in, each partici-pant was assigned certain technologies inwhich it was to take the lead. Matra-Harris, forexample, would specialize in c-MOS since thiswas Harris' strength; later the joint venturenegotiated a further agreement with Intel,largely to gain the latter's n-MOS technology.ICs were new technologies for both Matra andSaint- GobaLi, which were picked for the pro-gram in part because of their success in otherfieldsin the case of Matra, its high-technol-ogy experience in aerospace was a particularattraction. Saint-Gobain-Pont-a-Mousson, a ma-jor producer of glass and chemicals, had de-cided of its own accord to diversify into elec-tronics; in addition to participating in Le Plandes Composants, the company purchased sub-stantial interests in CII-HB and the Italian com-puter and office equipment firm Olivetti dur-ing this period. More recently, the French havedecided that, if one national champion is toofew, five are too many; since the Mitterrandgovernment came to power, extensive-discus-sions aimed at consolidation have been under-way., Three centers of excellence in microelec-tronics seem likely to emerge. Both Matra andSaint-Gobain have been nationalized, withSaint-Gobain evidently forced out of elec-

398 International CompetitIven&: mics

tronics.52 Eurotechnique has been sold toThomson, which purchased the shares of bothSaint-Gobain and National Semiconductor,while retaining a technology exchange agree-ment with the American firm.

The R&D portion of !.9 Plan des Composants,known as Le Plan Circuits Integres, channeledaboUt $150 million in government funds as di-rect 'grants and loans to the five companies.The nioney supported work on very large-scaleICs, ranging from circuit design to the develop-ment of 'processing equipment. Research cen-ters, inClUding the Electronics and InformaticsTeclincloi. Laboratory of the French AtomicEnergy Col nission, were strengthened, whileLe Plan Circuits Integres also supports micro-processor ap\ilications through a new Informa-tion Agency.53

It is 107-) earl, to judge the success of Le Plandos C . .-Josants in building a viable commer-cial industry, b6t in terms of technology Frenchsemiconductor\ firms have mane great prog-ress. Eurotechnique manufactured its first ICsat the end of 1980 and has since expanded out-put at a high rate. EFCIS's production of ad-vanced devices 1?egan about the same time. De-spite rapid incrOses in production,_ however,the French entrants remain small on a worldscale (see ch. 4, table 32), suffering from thinproduct lines aind limited distribution net-works. Still, the technical know-how they haveacquired from Atnerican firms places them inElthiantageous positions compared to otherEuropean semiconductor manufacturers.

In recent years, the French bureaucracy hasalso given a gobd deal of attention to minicom-puters and peripherals through Le Plan Peri-nformatique. Moreover, the components pro-

vam has been linked to a major push into tele-communicationsincluding developments

52See "Possible Strategies fir Executing MicroelectronicsPlan," West Europe Report, Science and Technology. No. 112.Joint Publications Research Service JPRS 81340, July 22, 1982,

Absorbs-Eurotechnique-FinancialTimes, Jan. 21, 1983. p. 14. .

""190 Million Francs in Next Five Years for VLSI Research,"West Europe Report, Science and Technology, No. 89, Joint Pub-lications Research Service JPRS 80022, Feb. 3, 1982, p. 7; "LeDeveloppement des Applications de L'informatique," Lettre 101,Oct. 7, 1980.

such as videotext7that French planners em-barked on in the .rnid-1970's; the PTT's am-bitious projections 'envision 25 million ter-minals in French homes by 1990, pointingtoward a rapidly growing market fOr semicon-ductors. As part of its telecommunicationspolicy, the government has forced the sale ooftwo foreign-owned companies (subsidiaries ofITT and Ericsson) to the Thomson group.

Recent Developments

French industrial policy has been in some-thing of a turmoil since Mitterrand's election.The outlines of the Socialist Government's pro-gram remain murky, although the intent is toemphasize electronics. Initiatives in semicon-ductors, computers, communications, and con-sumer products are likely to be even more tight-ly coordinated than in the past. And, while thethemes of nationalization and merger policypredate Mitterrand, the Socialists have carriedthis aspect of French industrial policy stillfurther.

Even beforeMitterrand came to power, theEleventh Five-year Plan (1981-85) had targetedelectronics for special support. The plan sin-gled out six fields for massive governmentassistance, with electronicsranking third inFrench exports of manufactures, after machinetools and chemicalsviewed as a criticalsector." Under the plan, total R&D expendi-tures in France are scheduled to increase to 2.5percent of GNP by 1985. Currently, France ismaking a more concerted effort than any, otherEuropean country to strengthen its technolog-ical base and promote high-technology in-dustries, with considerable attention to train-ing greater* numbers of engineers and tech-nicians. Le Plan des Composants indicated thatthe French had learned from the mistakes ofLe Plan Calculand also from the commercialfailings of the Concordewith French indus-trial policy as it affects electronics and otherhigh-technology sectors passing into a newstage ;nne-marked by-a-more suphisticated-un

"Rapport de la Commission Industrie, Commissariat Generaldu Plan, July 1980. p. 48. According to this report, electronicshas received about 10 percent of all direct sector-specific aidto French industry in recent years (p. 113).

el 0

Ch. 10National Industrial Policies. 399

derstanding of international competition incommercial Products and technical develop-ments. This shift began during the 1978-80period. tinder Prime Minister Barre, the gov-ernment 'claimed to be "decontrolling" mar-kets, fo: instance cutting back on price con-trols. With a deemphasis on planning and na-tional champions, came more stressat leastin official rhetoricon market forces. Is Mit-terrand likely to reverse this trend?

A fundamental plank in the Socialist plat-form had been nationalization of companieslike CII-HB. As Mitterrand himself explains theSocialist strategy, the object is first to win backthe domestic market in key industries such assteel, machine tools, semiconductors, andsmall computers.55 In conjunction with furthernationalization in the financial sector, the an-nounced philosophy was "flexible" national-izationwith the government providing con-.siderable support while promising to eschewextensive involvement in business affairs oreconomic planning at the micro-level. In ap-pearance, this is not a sharp turn from the past;despite, the lengthy history of planning inFrance, nationalized firmsso long as theyhave performed adequatelyhave operated rel-atively free of direct intervention by the bu-reaucracy.

In R&D and technology development the newgovernment has also moved ahead in bold ifseemingly disorganU.od fashion. Research sup-port has been increased under the current5-year plan, and is to include a new microelec-tronics project as a fellow -on to Le Plan Cir-cuits Integres, plus more money for computersand data processing. The government expectsto put up two-thirds of the $500 million itbelieves must be invested in microelectronicsover the 1982-86 period.56 As in the past, muchof the money will come from the Ministry ofDefense. And as also in the past, the new mi-crioelectronics plan is but one piece of a much

""Mitterrand: Why Nationalization Will Work.- Wall StreetJournal. Oct. 7. 1981. p. 27.

54"Microelectronics Plan: Win Market. Technology Independ-ence." West Europe Report, Science and Technology, No. 113,Joint Publications Research Service If' "S 81392. July 29. 1982.p. 10.

larger effort aimed at strengthening the entireFrench electronics industry.

While the overall outlines remain vague, ttiegovernment is promising that investments inelectronicsfrom both public and privatesources, and including investments by foreign-owned firms (IBM, Texas Instruments, andMotorola are among the American electronicscompanies with a major presence in France)will total $20 billion over the 5-year period1982-86. The Government Program for Devel-opment of Electronicspresented in Sep-tember 1982 after an extensive study by anElectronics Industry Task Force.is .to be coor-dinated by an Interininisterial Committee forElectronics, with representatives from the min-istries of Industry and Defense, the Plan, andthe PTT.

A primary vehicle will be 9 "national proj-ects," chosen from 14 originally recommendedby the Task Force. These national projects,which will get extensive government support,are intended to link private and nationalizedfirms, as well as the labor and user Comm,.ties. The nine projects have the folloutitles:57

consumer electronics;information displays;local networks;cable TV net,very large-sc,lie ICs (fabrication as well asdesign);central processing units for small comput-ers;computer-assisted education;computer- assisted engineering; andcomputer-assisted translation.

The list is noteworthy for emphasizing com-puter systems from the perspective of userneedsnot only the last three projects, but alsothat on local networks. All are Software-inten-

"See R. T. Galkgher. "$20 Billion for French Electronics."Electronics. Sept. 8. 1982. p. 104: "Fourteen Projects." WestEurope Report. Science and Technology. No. 116. Joint Publica-.tions Research Service JPRS 81575. Aug. 18. 1982. p. 14. Amongthe five that were droppednot necessarily permanently-7-wasa supercomputer effort.

403

400 international Corripetitivene,3S in Electronics

sive, a field in which France ;s I n a relativelygood position. The plan does not neglect con-sumer products. France hopes to increase itspresence in consumer electro.hus marketsthroughout Europe, with Thomson moving ag-gressively into new generations of products likeVCRs, electronic toys and games, and home in-formation systems."

The Socialist Government faces severe obsta-cles in implementing such a vast program. Inaddi ion to the $20 billion in planned invest-ments, a considerable increment compared torecent expenditures within the industry,France has an inadequate supply of men andwomen with training and skills in .electron;cs;the shortfall is reckoned at more than a thou-Sand engineers yearly and at least three timesas many technicians. The development plan.contemplates an extensive training effort, in-cluding the establishment of several newschools. Moreover, foreign firms with invest-ments in France may resist some elements ofthe program. Joint venture participants, for in-stance, could prove less willing to transfer tech-nology when the partner is a nationalized con-cern. But in the end, money will probably bethe limiting factor; boosting France's R&D rxpenditures from 2.percent of GDP hi 1982 to2.5 percent by 1985 is extraordinarily ambi-tious. And, with nearly three-quarters of R&Dcarried out in government-controlled inFtitu-,tions, France runs a real risk of stifling innova-tion and new ideas.

Future Prospects

While the hallmarks of French industrial pol-icy remain the.samean elite corps of officials,centralized policymaking, and a preference forsectoral policy along with a tradition of stateintervention in the affairs of industryMitter-rand's philosophy does represent 1 turn awayfrom the market orientation of Gi:;card d'Es-taing. It is too soon to assess the effectivenessof the new avenue, but past results give someinsigh-t§.-Gwvernment-elfo-rts under Le-P dlt

""First Details Published on Electroni::s Plan." West EuropeReport. Science and Technology. No. 12a Joint Publications Re-search Service JPRS 81904, Sept. 20, 1982, p. 7.

Calcul must be termed a failure, although CII-HB's troubles had multiple sources. In semi-conductors, Le Plan des Composants seems tohave functioned much better. Even so, thelargest French producerThomsoncontrolsonly a quarter of the domestic market, with amarket share in all of Europe that is perhaps7 percent. Most of Thomson's sales are ;ndiscrete semiconductors; the company has nomore than about 2 percent of the European ICmarket. Although Thomson appears to havebenefited considerably from technology-assistance agreements with Motorola, as haveMatra arid Eurotechnique through their jointventure with American partners, French elec-tronics firmsalong with most European man-ufacturersremain heavily dependent onforeign sources of MOS and microprocessortechnology.

The history of French electronics policyshows that strong government direction can-not by itself produce a competitive industry.At the same time, the French seem to be learnL.ing how to make their electronics policy func-tion more effectively.

United Kingdom

In contrast to the French, with their relianceon centralization and government action, Brit-ain's industrial policy has been closer to thatof the United Stateslargely ad hoc, not wellcoordinated. There is at least one major dif-ference: the United Kingdom during the 1970'sbegan to experiment with a variety of novelmeasures intended to directly affect the actionsof industry. Ranging from programs to encour-age applications of microprocessors to govern-!tient investment in the semiconductor ventureInmos, these initiatives are far different fromthe arms-length approach to industrial policyof the United States (U.S. policies related to na-tional defense are, as usual, the exception). Atthe same time, these policies some of whicha tt racted- considerable-attention-i n-other-partsof the worldwere pursued withlittle senseof direction. Only in its sup, Jrt of Interna-tional Computers Ltd. (ICL) through procure-ment practices, R&D funding, and other con-


vadtional policy tools has Britain shown muchconsistency over the longer term in policiestoward elearonics.

Measures to aid ICI. in several respects shn-ilar to the somewhat earlier French effort tobuild and strengthen CII-HB, date from the late1960's. In the latter part of the 1970's, theUnited Kingdom's electronics policies, as inmany countries, turned .owards semiconduc-tors; a group of programs -Ire developed topromote IC technolugy and pplications. Evenso, neither today nor at any point over the pastdecade does the British example show muchevidence of a coherent view of industrial pol-icy..

Early ExperimentsCertainly Britain has had ample incentive to

try new approaches; since the early 1960's, pol-icymakers in the United Kingdom have soughtways of grappling with the nation's lacklustereconomic performance by most measures thepoorest among industrialized nations. Duringthe 1950's, macroeconomic policies had beenassumed sufficient for, revival. Rut continuinginflation, along with persistent wage disputes,convinced the ruling conservatives to move to-ward a more active government role. The Na-tional Economic Development Council (NEDC)was established in the early 1960's as a forumwhere business, labor, and government couldair their ideas about the future direction of theeconomy.5 Inspired by the prestigious FrenchCommissariat du Plan, NEDC was empoweredto produce 5-year plans intended to reduceuncertainty about the directions of governmenteconomic policy. Planning responsibility fellmainly.on a National Economic DevelopmentOffice attached to the NEDC.

For an outline of the origin and role of the NEDC. see T. Smith."The United Kingdom." Big Business and the Stare. op. cit., pp.

np tho_narlier yprirq of Rritain's indlts-trial policy is drawn from this source. Also see S. Blank. "Brit-ain: The Politics of Foreign Economic Policy, the Domestic Econ-omy. and the Problem of Pluralistic Stagnation," Between Powerand Plenty: Foreign Economic Policies of Advanced IndustrialStates. P. J. Katzenstein (ed.) (Madison. Wis.: University of Wis-consin Press. 1978), pp. 114ff.

The Labor Government which came to pow-er later in the decade continued this generalorientation, and picked up the pace by estab-lishing a number of Economic DevelopmentCommittees to deal with specific industries.The Electronic Development Committee, set upin 1964, produced a series of reports that iden-tified problefns and proposed strategies forovercoming them: But by the end of the dec-ade, the planning experiment had run afoul ofpersistent conflicts with the macroeconomicpolicies that Britain's leaders were determinedto pursue; economic, planning came to beviewed as a failure, and the visibility and in-fluence of the Economic Development Com-mittees waned.°0

The More Reconfi ContextSince the beginning of the 1970's, U.K, in

dustrial policy has been a Bodge-podge. As inill( 'United States, consistency has been foundmoutly in the area of national defense. A hostof government offices, themselves subject toperiodic reorganization and changes in direc-tion, have been involved in policies affectingBritain's electronics industry. The NationalResearch Development Council, set up as earlyas 1.946 to provide financial support for jointresearch ventures under the Ministry of Tech-nology, is one example. The Ministry of Tech-nology also had jurisdiction over the IndustrialReorganization Corp. (IRC), establi..,ned in 1966to aid industrial restructuring. Under the authority of the Industrial Expansion Act, theMinistry of Technology engineered the merg-ers creating the computer firm ICL, as dis-cussed in more detail below. The IRC likewiseprovided financial backing and other encour-agements for a series of mergers that enlargedGEC, the British General Electric Co. But theinterventionist IRC was abolished in 1971,about the time the Ministry of Technology be-came part of the larger Department of Tradeand Industrywhich has since again been di-

"They still exist. however. The committee for electronics re-cently issued a report urging a comprehensive spctoral policyfor the industry. See "Prescription for Electronics." FinancialTimes. Apr. 30. 1982, p. 16.

4 05


vided, leaving a Department of Industry anda Department of Trade,

Among other agencies active in industrialpolicymaking, the National Enterprise Board(NE13) has had considerable leverage becauseof its ability to provide direct financing toBritish firms. Esablished in 1975, NEB hasconcentrated on startups such as the semicon-ductor manufacturer Inmos, to which it gaveabout $90 miilion in equity capital (ch. 7). Inquite different .realms, the Science ResearchCouncil and the Advisory Council for AppliedResearch and Development, set up in 1976, areintended to supply policy_ guidancr nr, suchtopics as applications of new lechnowgies andthe education and triii)ing of engineers.

The number of government bodies involvedin Britain's industrial policy provides one ex - --

planation for the random approach to pro -grains in electronics. In France, relatively clearlines of authority link the various parts of thebureaucracy dealing with electronics; certainagencies have the lead role in certain areas. Bycomparison, the British approach is uncoordi-nated. In further contrast to the situation inFrance, Britain has never been very comfort-able with government intervention in the af-fairs of businessrather surprising consider-ing the size of the public sector. Not only dogovernment plus the nationalized firms employabout a quarter of the British labor force, butpublicly owned enterprises account for morethan 10 percent of the country's industrial out-put and in recent years about a quarter of totalcapital investment.'" Nationalized firms in in-dustries like steel and automobiles have re-ceived more attention from British policymak-ers than electronics.. Still, the United King-dom's approach to the electronics industrydoes reflect a belief that government canstrengthen existing firms as well as create newones.

"'P. Maunder. "Government Intervention in the Economy ofthe United Kingdom." Government Intervention in the Devel-oped Economy, op. cit., pp. 131-137.

ICL

The formation of ICL was preceded by muchless rhetoric concerning the need to create na-tional champions than in France, but theemergence of ICL in 1968 was similar to thatof CII-HB. ICL benefited not only from gov-ernment fin,:Tscino, but from aid for R&D andthe promise of public sector purchases of itsproducts. Britain's Government encouragedthe series of mergers by which the companywas formed, and supplied about $12 million ayear until 1976 to stimulate its growth,"Despite this, CL never emerged as a viablecoml)etitnr in le world computer industry. Al-though still holding more than a' third of theU.K, marketlargely the result of governmentprocurements coupled with "Buy British" per-suasion aimed at private firms--ICL has hadlittle success outside the United Kingdom.Within Britain, the Central Computer Agency,responsible for government purchases, gaveperhaps 90 percent of its orders to ICL duringthe early 1970's." This is a major reason whythe United Kingdom joins Japan as one of onlytwo countries where American computer man-ufacturers and their subsidiaries do not haveat least half the installed base,

ICL is known for its software, butlike mostcomputer manufacturers outside the UnitedStatesmissed the shift. toward small systems.The company has also been handicapped bythe lack of a strong local semiconductor in-dustry. Since the latest government initiativea package of loan guarantees totaling nearlyhalf a billion dollars, and the installation of anew management team headed by a longtimeexecutive of Texas Instruments' U.K. subsid:siarythere have been signs of revival.'" -

Not long before jumping back in to try to saveICLin part because mergers or takeovers in-volving American companies were rumored,

62"Technology and Trade Policy: Issues and an Agenda forAction," op. cit.. p. 58.

63C (IP rarmoy, "Subsidy Pniiriec in Flritain, Franrpiinri WectGermany: An Overview," International Trade and IndustrialPolicies. S. J. Warnecke (ed.) (New York: Holmes & Meier, 1978),p. 38.

"E. Bailey, "Britain's Role at Ailing ICL," New York Times.May 18, 1981, p. Dt; S. Love, "New Talent Spurs Britain's ICL,"Wall Street Journal, Mar. 1, 1982, p. 27.

o

Ch. 10National industrial Policies 403

the Thatcher government had sold off the pub-licly held 25 percent of the' cornpany's stock:ICL's checkered past thus illustrates the "stop-go" quality of industrial policy in the UnitedKingdom. The uneasy relationship betweenInmos and the government has followed a sim-ilar pattern, one in which Inmos faced con-siderable uncertainty over whether the Thatch-er administration would provide the second in-stallment of capitalanother 25 millionpounds :hat the company was counting on.

Research and DevelopmentBeyond support of ICL, the U.K. computer

industry benefited during the late 1960's andearly 1970's from R&D funding providedthrough the Advanced Computer TechnologyProject, which provided up to half the costs ofprojects dealing with hardware or software.The British have also attempted to aid theirelectronics industry through efforts like apreproduction order program, in which thegovernment purchases newly developed prod-ucts ,and leases them to usersor "clients."After a trial period, the client, prepares a reporton the new product and 'hen must either buyor return it. Other programs have supportedsoftware development and marketing, as wellas microelectronics. In addition to contractedbasic research, paid for by both civilian anddefense agencies, commercial product develop-ment has been financed through government

cts, particularly to ICL.Nonetheless, the United Kingdom has been

a poor performer in R&Dmore precisely, indevelopment. Although British scientists con-tinue to do excellent basic research, industryhas been reluctant to invest heavily in R&D di-rected at commercial products and processes;between 1967 and 1975, real R&D expendituresby industry declined. Furthermore, sectors likeelectronics have suffered from a lack of capableengineers. While government R&D expendi-tures have been heavily concentrated on elec-tronics and aerospacein 197573D percentwent to electronics and communications alonethis spending, largely motivated by militaryneeds (table 77, ch. 10), has had little percept-ible effect on the competitiveness of British

firms. Tellingly, electronics and communica-tions manufacturers have spent less of theirown money on R&D than the government hascontributed; in 1975, private firms spent 113million pounds on R&D, publicly held corporations 36 million pounds, and the government130 million." In the United States, the impactsof military spending on electronics have beenfar overshadowed by the vigor of the commer-cial industry; the British case has been vastlydifferent.

Other Policies Toward ElectronicsAmong the more intriguing programs of the

U.K. Government have been those aimed at uti-lization of microelectronics. In the midst of alengthy debate on the q destion of whether thecountry needed an indigenous capability to de-sign and manufacture advanced ICs, the Mi-croprocessor Applications Project (MAP) wasestablished to encourage companies in any in-dustry that could to incorporate these devicesin their products. MAP, which began in 1978and has been somewhat reluctantly continuedby the Thatcher government, funds up to 25percent of the costs of product development.Increased support is provided for microelec-tronics-related programs in schools and col-leges, principally teacher training. A third ele-ment consisted of a consciousness-raising cam-paign aimed at 50,000 managers in private in-dustry, with MAP funds supporting seminarsto educate corporate decisionmakers on the vir-tues of the new terthnology. Governmentspending through MAP totaled nearly $100million over a 3-year period."

A related program known as MISPthe Mi-croelectronic Industry Support Programme,also started in 1978aids firms in developingand manufacturing ICs. MISP was stimulatedby a report prepared for the NEDC whichstressed the importance of design and process-ing expertise; a central goal was mass produc-

"K. Schott, Industrial Innovation in the United Kingdom,Canada and the United States (London: Contemprint, July 1981).p. 12.

"'Microelectronics, The New Technology." Department ofIndustry, London. 1981. p. 23.

40?


tion'capability in standard devices."' Amongthe steps t 'iken wend the establishment of a jointventure between 0,13 British firm GEC and Fair-child, then still American-owned, to manufac-ture a broad-based line of ICs. More ambitiouswas the decision to establish a greenfield firm,Inmos.

NEB's announcement in June 1978 that In-mos would receive equity funding from thegovernment to design and manufacture n-chan-nel MOS circuits, starting with memory chips,generated a good deal of controversy withinGreat Britain (NEB has considerable independ-ence in making such decisions). The attemptto replicate a merchant semiconductor firm onthe American modelcomplete wth execu-tives experienced in U.S. companies and a design center in Colorado Springswas ,C. movedirectly into the central arena of worldwidecompetition. The risks were high. Ininos wasto begin production of 64K RAMs in 1981atarget which slipped, but as it turned out, nomore than those of a number of well-knownAmerican firms.

Inmos is a unique experiment; the govern-ment committed 50 million pounds, split intotwo installments, with the hope not only ofcreating a first -rank semiconductor company,but also of luring talented British engineersback from employment with foreign firms (onereason for the Colorado Springs location). Stim-ulating end-users in Britain was another majorobjective. While there are still many doubters,Inmos appears to have had reasonable successin developing its first products. The companyhas plans for a new family of microprocessors,as well as a broad line of memory chips. Be-coming profitable may be more difficult.

Since the election of the conservativeThatch-er government in 1979, efforts such as MAPand MISP have been scaled back. The conserv-.atives' review of the Inmos venture revivedpublic debate over the company's prospects.After-a Lonsidelableperiod-ofttmei La hity , dui-ing which it appeared that NEB's holdings in

"'See Microelectronics Into the 1980's (Luton. England:Mackintosh Publications Limited. 1979), p. 27, for a summaryof the Sector Working Party report to NEDC.

the firm might be sold, and following a deci-sion by Inmon executives to locate a produc-tion facility in the depressed area of SouthWales, funding was continued. In late 1981,NEB reported $22 million in pretax losses,more than half accruing from its holdings inInmos; losses are to be expected during theearly years of such an enterprise, and it is stilltoo early to judge the success of this recent en-try into the world semiconductor industry, butthe qualms of the conservative government arenot surprising.

Has Britain's Approach Worked?

The answer, implicit in much of the discus-sion above, is that it has not. While some of theinitiatives in electronics may eventually havepositive results, U.K. industrial policy as awhole has suffered from lack of consistencyeven during periods when the same party hasbeen in powerand from a rather odd, if notchaotic, mixture of policy instruments. Eventhe direct beneficiaries, British electronicscompanies, have not been very enthusiasticabout the government's support efforts, view-ing them as favors likely to be withdrawn onshort notice." Some executives in British in-dustry could be described as not only skepticalbut cynical about their government's policies.Nothing like the symbiotic relationship be-tween business and government in Japan, oreven France, has emerged in the United King-dom.

The grab-bag character of U.K. policies to-ward electronics has stemmed in part from theinconclusive nature of debate over the need fora continuing British presence in semiconduc-tor manufacturing. Many took the positionthat, so long as British industry applied ICsin its products, there was no need to havebme-grown design and production capabili-

ty. Others held that, lacking an IC design andproductionobase, applications would always lagthose in other countries. Rather than coming

"See D. Imberg and J. Northcott. Industrial Policy and Invest-ment Decisions (London: Policy Studies Insitute. 1981). pp. 72-73.Also J. Northcott and P. Rogers. Microelectronics In Industry:What's Happening in Britain. No. 603 (London: Policy StudiesInstitute. March 1982). especially ch. 8.


to a decision; the British have tried to have itboth wayssupporting Inmos, though neverwhole-heartedly, while also pursuing applica-tions and technology diffusion through MAPand MISP. Similar patterns, over a longer timeperiod, have characterized the government'sdealings with ICL; after many years of publicsuppor',r, the Thatcher government withdrew,only to find itself forced to the rescue of thefaltering computer manufacturer. Industrieslike steel and automobiles show similar oscilla-tions in government attitude.

The fact is that foreign firms have alreadycaptured major shares of most British elec-tronics markets, except where the government,itself is the customere.g., the defense sector.Outside the government market, ICL presentslittle challenge to its American and Japanesecompetitors, just as British firms now hold onlya small share of U.K. semiconductor sales.Thus, ICL's agreement with Fujitsu, entailingmarketing of Japanese-built mainframes inEngland, also involves purchases of Fujitsusemiconductors. While ICL has also negotiatedfor rights to U.S. and Canadian technology,Britain does not seem as well-placed as Franceto make use of foreign know-how, and may findthat it is already too late for technologicalindependence.

In sum, many of Britain's industrial policyefforts in electronics seem to have been too lit-tle and too late. The formation of Inmos andthe creation of MAP and MISP came at theclose of the 1970's, by which time Americanand Japanese suppliers were firmly establishedin the U.K. market.

Industrial policy has been doubly difficult be-cause of the stagnant British economy. As eco-nomic troubles continued, the government cutback its R&D support, making progress in in-dustries like electronics still less likely. Recent-ly, the Thatcher administration has tried tostreamline industrial policymaking by mergingthe National Research Development Corp. andthe NEB into a "British Technology Group."One goal has beet z -) temper the activist poli-cies of the NEB, which enjoyed considerableautonomy in the past. There is no indication

yet that this will produce positive results. Tobe fair, industrial policiesof whatever stripeare a limited tool when the overall economicsituation has been as grim as Britain's. WhileU.K. industrial policies may seem neither effi-cient nor effective, they have perhaps beenasked to do the impossible.

West Germany

Industrial policy in the Federal Republic ofGermany (FRG) has been distinguished by re-liance on the market. Objectives have beenallowed to remain vague beyond the level ofmacroeconomic policy, where stability hasbeen paramount. But if private sector actionshave been central, this does not mean the roleof the pubic sector has been negligible. Follow-ing a "social market philosophy," the West Ger-man Government has helped reconcile nation-al, regional, and interest group concerns: TheAct for the Promotion of Stability and Eco-nomic Growth (Gesetz zur Fi5rderung der Sta-bilifdt and des Wachstums der Wirtschaft) pro-vides a set of tools to coordinate economic pol-icymaking among government, management,and labor aimed at "macroeconomic equilibri-um." While avoiding extensive planning, pol-icymakers have paid consistent attention tostructural adjustment; since the mid-1960's andthe tenure of economics minister Karl Schiller,it has been widely accepted that policy instru-ments could be deployed to "rationalize" mar-kets and ease structural change. Especiallysince the mid-1970's, the FRG has also pro-vided considerable support for R&D. Althoughproponents of an avowedly sectoral approachto industrial policy have become more vocal,it is still true that industrial policies are market-oriented, with limited reliance on public own-ership compared to a number of other WesternEuropean nations, and a strong commitmentto open international trade. Nevertheless, theGerman Government- has sometimes takenstrong and direct action on the sectoral levelwhen--eeononns- problems_have_arisen___

The Institutional SettingEconomic and industrial policymaking in the

FRG combines elements of decentralized deci-

403


sionmaking with representation by major in-terest groups, including labor. The ministriesof Finaince, Economic:3, Labor and Social Af-fairs, and Research and Technology are amongthe mere influential in terms of policies affect-ing industry. Macroeconomic policies are de-veloped and implemented by a number of agen-cies. The Ministry of Finance submits 5-yearplans. Since the early 1960's, a five-personCouncil of Economic Adviserscomprised ofacademics not otherwise'attached to the gov-ernmenthas been responsible for macroeco-nomic forecasting. The Council, also preparesannual reports on the health of the West Ger-man economy. Money supply is the responsi-bility of the Deutsche Bundesbanklegally in-dependent of the government, though closelytied to it. Policies and analysis related toeconomic and industrial development are cen-tered in the Ministry of Economics. TheLander (state) governments help formulate eco-nomic as well as regional development policies.A joint Federal-Lander panning committee, forinstance, draws up regional action programsidentifying growth points (schwerpu.!...torte) tobe promoted via investment grants. West Ger-many has emphasized regional developmentperhaps more heavily than any other Westernindustrial nation, with the Lander Govern-ments central to these efforts.°9

As chapter 7 pointed out, financial institu-tions have a special place in the West Germanpclicymaking structureas they do in Franceand Japan. Executives of the central Bundes-bank keep in close contact with public officials,and normally act in support of the govern-ment's economic policy. The Bundesbank'scontrol of the money supply gives it direct in-fluence over the value of the deutsche mark.During the years of rapid economic expansion,particularly the early 1960's, the bank helpedmaintain an undervalued currencya strategythat strengthened the export competitivenessof German goods but earned a good deal of crit-icism from the cnnntry'c trading_partnerq

The Federal Government also holds majori-ty shares in five banks, while cities and stateshave their own financial imititutions. One ofthe nationalized banks, the Kreditanstalt fin.Wiederaufbau, is a development bank that pro-vides funds to commercial lenders.'') While thefinancial communities are major seats of in-fluence over industrial policy in both West Ger-many and France, they function quite different-ly in the two countries: rather than selectivecredit for favored firms and industries as inFrance, German banks have supported fiscaland monetary policies oriented toward aggre-gate growth.

If economic and industrial policymaking inWest Germany is less centralized than inFrance, the lines of responsibility are moreclearly drawn than in Great Britain. While theResearch and Technology Ministry (BMFT),

tends to approach industrial policy witha perspective quite different from that of the.Economics Ministry, the division of authorityis more or less predictable and consistent. Ger-many's parliamentary system has seen fewchanges in government since 1949; when a dif-ferent party has come to power, overall objec-tives such as maintaining the country's exportstrength while controlling inflation have beenretained.

Policymaking Processes

A distinctive feature of the German systemis the broad representation of interests, the ef-fort made to integrate diverse points of view.The Stability and Growth Act empowers theFederal Government to provide "orientationdata" for policy measures to be "simultaneous-ly and mutually agreed upon" by Lander andlocal governments, labor unions, and employ-ers' associations. In the late 1960's, "concertedaction" incomes policies were developed,aimed at consistency in approach among gov-ernment bodies and socioeconomic groups.onbudgetary matters as well as wages and prices.Cannarteci-action-was-not-an a empt-tcr-sup-----

't0G. de Carmoy, "Subsidy Policies in Britain, France and WestGermany: An Overview," International Trade and Industrial Pol-icies, S. J. Warnecke (ed.) (New York: Holmes & Meier, 1978),p. 52.

7°E. Owen-Smith, "Government Intervention in the Economyof the Federal Republic of Germany." Government Interventionin the Developed Economy, op. cit., p. 176.


plant the monetary and fiscal policies of theFederal Government, but was intended as anadjunct and complement to these, the basic ob-jective again being the creation of an environ-ment conducive to economic growth. Like co-determinationwhich ensures labor a voice inplant operationsconcerted action sought tointegrate labor and other interest groups intothe mainstream of policy formulation. Today,concerted action has fallen into disuse, butspeculation on its revival regularly surfaces; ifnothing else, this indicates the persistence ofthe view in Germany that sound economic andindwArial policies depend on broadly basedconsensus-building.

Indeed, institutionalized participation bymajor social groups appe I's to offset much ofthe fragmentation that otherwise might seemto characterize industrial policymaking in theFRG. As in a number of other Western Euro-pean countries, notably the Scandinavian na-tion:I, Germany's industrial policy is markedby concern with labor issues. In 1974, theBMFT and the Ministry of Labor and SocialAffairs set up a joint research program on "hu-manization of the workplace." Directed notonly at health and safety issues, the programaims as well to identify and encourage organi-zational changes that would increase job satis-faction (ch. 8). A number of studies gponsoredby the program, which is oriented s'±rongly to-ward field experiments and employee partici-pation, have explored impacts of autoLoationand computer technologies."

The systematized participation of labor inWest Germ -.ny is especially noteworthy,in con-trast to Japan or France. In Japan, organizedlabor is in factif not always in appearancerelatively powerless; the consensus so clearlyvisible in Japan comes, not from full participa-tion, but from a rather passive acceptance byother groups of policies that business and gov-ernment have agreed on. In France, organizedlabor is vocalwith a marked radical castbut labor participation in setting policy has notbeen internalized as in Germany. French

""Research on the Humanization of Work," Ministry for Re-search and Technology and Ministry of Labor and Social Af-fairs, document No. 2181/74v.

unions traditionally exert pressure on thegovernment through political activism, oftenconfrontational. Even with Mitterrand and the_Socialists in power, this is not likely to changemuch.

Policies Toward Electronics

Despite its stress on macroeconomic tools,West Germany has, over the years, instituteda considerable number of policies directed atspecific industrial sectors. Some have beenin portions of the economy where governmentownership has been widespreade.g., energyand banking. In contrast, sectoral involvementin electronics has been mostly restricted, toR&D; compared to both France and the UnitedKingdom, military involvement has not beenprominent. Moreover, in further contrast withthese two countries, when FRG officials at-tempted to encourage a "rapprochement"among Siemens and several other computermanufactufers, the large and powerful Siemensconcern resisted quite successfully." WhenAEG-Telefunkenafter Siemens the country'slargest electrical and electronics producerfellG1-1 hard times, the private .:sector at first dealtwith tilt: crisis on its own. A consortium of 24ommercial banks engineered a massive res-

cue effort, with financing totaling more than'half a billion dollars." Only when the bankers'efforts proved insufficient did the governmentstep in with a package involving further loanguarantees and export credits.74 As this implies,and as chapter 7 described in more detail, co-operation among induStry and financial institu-tions in the Federal Republic has been com-monand an increasing subject of parliamen-tary scrutiny and public criticism, on groundsthat the power of the banks is too great.

Despite efforts such as .the aborted Tele-funken rescue, government influence has notbeen exercised as directly in electronics as in

"Jequier, op. cit., p. 217.73K. Done, "The Last Chance Rescue," Financial Times, June

14,1982. Under West German law, banks can own equity in pri-vate firms and act as brokerage houses (ch. 7). German banksheld about 40' oercent of Telefunken's stock.

74See "Germany's Telefunken Insolvent," New York Times,Aug. 10,1982, p. Dl. Nonetheless, the firm entered bankruptcyin mid1982.

411


sectors such as energy-(where subsidies havecontributed to high domestic coal prices), steel(where firms such as Salzgitter are publiclyowned). or shipbuilding (where a range of pol-icy initiatives have been marshaled to shelterthe industry from decline)." Still, since the late1060's the West German Government hassought ways of strengthening the nation's com-puter industry. While on the whole the Germanelectronics sector has been the strongest of anyin Europe, it has shared the common weaknessin computer. Part of the reason appears tohave been that the bigger electronics firmsSiemens, AEG-Telefunken, SELwere alreadyheavily committed to other lines of businessconsumer products, telecommunications, elec-trical machinery. Much like such Americancompanies as RCA or General Electric, theselarge and diversified enterprises never devel-oped much strength as computer manufactur-ers. From the standpoint of the German Gov-ernment, there really was no computer indus-try as such to support. As a result, it proveddifficult to devise effective policies for encour-aging either the technology of computing orcommercial production. As in many othercountries, government procurements havebeen channeled to local firms. Nevertheless, incontrast to Japan and France, the FRG haslargely avoided attempts to shield the industryfrom foreign competition, relying instead ondomestic supports and subsidies.* A majorthrust of German efforts has been to stimulateutilization of computers through training pro-grams and applications support.

Although benefiting from government fund-ing amounting to more than 100 milliondeutsche marks (something over $50 millionin fact a relatively small fraction of West Ger-many's total subsidies during the 1970's forcomputers and information processing), Tele-

73See Owen-Smith, op. cit.: p. 174 on coal prices; p. 184 onSalzgitter; p. 173 on shipbuilding.

`Even so, a recent trade dispute shows thatin Germany aselsewhereforeign firms are often discriminated against, In acase similar to AT&T's choice of Western Electric over Fujitsuin fiber optics, Bremen University was forced to reverse a deci-sion to purchase a computer E y sie m from Burroughs. The con-tract went to Siemens at a price comiderably above the Americanbid. See "'Technology and Trade Policy: Issues and An Agendafor Ac',1on," op.' cit., p. 49.

funken never achieved much success in com-puter systems. Siemens remains the largestGerman-owned computer manufacturer, some-what ahead of Nixdorf in sales (ch. 4, table 42).But Siemens' production is far less than thatof IBM's German subsidiary; Siemens hasnever appeared to view computers as a majorpiece in its corporate strategy. The companyhas only about 20 percent of the German com-puter market, and less than 10 percent forEurope as a whole. Nonetheless, Siemens con-tinues to receive by far the largest share ofgovernment funds for R&D in computer tech-nology." The contrast with Nixdorfa man-ufacturer of business-oriented rriaicomputersis striking. Nixdorf is an aggressive world-wide competitor in its chosen markets, muchin the American mold; the company has ac-complished this with little government assist-ance.

Again in common with other European elec-tronics firms, a number of German manufac-turers have pursued ties with American andJapanese enterprises, one aim being technol-ogy acquisition. In 1978 Siemens purchased 20percent of Advanced Micro Devices. More re-cently, the company negotiated an agreementwith Japan's leading producer of computers,Fujitsu; Siemens now markets several of Fu-jitsu's IBM-compatible mainframes in Europe.Such arrangements have brought criticism ofgovernment support for Siemens as failing topromote an indigenous computer industry.

Research and Development SupportFinancial subsidies for Siemens' computer

efforts have been part of a considerably largerprogram of technology development in theFRG. Total R&D expenditures grew more than60 percent in real terms between 1969 and1980, increasing from 2.1 percent of GNP to2.3 percent; the West German Government has

766,. 1978, the West German Government had supplied Siemensawith cumulative total of 351 million deutsche marks (nearly

$200 million) for the development of large- and medium-sizedcomputers. See "Sixth Report of the Federal Government on Re-search," Federal Minister for Research and Technology, Bonn,1980, p 82. Thd computer support programs of the BMFT nowseem widely viewed as failures; they have been drastically scaledback.

Ch. 10National Industrial Policies 409 -

strongly supported R&D, which has accountedfor 3 to 4 percent of the Federal budget dur-ing the pa_ I decade." Since the initiation of thesecond data processino program in 1969, theelectronics industry, and particularly the com-ponents sector, has.received a substantial frac-tion of this spending; 30 percent of GermanR&D funds have gone to electronics (and elec-trical equipment) -- % greeer fraction than inany of the other cc. listed in table 77. Butwhile governmeL _uppurt for electronics as awhole increased steadily during the latter partof the 1970's, it appears to have peaked at theend of the decade, with the more :recent con-traction stemming from disappointing resultsin computer technology."

0

Disillusionment with support for computersystems has led industrial policymakers in theFederal Republic to reorient their programstoward microelectronics. Here funds havegone toward device physics and processingtechnologies, as well as IC design and develop-ment. Between 1974 and 1978 relatively modestsums were spent by the government on Micro-electronics R&Dabout $30 million annually.A somewhat more ambitious effort began in1979. Like the computer R&D which it sup-planted, VLSI support at first centered on thelarge electronics manufacturers. In its first pro-gram, the government contributed about $300million over)the period 1976-82, ;with industryparticipants putting up matching funds andcompeting on a proposal basis." The GermanVLSI program has been much less centralizedthan Japan's; participants work independent

""Sixth Report of the Federal Government on Research," op.cit., p. 75; "FRG's Position in World R&D Community Assessed,!'West Europe Report, Science and Technology, No., 72, Joint Pub-lications Research Service JPRS 78876,,Sept. 1, 1981, p. 40.

"See "Sixth Report of the Federal Government on Research,"op. cit.. p. 53. While the budget, for government expenditureson electronics R&D was schedided to increase from about 350million deutsche marks in 1975 and 1976 to more than 600,mil-lion in 1980 and 1981, expenditures in these later years werecut back considerably from the amounts originally planned.

" M . Gold, "West Germany Reported About To Launch $300Million VLSI R and D Plan," Electronic News, Apr. 30, 1979,p. 1; "Eurcipean Semiconductor Industry: Markets, GovernmentProgratris," West Europe Report, Science and Technology, No.134. Joint Publications Research Service JPRS 82686, Jan. 20,1983, p. 32.

99-111 0 - 83 - 27

ly, with the responsibilities of government of-ficials limited mostly to coordination andavoiding duplication. Siemens has received 25to 30 percent of the money, with Telefunkenand Valvo each getting 10 percent or more. In-dustry seems to have regarded the program asuseful but not of great impact; the major Ger-man electronics firms have traditionally hadstrong commitments to R&Dincluding basicresearchand government money appears tohave gone mostly to efforts that the private sec-tor has judged marginal. Indeed, a principal ra-tionale has been to finance projects with timehorizons too long for industry to justify.

While most of the money in this first majorVLSI program went to big companies, the FRGhas also paid a good deal of attention to smallerfirmsof which there are more than a thou-sand in electronics.8° In contrast to the marketorientation of other German industrial policyinitiatives, the BMFTa relatively new agen-cyhas designed an array of sector-specificprograms aimed at small enterprises andgrowth industries like electronics and biotech-nology. Small technology-based firms in theFederal Republic often face difficulty in rais-ing capital. As in most countries other than theUnited States, venture capital markets are min-iscule. Viewing this as an obstacle to innova-tion; and with motives much like those lead-ing to the creation of the National EnterpriseBoard in Great Britain, the FRG Governmentset up a venture financing company (DeutscheWagnisfinanzierungsgesellschaft) in 1975. Thisorganization purchases minority interests inGerman firms --with the intent of backing in-novative developmentswhile giving the pro-prietors preferential rights to buy back the equi-ty if their business succeeds.

The turn toward support for smaller compa-nies has also been reflected in the BMFT's latest

w"General Scheme of the Federal Government's Research andTechnology Policy for Small and Medium-Sized Undertakings."Ministry for Research and Technology and Ministry of Econom-ics. 1979 updating. Also see Innovation in Small and MediumFirms (Paris: Organization for Economic Cooperation and De-velopment, 1982), pp. 133-139; and G. Kayser, "Small BusinessPolicy in the Federal Republic of Germany," European SmallBusiness Journal, vol. 1, winter 1983, p. 39.

413


semiconductor program. The 3-year effort, be-ginning in 1982 and funded at about $45 mil-lion per year islike a number of Britain'smore recent initiativesdirected primarily atapplications.," This new program comes on topof a 40-percent increase in microelectronicsR&D support that had already been scheduled.Most of the applications money will be chan-neled to small firms, with one of the objectivesbeing job creation; of 1,000 grant applicationsreceived during the first 6 months, two-thirdswere from companies with fewer than 200 em-ployees. Administration is the responsibility ofthe VDI Technology Center, established bythe BMFT in 1976 specifically to help small-and medium-sized firms develop and apply mi-croprocessor technology.

The Fraunhofer Gesellschaft"West Germany's attentiveness to smaller en-

terprises does not stop with microelectronics.The Ministry of Economics supports more than80 industrial research associations, while theFraunhofer Gesellschaft (Association of In-stitutes of Applied Research, FhG)comprisedof some 25 institutes which function as R&Dlaboratorieshas as one of its major respon-sibilities the diffusion of technology to in-dustry, especially small companies.

Strengthening the FhG, which was foundedin 1949 to perform applied research and engi-neering development on a contract basis, hasbeen one of the more intriguing BMFT initia-tives. The FhG remained small unti!, a govern-ment decision in 1969 made the chief vehi-cle for support of applied research. At thispoint funding began to increase rapidly. A re-examination of FhG goals in 1973-74 led to astrengthening of its mandate for transferring

"'See "Increased Government Funding for Microelectronics,"West Europe Report, Science and Technology, No. 92, Joint Pub-lications Research Service JPRS 80133, Feb. 18. 1982, p. 5;"Special Microelectronics Program," West Europe Report,Science and Technology, No. 113, Joint Publications ResearchService JPRS 81392, July 29, 1982, p. 13:

"=Much of the information in this section is based on inter-views. See, also H. Keller, "30 Jahre Fraunhofer-Gesellschaft:Rilck-und Ausblick," FhC Berichte 3-79 (Munich: Fraun-hoferGesellschaft, 1979), p. 3, and Vertragsforschung filr Wirt-scha ft und Staat (Munich: Fraunhofer- Gesellschaft, 1981).

.technologies to the private sector, as well asdeveloping them.

Joint government-industry financing on aproject basis is the rule. FhG instituteswhichtogether employ more than 2,500 peoplepro-vide technical advice to smaller firms, coop-erate with universities, and function as tech-nology conduits. Institutes are organizedaround technical disciplines; one concentrateson semiconductor devices and processing tech-nology (the Institute for Solid State Technologyin Munich), another on computer systems (theInstitute for Information and Data Processing,Karlsruhe). Several others work in areas lessdirectly related to electronics.

The Institute for Solid State Technology, oneof the more successful of the Fraunhofer lab-oratories, can serve to illustrate the FhG model.Loosely associated with the Technical Univer-sity of Munichthe Institute's director holdsa chair there, and perhaps 20 students workat the laboratorythe Institute employs near-ly 100 people, about half of them engineers orscientists. This makes it the largest organiza-tion of its type in West Germany, and perhapsin Europe. Founded in 1974, housed in its ownbuilding away from the university, and grow-ing largely through the initiatives of its direc-tor, internationally known for his research insemiconductor technology, the laboratory gets70 percent of its annual fundingabout $5 mil-lionvia separately budgeted R&D projects.The BMFT typically provides a major share ofproject budgets, the remainder coming fromone or more industrial sponsors. In essence,the government shares risks with industry.Two of the Institute's staff members are paiddirectly by the BMFT to advise and consultwith small- and medium-sized' companies.Much of the laboratory's work is concernedwith processing technology; prototype circuitscan be fabricated, along with small lots of spe-cialized devices such as sensors and ICs formedical applications, The Institute also oper-ates an X-ray lithography facility at West Ger-many's synchrotron storage ring in Hamburg.

Perhaps the most noteworthy aspect of theFhG and its mandate from the BMFT is the ori-

4iil


entation toward commercial technologies. Theinstitutes are not basic research organizationsthat function remains with the Max PlanckGesellschaft. Nor do they function as govern-ment laboratories, although the relationshipsbetween individual FhG institutes and govern-ment agencies vary considerably; the ties areclosest among the six that carry out R&D fi-nanced by the Federal Ministry of De ense. Theinstitutes are a conscious attempt to speedcommercialization of new technologies and dif-fuse R&D results through industry. One way inwhich the FhG does this is simply to provide avenue for bringing representatives of Federal,Liinder, and local governments together withindustry and the universities. The Fraunhoferexperiment is an attempt to compensate or theweak links that exist in Germanyas in mostcountriesamong these groups, especiallywhere commercial technologies rather thanbasic research are involved. Likewise, the deci-sion to accept defense-related projects in 1955was based on the belief that it was better notto isolate defense R&D, but to combine it withcivilian work in hopes that each would benefit.(Defense-related projects now account for 20to 25 percent of the FhG's effort.)

Within Germany, the Fraunhofer Gesell-schaft has won high marks for facilitating tech-nology transfer while avoiding direct govern-ment involvement in decisions on directionsand priorities, but its comparatively smallbudgetabout 230 million deutsche marks in1981, something over $100 millionlimits theassistance that flows to any one industry."

The FutureIf a joint strategy for the European Commu-

nity in electronics comes to passa prospectthat seems slight, as discussed in the next sec-tion, but not so improbable as a few years agoWest. Germany's industry would probably be

",Vertragsforschung kir Wirtschaft und Steal, op. cit. In 1978,35.7 million deutsche marks 1818.7 million) went to the fourFraunhofer institutes ,involved in work related to microelec-tronics and information processing; see "Cooperative R and DPrograms To Stimulate Industrial Innovation in Selected Coun-triesWest Germany," Department of Commerce, National Bu-reau of Standards, Office of Cooperative Technology, November1979, p. 69.

the best placed of any in Europe. But in themore likely event, progress in electronics in theFederal Republic will dependas it has in thepaston domestic actions, public and private.

Past government policies, when directed atelectronics, have not been notably successful.Nonetheless, German industry has a soundbase to work from. Siemens, if not a leader incomputers, probably has the best semiconduc-tor technology of any company in Europe.(P is strong in linear circuits because ofits emphasis on consumer electronics, but Sie-mens was virtually the only European man -.ufacturer that recognized the importance ofMOS ICs at an early date.) Germany's domesticproduction of ICs has grown as a percentageof consumption in recent years, a sign ofSiemens' continued technical strength and per-haps of positive results from government R&Dprograms." But the entire consumer elec-tronics sector in West Germany, not just AEG-Telefunken, has faltered under the pressure ofJapanese competition. ZVEI, the Central Asso-ciation for the Electrotechnical Industry, hasclaimed that increased sales of imported homeentertainment products have been a directcause of shrinkage by such firms as GrundigAG, and consequent losses of jobs; at the endof 1982, Grundig and Philips filed an antidump-ing complaint against Japanese producers ofVCRs.85 In computers, West German firmshave less than 5 percent of world sales. Nix-dorf has chalked up respectable profits and ex-ports by concentrating on smaller business-oriented systems; the company has done thison its own, without significant government aid.As the example of Nixdorf showsa lessonrepeated in other countriesindustrial policyis no substitute for well-managed private firms.

West Germany has thus maintained its posi-tion in the second tier of the world electronicsindustry. Can it compete in the years; aheadwhen faced with both American and Japanese

"G. Dosi, Technical Change and Survival: The Europea» Semi-conductor Industry (Brighton, U.K.: Sussex European ResearchCentre, Sussex European Papers, May 1981.)

J. Gosch, "German Consumer Firms Face Bad Times,''Elec-tronics, Sept. 11, 1980, p. 97; "Japanese VTRs Are Target of ECAntidumping Case," Wall Street Journal, -Disc. 24,_1982, p. 9,

41.5


firms? At present, many of the industry's prob-lems stem from the brciader dilemmas of theFRG economy; high interest rates and low prof-it margins have made it difficult for German

mnpanieS, which have traditionally borne theoulk of such expenses themselves, to maintainhigh levels of spending for R&D and new cap-ital investment. In recent years, Siemens hasaccounted for as much as 12 percent of all WestGerman industrial R&Dfor many observersthis alone signifies imbalance."

West German firms also face a critical deficitin technical manpower, despite mounting un-employment in the nation as a whole. Accord-ing to the Association of German Engineers,16,000 jobs have been vacant for lack of peopleparticularly in electrical, mechanical, andcivil engineering; in 1980 only 3,600 studentswere enrolled in technical universities able toaccommodate 4,700.87 Such problems are in nosense unique to West Germanythe question'is whether government policies will help toresolve them.

How Effective Are West GermanIndustrial Policies?

Industrial policy', has a less distinct identityin the FRG than in\many other countriesatleast it is harder to summarize. On the onehand, the approach has been more market ori-ented than in France;',certainly planning andcoordination on the French model are absent.On the other hand, the West German Govern-ment has consistently supported industrialdevelopment through macroeconomic meas-ures and by integrating a broad range ofperspectives and interests into the policymak-ing process (critics in some countries mightregard this as !a weakness). The role of the

"According to Siernens' annual report for 1980, the company I

spent over 3 billion; deutsche marks (about $1.6 billion), more Ithan 9 percent of worldwide sales, on R&D. Over 90 percent ofthe money came from Siemens' own funds, the rest from govern-ment contracts and grants. For comparisbn, U.S. firms during f odic proposals that the European Communityhe same year spent the following amounts as a percentage of/

sales: Amdahl, 15.8 percent; IBM, 5.8 percent; Data General, 10.0' /(EC) d4velop a j,k)int polidy toward electronics.percent. See "Spending for Research Still Outposes Inflation,'} Rapid increases \in consumer electronics ship-Bto

ithin.,:ss Week, ',sly 6, 1981, p. 60. Siemens is clearly committed1 mentslfrom Japan have Stimulated talk of im-

eeping up in technology.

government, then, is far from laissez-faire. Incontrast to the British case, sectoral initiativeshave been pursued with a good deal of consist-ency over time, although such policies have notnecessarily entailed extensive involvement bygovernment officials. The West German casedoes underscore the critical importance of ag-gregate policies as necessary (if perhaps notsufficient) to sectoral development.

Industrial policy in Germany has benefitedfrom a better sense of timing than in the UnitedKingdom. Government support for R&D inelectronics began to pick up in the late 1960's,and has continued to growthis despite an on-going debate between the BMFT, which favorsexpanded sectoral thrusts, and the EconomicsMinistry, which continues to stress aggregatemeasures. While R&D programsincludingfunding for VLSI research and the efforts ofthe FhGhave not advanced the competitiveposition of the German electronics industry inany very obvious or dramatic sense, they ap-pear to have nurtured it in a variety of lessdirect and visible ways. Unlike electronicspolicies in nations which have tried to leapfrogthe competition, the German approach hasbeen one of broad support for more basic kindsof research, in the hope of returns over thelonger run.

As the Federal Republic struggles with ris-ing unemployment and continuing economicstagnation, such policies will be severely tested.Formulated in a time of overall growth, thereis no guarantee that the FRG view of industrialpolicy will prove adequate to deal with theadverse conditions promised by the rest of the1980's. Germany's problem is much the sameas that faced by the United States.

The European Community

In West Germany and elsewhere in Europe,concern over technology gaps vis a vis Amer-ican and Japanese competitors has led to peri-

Tagliabue, "Germany's Economy Stumbles," New .York port restraints, but a cominon effort in R&D hasTimes. Apr. 13, 1981, p. Dl. been the most frequent suggestion. A 1980

1


study by Siemens, for instance, held out littlehope for indigenous semiconductor industries;the conclusion was that continued growth inEuropean sales would probably benefit Japa-nese firms the most, with the U.S. market sharedropping from three-quarters to less than two-thirds by 1985.88

Although much of Europe has suffered sim-ilar problemsthe twin maladies of recessionand inflation, a perceived slowdown in tech-nological advance, rising labor costs, unem-ployment, low rates of capital investment, slip-ping competitivenessjoint responses havebeen slow to appear. In 1980, the EC's industrycommission proposed a European strategy inelectronics that would have included govern-ment-funded programs to develop semiconduc-tor processing equipment, as well as an ad-vanced communications network linking themembers of the Community.89 The proposal,which would have required modifications tonational procurement policies, stalled when theFrench dismissed it as insufficient while theBritish dithered over the implications of expos-ing ICL to open procurements. So, while theEC countries nave recognized the need for amore unified approach, national concerns havethus far remained paramount.

The latest attemptwhich bears the nameEsprit (European Strategic Programme of Re-search in Information Technology)got under-way in mid-1982. At first directed chiefly atsemiconductor processing, in part becauseEurope has been heavily dependent on im-ported processing equipment, Esprit will alsosupport work on chip architectures for VLSI,device modeling, and computer-aided circuitdesign and testing.99 The program has been

""Growth of Electronics Market in Europe Seen BenefitingJapan," New York ,Times, Nov. 28. 1980, p. D3.

""Europe's Electronic Strategy is Modest, But It Still Isn'tEasy." The Economist, July 26, 1980, p. 63. Over the years. theEC Commission has produced a variety of elaborate proposalsand studies, to little evident effect. See, for example: "New Infor-mation Technologies," Sept. 1, 1980; "Proposal for Council Reg-ulation Concerning Community Actions in the Field of Micro-electronic Technology," Sept. 1, 198C; "The Competitiveness ofEuropean Community Industry," Mar. 5, 1982; all Commissionof the European Communities. Brussels.

"D. Fishlook, "Why Europa Wants Esprit," Financial Times,Aug. 3, 1982, p. 13; J. Smith, "Can Europe Cooperate on Re-search?" Electronics, Aug. 25, 1982, p. 85."

carefully designed to avoid areas where coun-tries and companies compete directly. Fund-ing, planned ,`o be about $45 million over 3years, will be contingent on substantial con-tributions from the industrial participants,which number a dozen of Europe's largest elec-tronics firms (1CL, Siemens, Nixdorf, CII-HB,Philips. Olivettithe planning effort beganwith company managements rather than gov-ernment officials). EC planners hope the effortwill expand within a few years to encompassmore ambitious targetse.g., projects analo-gous to Japan's government-sponsored R&Dventures in supercomputers And fifth-genera-tion systems. It remains to be seen, however,whether the Europeans will manage to coop-erate effectivelyand, if they do, whethercooperation in basic research will make-muchdifference, given that many of the large Euro-pean electronics companies have always per-formed high-quality research but have had dif-ficulty translating the results into commercialproducts.

Japan

japan is the exception to many rules in theinternational electronics industry. Governmentpolicies evolved along with the industry; theyhave consistently supported private firms,directly and indirectly. Subsidies have beensubstantial, though not inordinately large com-pared with other countries. Both financial sup-port and indirect measures have been careful-ly targetedbenefiting some parts of Japan'selectronics industry much more than othersafeature that has attracted much attention in theUnited States. Consumer electronics, for exam-ple, has not been a major focus of governmentpolicy compared to microelectronics and com-puters; nevertheless, diiring the period of con-solidation and concentration that extendedthrough the 1960's, the government maintaineda series of barriers to imports and foreign in-vestment that effectively limited competitionin consumer electronics to local firms.'" Lib-

r.""Sources of Japan's International Cu-npetitiveness in the Con-sumer Electronics Industry: An Examination of Selected Issues,"prepared for OTA by Developing World Industry and Technol-ovy, .1.nc. under contract No. 033-1010.0, pp. 31-46; see also TheU.S. Consumer Electronics Industry (Washington, D.C.: Depart-

41Y


eralization began only in the late 1960's, as thegovernment's attention turned elsewhere; bythis time Japan's consumer electronics industryhad become well established. Foreign invest-ment controls on monochrome TV productionfacilities were relaxed in 1967, on color pro-duction 2 years later (in some contrast withEuropean governments, Japan limited inflowsof foreign capital as well as products). Like-wise, the tariff on color TV imoortsformerly30 percentdropped to 71/2 percent in 1971.Similar measures were adopted to protect thefledgling computer and microelectronicsindustries.

TV manufacturers clearly benefited fromgovernment support of broadcasting, from thearray of direct and indirect trade barriers thatJapan erected during the postwar years, andfrom policies that encouraged exporting. Still,direct and positive supporte.g., for R&D andproduct developmentwas modest comparedto the attention lavished on information proc-essing.Beginning in the 1960's, computers andsemiconductors have been at the center of pol-icies toward electronics and "the informationindustry." As these sectors grew, Japanese pol-icyakers shifted directionaway from thecomplex of measures for protecting domesticindustries that had been the hallmark of thegovernment's approach during the 1950's and1960's-, toward more positive measures. Ratherthan simply sheltering local companies, thegovernment sought to actively strengthen Ja-pan's capability in data processing, with theaim of moving into world markets. Financialsubsidies, primarily for R&D, were a major ve-hicle, along with other, less direct supports for..research, as well as measures to encourage andfacilitate applications of new technologies. Anexample of the latter is the Japan ElectronicComputer Co., which buys data processingequipment from computer manufacturers andleases to users (ch. 4).

Today the information industries are viewedas the flagship of the knowledge-intensive sec-

ment of Commerce. September 1975), pp. 12-13. and UnitedStatesJapan Trade: Issues and Problems (Washington. D.C.:General Accounting Office. ID-79-53. Sept. 21. 1979). ch. S.

tors at the core of Japan's emerging industrialstructure, the structure that will keep the coun-try's economy growing and competitive intothe next century. A unique feature of elec-tronics policy in Japansince copied by othernationsis official sanction and promotion,not only of the industry as such, but of elec-tronics as the epitome of a broad array ofemerging technologies (including CAD/CAM,robotics, composite and ceramic materials, andbiotechnology); the policies of Japan's Govern-ment toward electronics are in fact aimed atgoals transcending conventional sectoralboundaries. These policies, for years, have alsobeen consciously directed at leapfrogging othernation's technologiesanother aspect of theJapanese strategy that governments elsewhere,particularly in Asia, have tried to emulate. Inseveral respects then, Japan's use of the toolsof industrial policy has been innovative; Jap-anese policymakers have been both more am-bitious and more experimental than, for in-

41 d


Light emerging from glass filaments usedin fiber-optic communications

Ch. 10 National Industrial Policies 415

stance, their court lerpo rts in France or theUnited Kingdom.

The efforts of people like YonejI Masuclahave fed the broad consensus which evolvedamong :coders in Japanese business and gov-ernment concerning the critical importance ofelectronics, and particularly computers. Activesince the mid-1960's on advisory councils tothe government, Masuda was responsibleasExecutive Director of the Japan ComputerUsage Development Institutefor the 1972 re-port, "The Plan for an Information Society:Japan's National Goal :award the Year 2000."Respected academic and author of more than20 hooks, as a government advisor Masuda ad-vocated a comprehensive national plan for"computerization" in Japan, including govern-ment investment in future-oriented projectssuch as a "coinputopolis," or computerizedcity, and a computer peace corps. Masuda'sideaswhich are well within the mainstreamof this brand of futurism, based on the assump-tion that the production of information willgradually overshadow the production of ma-terial goods, eventually comprising the nextstage in economic developmentheavily influ-enced I'vlITI's (the Ministry of InternationalTrade and Industry) vision of a future infor-mation society.92 Most Japanese policymakerstake a more pragmatic view, but the visionaryoutlook of Masuda and others like him helpedcrystallize a broadly based consensus on theimportance of computer technology.

The Institutional SettingIn contrast to the United States or the United

Kingdom, a well-defined group of governmentagencies in japan bears the responsibility forofficial policies toward the electronics in-dustry. Both policy development and imple-mentation are centralized in MITI, specifical-ly its Information Machine Industries Bureau.Satellites attached to MITI include the Agen-cy for Industrial Technology, with functionsin R&D, and the Information Processing Indus-tries Advisory Council; a prestigious group

92Y. Masuria, The Information Society as PostIndustrial Socie-ty (Tokyo: Institute for Information Society, 1981), p. 29.

with membership drawn from the private sec-tor.

The only other public agency with significantongoing jurisdiction related to electronics is theScience and Technology Agency (STA), underthe Prime Minister's Office. In size and re-sources, STA cannot rival MITI. It does, how-ever, coordinate the government's budgetaryoutlays for R&D and related exper ditures, pre-paring, for example, an annual, ",science andTechnology White Paper." STA also funds re-search projects, including contract research byprivate firms, through its New Technology De-velopment Corp.93 STA influence over nuclear,ocean, and space technologies has been moreextensive than in electronics.

This is not to say that other government agen-cies do not develop policies that affect the Jap-anese electronics industry. They do, but on aless regular basis than MITI and STA; more-over, the influence of other agencies tends tobe less direct. The Ministry of Finance (MOF)has jurisdiction over macroeconomic matterse.g., fiscal and monetary policy. In recentyears, growing budget deficits have forced theMOF to weigh proposals for sectoral assistancemore carefully; competition for funds amongelectronics and other industriesas well aswith government objectives other than indus-trialdevelopment--has become stiffer. TiMOF also exercises a good deal of influencover the Bank of Japan, while public corpora-tions such as the Japan Development Bank canchannel funds to favored companies throughloans and grants (ch. 7). Long-term projectionsby the Economic Planning Agency includeforecasts of output by sector of the economythat are widely regarded as reliable guidepoststo future business prospects. While neitherpublic nor private banks need subscribe to thegovernment's investment priorities; they oftenput money into sectors targeted by such plans.

An independent body, the Fair Trade Com-mission (FTC)though peripheral in industrialpolicy compared to MITI or the MOFhas

91Cagakt, Gijutsu-cho/Kankyocho (Science and TechnologyAgency/Environment Agency) (Tokyo: Kyoikusha, 1979). pp. 48,77.

413

416 snternational Competitiveness in Electronics

often resisted policies formulated by thoseagencies. Examples include legislation exempt-ing sectors like electronics from provisions ofJapanese antitrust law to facilitate "collabora-tion" among firms for "rationalizing" the in-dustry." The FTC has repeatedly, though sel-dom successfully, opposed WTI recommenda-tions for antitrust exemptionsbut in contrastto the Japanese petroleum industry, where theFTC has frequently investigated particularcompanies, electronics firms have seldom beenscrutinized apart from matters of rebates andresale price maintenance. Even the public out-cry over price-fixing among color TV manufac-turers fueled by media reports of dumpingcharges against Japanese firmsin the UnitedStateswas assuaged informally ?ether than byFTC decision; MITI persuaded the companiesinvolved to lower domestic prices by 15 per-cent. Legal challe,igas to the business activitiesof Japanese electronics firms have come pri-narily from abroad: in the United States alone,opanese electronics companies have been in-volved in more than 30 lawsuits, the mejorityover dumping."

In addition to these traditional actors, otheragencies and organizations have recentlyfound more prominent roles. The intermin-isterial Council for SCience and Technologyhas been active in developing and coordinatinglarge-scale R&D programs. The Ministry of Ed-ucation has lailriched its own 3-year VLSI proj-ect. Diet (pug liamentary) committees dealingwith sciervx and technology have becomemore visib'e. Local governments have startedto court new technology-based industries;Kawasaki has put together a plan calling fortransformation into a "microcomputer city,"while Hiroshima has organized a council tostudy the impacts of high technology on itsestablished industrial base. Given this prolifera-tion, science and technology policy in Japanmay become more politicized in the years

"Kijoho no Kaisetsu (An. Explication of the Law for SpecialMeasures for Specified and Information Industries) (Tokyo: Min-istry of International Trade and Industry, 1979).

"Denshi Kogyo Nenkan, 1979 (Electronics Industry Annual,1979. Ministry of International Trade and Industry) (Tokyo:Denpa Shuppansha. 1979), p. 303.

ahead (in some energy research areas, such asnuclear power, this has already occurred).

Policymaking in JapanJapanese indus :ial policy is built on close

consultation among business leaders and gov-ernment officials. Corporate executives rou-tinely participate in both formal and informaldiscussions concerning policies toward elec-tronics. It is an overstatement to claim, as someobservers have, that in Japan industry tellsgovernment what to do, while in France gov-ernment tells industrybut this does conveya sense of the difference. The InformationProcessing Promotion Advisory Council, forinstance, brings together representatives ofJapan's leading electronics firms to discussMITI proposals. While such advisory councilsmeet relatively infrequently, and rarely havea determining voice in policy development,they serve to mobilize business interests andhelp form a consensus in support of the even-tual outcome. Advisory councils are only onesuch forum. Representatives of the many elec-tronics industry associations in Japan interactwith officials from MITI and other agenciesthrough a wide network of public and semipub-lic institutions. Several organizations bringtogether government, industry, and universi-ty leaders to stimulate work On computer soft-ware; the Information Technology PromotionAssociation (IPA), for one, had a 2.78 billionyen budget (about $13 million) in 1980, raisedfrom both public and private sources. Es-tablished in 1970, IPA organizes programsthrough which private corporations and IPAstaff conduct joint research on problems suchas computer-aided design or software packagesfor small businesses."

Similarl/the Japan Information ProcessingDevelopment Center (JIPDEC)a semipublicorganiztion with a staff of 150, the bulk ofwho are engineerswas established in 1967wi the support of MITI and the Ministry ofP sts and Telecommunications. JIPDEC's pri-

,"Konpula Hakjusho-1979 (Computer White Paper-1979). Nihon)oho Shari Kaihatsu Kyocini (Japan Information Processing De-velopment Associati;:n) (Tokyo: Konputa Ejisha, 1979). p. 94.

Ch. 10National Inoustrial Policies 417

mary mission is the marketing of software.Loans and grants for some of its programs hivebeen provided by IPA. Operating with a Si)million budget. JIPDEC carries out surveys oninformation processing, conducts R&D, sup-ports technical training and education, and en-courages information exchange through sem-inars and publications. Examples of JIPDECprojects include a microcomputer promotioncenter and an Institute of Information Tech-nology for retraining technical specialists.JIPDEC activities also led to the fifth-gen-eration computer project.

Government-Sponsored Research andDevelopment Projects

The fifth-generation computer effort typifiesJapan's approach to R&Dbringing togetherprivate sector firms, along with selected publicinstitutions. With funding from MITI and thebicycle racing association, the fifth-generationprojectwhich has attracted worldwide pub-licityis overseen by a 22-member panel in-cluding representatives from Tokyo Universi-ty, companies such as Fujitsu, and MITI.97About half the roughly $500 million budgetedfor the 10-year effort is to be provided by thegovernment. A research association (kenkyukumiai) was set up in 1979 to mobilize nine Jap-anese companies for R&D on microelectronicsdevices and peripheral and terminal equip-ment, as well as softwareall aimed at majorstrides in computing technology. JIPDEC's rolehas been largely facilitative; the research as-sociation now carries the primary responsibili-ty. The association's administrative staff hasbeen drawn from employees of the participat-ing companies, who are dividing the R&Deffort.

As discussed in more detail in chapter 5, thefifth-generation computer project is far froman independent or all-inclusive effort; its workis proceeding in a context of government-sub-sflized R&Das well as company-funded re-searchaimed at related aspects of informa-

""Fifth Generatimi Computers," JIPDEC Report, Japan Infor-mation Processing Development Center, summer 1980. The dis-cussion following also draws on interviews with MITI officialsin the Information Machine Industries Bureau.

tion processing. Likewise, the project is onlyone of a number of follow-ons to earlier MITI-sponsored activities such as the VLSI R&D pro-gram (discussed in ch. 5, as well as below) andthe Pattern Informatio-i Processing SystemProject (PIPS)." Such R&D efforts complementone another; they involve shifting groups ofpublic and private sector participants drawnfrom a wide range of institutions, In parallelwith the fifth-generation computer project,MITI is sponsoring the supercomputer effortmentioned earlier, along with a 10-year pro-gram on advanced microelectronic devices andwork on optical measurement and control. De-spite the funding that MITI provides, the Min-istry's officials seldom attempt to guide ordirect research, but confine their participationto helping shape objectives and to administra-tive functions.

Compared with other countries, Japan's ap-proach to aid for electronics is unique in atleast three ways: 1) government-supported pro-grams are multiple but carefully coordinatedwith one another; 2) they are oriented towardfacilitating the activities of industry, ratherthan telling industry what to do; and 3) the timehorizons are unusually long. The last point iscritical: the 8- or 10-year planning horizons formany current Japanese R&D projectswithevery indication that, while projects will beadapted to evolving circumstances, continui-ty will be preservedpoint to the depth of thegovernment's commitment Certainly there arefew analogs in the United States, even in de-fense researchwhere the 6-year VHSIC pro-gram is the exception, not the rule.

Cooperation in Research and Development

Observers in the West often misconstrue thenature of Japan's "cooperative" R&D efforts.While corporate leaders and government of-ficials do in some cases work closely with one

"PIPS has been much less visible in the United States thanseveral' of Japan's other R&D efforts, but it played a major rolein laying groundwork for the fifth-generation computer project.See H. Nishino, "PIPS (Pattern Information Processing System)ProjectBackground and Outline," Proceedings of the 4th In-ternational Joint Conference on Pattern Recognition, Kyoto, Nov. ,7-10, 1978, International Association for Pattern Recognition,p. 1152.

421

:116 International Competitiveness in Electron,cs

another during ongoing projects, the moreusual pattern has been a carefully planned divi-sion of Libor. MITI bureaucrats help initiatenew projectsafter lengthy preliminary discus-sions with industry advisory committeesbywinning budgetary approval. They also mon-itor ongoing programs, evaluating progress andjudging success. Government officials are oftendetailed to organizations like JIPDEC. Programadministration is normally delegated to repre-sentatives of participating firms, with ther,, search itself divided among these firms. Peo-pio from different companies seldom work sideby side.

The two government-supported VLSI proj-ectsparalleling one another in timeillustratethese patterns. The first, oriented toward com-munications, was carried out by the public cor-poration Nippon Telegraph & Telephone (NTT)under the aegis of the Ministry of Posts andTelecommunications. The second, directed atapplications of ICs to computers and much bet-ter known outside Japan, was sponsored byMITI; with 40 percent government and 60 per-cent private funding, the $300 million, 4-yeareffort took the form of a research associationlinking five participating firms. Three lab-oratories divided the work: a shared facilitymanaged by the VLSI Technology Research As-sociation; the Computer Development Labora-tory jointly run by Hitachi, Fujitsu, and Mit-subishi; and the NEC-Toshiba Information Sys-tems Laboratory. Staffs of the latter two lab-oratories came, not from the larger group ofparticipants, but from the companies operatingthem; the joint facility drew engineers and sci-entists from all five, as well as MITI employeesfrom the AIST. MITI was deeply involved inplanning and organization during the prelim-inary stages. Later, the teams from the par-ticipating companies independently carried outtheir assigned research tasks. Only in theassociation's joint laboratory was a real effortat cooperationwith technical people from dif-ferent companies working togetherunder-taken; this was a minor portion of the overallprogram, restricted to more fundamental re-

search.99 Individual firms did not cooperate oneither product designs or processing technol-ogy. Thus, while the MITI-sponsored VLSIproject has become known abroad as a "coop-erative" effort, the actual extent of interactionamong participating firms was limited; spokes-men for the Japanese electronics industry saythat dividing the research enhanced the overallsuccess of the project. It appears that the or-ganizational form involved a compromise be-tween attempts to encourage individual inter-actionswith objectives str.-..h as stimulatingpersonnel developmentand the more con-crete technical goals. Certainly as the workundertaken by joint R&D projects in Japanmoves toward development, interfirm coopera-tion declines; a MITI-orchestrated follow-on tothis VLSI project, which began in 1980 and em-phasizes chip designs and applications, takesthe fcrm of totally independent efforts by eachparticipant.

The work of the "Research Association forR&D on New Function Elements," also begin-ning in 1980, can be viewed as another follow-on to the VLSI project; it illustrates the way inwhich MITI-sponsored research efforts com-plement one another. This association's labora-tory draws on a larger group of companies.Matsushita, Sanyo, Sharp, Oki, and SumitomoElectricnone as strong in their technology asthe five companies that had participated in theVLSI projectwill all be involved in onemore of three major microelectronics develop-ment efforts.199 These are:

Three-dimensional circuit elementswhich can be visualized as more or lessconventional ICs stacked atop one an-other, increasing the density.High electron mobility transistors(HEMTs), one variety of which consists ofextremely thin layers of semiconducting

9,0Interview with Mr. Nebashi. IBMJapan and formerly at theVLSI Cooperative Laboratory, Nihon Keizai Shim bun, Jan. 19,1981, p. 1.

100"FY82 Government Projects in Electronics Listed." JapanRepoil, Joint Publications Research Service JPRS L110676, July22, 1982, p. 55.

materials such as gallium arsenide orgallium aluminum arsenide; these struc-tures carry the potential for higher switch-ing speeds. hence faster computers.Radiation- hardened devices suitable foruse in extreme environments such asnuclear powerplants or outer space (resist-ance to heat and vibration is a relatedobjective).

The first two especially will support both thesupercomputer and fifth-generation projects.

The Role of Universities

University-industry interactions in R&D arenu closer in Japan than in other countries

/ again perhaps in some contrast to the common// perception. Close collaboration is rare, even

though the rules prohibiting professors in thenational universities from working for privatecompanies can be circumvented. Contract re-search and consulting by university faculty aremore limited than in the United States.

Japanese policymakers universary expressthe wish that university-industry relations beimproved, and that sufficient numbers of welltrained professionals be available to meet theeconomy's needs. To date, however, little prog-ress seems to have been madenor, in fact,have new policy initiatives directed at suchconcerns emerged. As discussed in chapter 8,Japan's colleges and universities have for someyears been turning out more engineering grad-uates than in the United States. Nonetheless,as in other industrialized countries, there hasbeen concern over future shortfalls in the sup-ply of engineers and scientists; a resent surveycovering the hiring plans of more than 1,600Japanese.firms points to stiff competition dur-ing the-1980's for university graduates trainedin technical fields. 101

How Significant Are Supports andSubsidies in Japan?

As for any country, it is impossible to placea monetary value on the policy measures thatbenefit Japanese electronics companies. Nor

101"Daisotsu Danshi Nobi Niketa" (Number of Male GraduatesDeclines). Nihon Keizai Shimbun, Aug. 27, 1981. p. 1.


would an attempt at such an accounting bevery meaningful. Indirect benefitse.g., tem-porary exemptions from antitrust provisionsescape Quantification. Even when governmentfunds flow directly to industryas in the cost-sharing typical of joint R&D projects in japan,or the subsidies for the West German computerindustry during the 1970'sthe real questionsconcern the effectiveness with which themoney is spent.

Nevertheless, subsidies deserve special atten-tion in the case of Japan because the U.S. elec-tronics industry has argued that they have beena key to the competitive success of Japanesefirms. Research funding is only part of the totalpicture of industry-specific support, but astable 77 indicatesalid in common with otherindustrialized countriesmore than a quarterof all Japanese R&D expenditures, both govern-ment-funded and industry-sponsored, havegone to the electronics/electrical machinerysector. At the same time, government expend-itures on research are not high compared toother countries; considering only R&D, andcounting only expenditures directly related toelectronics, public funding is quite smallabout 1 percent of the total for 1978, accordingto the Japanese Government.702 This is hardlythe whole story; it does make the point thatR&D in Japan is primarily the responsibility ofprivate industry. Japanese R&D is heavily con-centrated on commercial applications; neithermilitary technologies nor basic research get theattention they do in other countries. Lookingat all R&D spending, the private sector in Japanprovides over 70 percent of total fundingmore than in the United States, where industryspending accounts for 50 to 60 percent (table77).

MITI's annual compilation of governmentsupports and subsidies for the "information in-dustry" is the most comprehensive listing of

nuln 1978, government bodies in Japan, including state andlocal, reportedly contributed 6.8 billion yen (about $34 million)to the total of 580 billion yen (about $3 billion) spent for R&Dor, "electrical machinery." This includes household el'ectricequipment. as well as communications and electronics. Mostof the R&D work is for development. See Kagaku Gijutsu KenkyuChose (Report on the Survey of Research and Development,Prime Minister's Office. Statistical Bureau) (Tokyo: Nibon TokeiKyokai. 1979), pp. 39-40, 94.

423

420 International Competitiveness in Electmni:s

programs related to electronics. (For the UnitedStates, no comparable data existin part be-cause no one agency has responsibility for suchprograms). For 1980, the Japanese Governmentbudgeted about S1.3 billion toward the develop-ment of the information industryexpendi-tures encompassing much more than just theR&D programs highlighted above; a large frac-tion of the total consists of loans and loanguarantees rather than direct grants.103 In-cluded in the total, for instance, is the morethan 5200 million that the Japan DevelopmentBank loaned to the Japan Electronic ComputerCorp. for lease financing; this aids Japanesecomputer manufacturers by reducing the fundsthey would otherwise have to commit to rent-al and lease arrangements with their cus-tomers. as well as absorbing risks associatedwith repurchasing.

The computer industry has received a sub.stantial share of direct subsidies. BudgeteeMITI expenditures for major projects closelyrelated to data processingincluding severalof those outlined aboveare listed in table 80.The table is not inclusive, and is intended onlyto give an idea of the magnitudes of typical gov-ernment expenditures. These sums are notlarge compared to R&D spending by industryitself in either Japan or the United States, orin comparison with government funding inother countries, Portions of such subsidies havefunded large-scale. long-term programs aimedat social applications of electronics technol-

thWenshi kogyo Nenkan 1979, op. cit.. p. 340.

ogies e.g., health care, regional energy saving,computerized traffic control systems. The fig-ures in the table also include money for con-ducting surveys on computer usage, adminis-tering qualifying examinations taken by com-puter technicians, and the costs to the govern-ment of special tax deductions extended tocompanies that train information processingspecialists.

Taken together, it is the comprehensive na-ture of such programsnet their spending lev-elsthat distinguishes Japan's policies towardelectronics and other targeted industries.'" Thevery fact that the government publishes an in-formation industries budget indicates the carewith which the bureaucracy monitors develop-ments in electronics and disseminates infor-mation among government, business, and fi-nancial circles. It is this attentiveness on thepart of government, and the fact that most pro-grams are coordinated by MITI, that sets Jap-anese industrial policies apart. Over the years,funding by the Japanese Government hasgrown, but the significance of MITI's initia-tives goes well beyond financial support; in-deed, to look only at the money spent is to un-

"The "Research and Development Project of Basic Technol-ogies for New Industries." established in late 1981. is anotherexample. The original plan called for total spending of about$460 million over 10 years: however, the first year's expenditumshave been scaled down by the finance-conscious MM. Privatecorporations are being funded to participate in one of 12 R&D"themes." such as biotechnology and advanced materials. The"New Function Elements" microelectronics projects mentionedearlier are also part of this umbrella program. See "A IST 1982."Agency of Industrial Science and Technology. Ministry of Inter-na.tional Trade and Industry. Tokyo. pp. 6 and 7.

Table 80.Japanese Government Expenditures on Selected ProjectsRelated to Computer Technology

Prgject

Budgeted expenditure(millions of dollars)a

1981 1982

Basic technology for nextgeneration computers $28 $22Basic software technology 24 20Microelectronics ("new function alements") 3.1 4.5Supercomputer, R&D 0.14 3.3Peripherals 4.8 2.6Fifthgeneration computer R&D 0.07 1.7

$60.1 $54.1

ariscal year basis. convened from yen at 220 to the dollar for 1981, 249 for 1982.

SOURCE: "PY82 Government Projects in Electronics Listed," Japan Report.JointPublications Research Service JPRS L10678,

424i


derestimate the impacts of such programs.They have considerable symbolic and psycho-logical value in galvanizing the efforts of manyparticipants behind a set of goals shared bygovernment and industry. Programs in elec-tronics have typically been aimed at breakingbottlenecks viewed as critical to continuedprogress. Both the VLSI projectwhich, as out-lined in chapter 5, was intended to help 7apancatch up to the United States in digital MOSICs=and the fifth-generation computer project,with its software push, have been designed toserve such purposes. The supercomputer proj-ect is quite different; not particularly. impor-tant in Any commercial sense, it is first andforemost intended as a highly visible symbolof Japan'', ability to compete technologicallywith the United Statesfrom the Japanese per-spective, supercomputers are one of the criticalpropaganda battlefields of the "computer war."

Comparing Japan's industrial policy with ef-forts in Britain or FranCe points to a major dif-ference: government policies in Japan are di-rected at further strengthening a private sec-tor that is vital and still expanding rapidly, notat revivifying a stagnant industry. Governmentprograms in Japan complement the dynamismand international orientation of the country'selectronics firms; they have contributed to, butnot created, their competitive ability.

Recent Trends

Majc. thrusts of Japan's industrial policyhave been aid and encouragement for exports

tional moves by Japanese electronics manufac-turers have for many years had the active sup-port of Japan's Government.

In the United States, many signs indicate thatJapanese manufactuiers are now often recog-nized as peers, Technical exchange agreementsbetween American and Japanese electronicscompaniesrather than outright purchases byJapanare on the upswing. Mitsubishi andWestinghouse have arranged a joint venture todesign and manufacture ICs. Hewlett-Packardis getting RAM technology from Hitachi. TheU.S. Department of Defense has persuaded Ja-pan to transfer defense-related electronicstechnologies to this country (although whatthese technologies will consist of is far fromclear). American semiconductor firms are set-ting up design centers in Japan, as well as pro-duction faciliteswhile Japanese firms do like-wise in the United States, each seeking to drawon the other's technical talent.

Movement toward cooperation amidst on-going commercial rivalries has not been con-fined to the initiatives of private companies.In response to criticism from the United Statesand elsewhere that MITI-sponsored electronicsR&D constitutes an unfair subsidy, Japan hassuggested steps in the direction of internationalcooPeration. For example, foreign firms havebeen invited to participate in discussions aimedat an enlarged fifth-generation computer proj-ect having the form of an international jointventure.'" SuOh proposalseven if carriddthrough,would not by themselves stem the ris-

of electronics and, to a lesser extent, overseas ---inglide of criticism aimed at Japan's industrialinvestment. The international activities of Jap- policies, as well as the country's indirect andanese electronics firms are especially visible nontariff barriers to trade. Still, if nothing else,in Asian markets, where interdependence is they are a sign of the confidence the Japanesegrowing (ch. 4). A study by the Electronic In- now:have in their own abilitieswhile also be-dustries Association of Japan forecasts strong ing a well-calculated public 1.1acions ploy.expansion elsewhere in the Far East, and urgers There ard two fundamental perspectives inJapanese firms to develop strategies of "accom-modation " promoting

United States on questions of J6, an'Japanese investments sub-:

sidies and indirect trade barriers. the oneand technology transfer, while importing low-technology, labor-intensive electronics prod- and, those who believe free flows o technol-

ogy to be a prerequisite for economy growthucts from other Asian nations.'" -These interna-

I05Denshi Sangyo no Kokusaika rio Hoko to sono Eikyo ni Khn-sun, Chosa Hokoku (Survey Report on Trends in the Internation-alization of the Eloctronics Industry and Their Influence, PartII on East and.Southeast Asia). op. cit., pp. 271-291,

1°6"Dai GoSekai no Conputa: NichiBeiQo no Kyodo KaihatsuShido" (Fifth-Generation Computer: Beginning of Joint U.S.-Japan-West European Joint DeveloPment). Nihon Keizai Shim-bun. Aug. 30. 1981.

425


and technical innovation call for equal accessby U.S. firms to programs sponsored by theJapanese Government. The other view is heldby those who would prefer restrictions on out-ward flows of U.S. technology in an effort toyeserve "technological security." As debatesin this country continue, the procurement,R&D, and customs and standards activities ofJapan's Government will be scrutinized bypartisans of both viewpoints.

How Effective is Japanese Industrial Policy?

Any judgment of the contribution of Japan'sindustrial policies. or government policiesanywhereto international competitiveness inelectronics rests in part on intangibles. Preciseevaluations are impossible. What is the"worth" of the networks for information trans-mittal.and consensus-building woven by MITI?What are the costs and benefits of the am-biguities and uncertainties surrounding an-titrust enforcement in the United States?

In judging the effectiveness of Japanese in-dustrial nolicy, the starting point is its basicthrusti -) cultivate rather than confine the na-

"tion's electronics companies.. The institutionalapparatus that has evolved over the years hascontributed far more than absolute levels offinancial assistance might indicate. The endresult' has been effective mobilization of institu-tional and human resources, 'comprehensive-ness in government efforts, a substantial degreeof policy integration without rigidity. The focusof Western observers on cooperation betweengovernment and business only hints at how thesystem works.

With few exceptions, Japan's Governmenthas used tho same policy tools to promote elec-tronics as other nations: in the early years,tariff barriers combined with controls on for-eign technology and capital flows; today, sup-ports and subsidies for R&D and commercial-

-ization While the highly publicized VLSI R&Dproject has been held out as a unique instanceof cooperation; one that would violate anti-trust laws in the United Statesunder closer

examination much of the appearance of inter-firm cooperation vanishes. The program waseffective because it was carefully craftcd tohelp Japanese firms overcome specific weak-nesses that MITI and industry leaders hadidentified: emphasis on linear devices, a legacyof production for consumer products; laggingcapability in the processing of large-scale ICs,because Japanese firms were dependent onsemiconductor manufacturing equipment fromthe United States; lack of experience in digitalcircuitry among engineers and technicians. Incontrast to governmentsupported R&D proj-ects in West Germany or the United Kingdom,the Japanese were able to define their needsand agree on a program that would help themcatch up to the United States. It is the consist-ent and coordinated attentiveness to the prob-lems and potentials of electronics (and otherindustries) that dish_ guishes the policies ofMITI and the rest of Japan's Government morethan the character of individual programs orpolicy instruments.

At a more general level, the long-term orien-tation of policies toward electronicstypifiedby the fifth-generation computer project--alsodistinguishes Japan from other countries. Fur-ther, development of the electronics industrywhile a goal in itselfhas been pursued forlarger reasons: electronics is viewed as the keyto Japan's overall industrial development, thefirst ingredient in the knowledge-intensive, en-ergy-efficient economy that the country's tech-nocrats are striving toward.

Japanese industrial policies have certainlynot been univeital tri u phsefforts to propup declining.sectorsTsY 1) or to counter inter-national market trends (petroleum) have notbeen particularly successful. But policiestoward electronics have complemented the dy-namism of private companies already well po-sitiOned both domestically and international-ly. It is the congruence of public policy andevolving shifts in industrial structure that, inthe end, is the hallmark of present-day Japanesepolicies toward the information industry.


Summary and ConclusionsAmong nations that have set out to promote

their electronics industries, the policy toolscome from a common list: R&D funding, in-vestment grants and subsidies, procurement,merger and trade policies. No avenue emergesthat can guarantee success in strengthening thecompetitive ability of a country's electronicsfirms. Under closer scrutiny, many of the pol-icies adopted by nations like Japansorrietimesthought to be unfair or uniqueare not so dis-similar from those used in other advanced in-dustrial economies, even the United States.Matters of timing, comprehensiveness, con-sistencyrather than, the types of policiesadopteddifferentiate the industrial policies ofvarious countries.

As competition in the international elec-tronics industry has intensified, governmentshave stepped in to help their own entrants. Inthe early 1960's, European and Japanese fearsover the "American challenge" sparked syste-matic attempts to protect and strengthen do-mestic computer manufacturers. At that time,the preferred policy approach began with tradeprotectiontariffs, controls on flows of foreigninvestment and technology, discriminatoryprocurements. Several countries encouragedmergers among computer firms. In the 1970's,as trade liberalization under the General Agree-ment on Tariffs and Trade continued, indus-trial policies shifted away from overtly protec-tionist and defensive approaches. Today, sup-ports Thr R&D, indirect subsidies such as taxincentives, and other less direct measures com-prise the fou.idations of public policies towardelectronics in virtually all countries. Except inthreatened sectors like consumer products,trade liberalization has been accompanied bya parallel movement toward policies with sec-ondary rather than primary effects on interne-

- tional flows of electronics goods. If there is anexception, it is the United Stateswhere, leav-ing aside defense-related policies, the mostprominent measures have continued to be reg-ulatory.

Can, then, industrial policies create com-parative advantage? The answer is clearly no.Competitive success in electronics, here andabroad, depends on many factors, of whichgovernment actions are only one. Taken alone,public policies are seldom as important as thecapabilities of a nation's private companies:human resources and their utilization, includ-ing the quality of management; costs and avail-ability of capital; technological ability in elec-tronics and the complementary infrastructure;overall market conditionsthese are more cen-tral to international competition. Public pol-icies can add or substract from them, but theability of governments to compensate for weak-nessesor to reverse declines in competitive-nessis circumscribed. Although they caneither help or hinder industrial development,public policies alone do not determinedirectly or indirectlythe competitive standingof electronics industries in any nation.

Today, policyrnakers in the U.S. Governmentmust decide whether to continue the ad hocapproach of years past or move toward meas-ures aimed more consciously at preserving andstrengthening the advantages that the Amer-ican electronics industry draws from its settingand structure. If the choice is to develop a morecomprehensive industrial policy, much can belearned from studying foreign experienceWest Germany's Fraunhofer Gesellschaft,Japan's VLSI project, Britain's schemes to pro-mote commercial applications. But no recipefor success emerges from the countries thathave experimented with industrial policy. It isone thing to-,say that policies toward elec-tronics should'be in tune with overall changesin industrial structure and international mar-kets; it is quite another to actually design andimplement an effective industrial policy amidst',the ongoing uncertainties and ambiguities that

the political and economic con-text.

Governmeip poliCies, thenas illustrated bythe countries examined in this chapterare

427


generally tailored to the level of technologicaland commercial development of firms in thelocal industry. It is no coincidence that the na-tion which led initially in semiconductors, incolor TV, in computersthe United Statesmade no attempt to devise a systematic policyorientation toward electronics. Where was theneed? Nor is it surprising that the countrieswith more comprehensive policies have gen-erally been those that have perceived them-selves at a disadvantage. But why have somecountries been more successful in mobilizinginstitutional resources to create sustained andcoordinated industrial policies than others?

A variety of forces work to enhance the abili-ty of government officials to design and imple-ment coordinated, timely, and comprehensivepolicies toward industries like electronics. Arelatively centralized policymaking apparatus,where a single agency or a select few have well-defined' responsibilities, is one. The grab-bagnature of British policies mirrors the agenciescharged, at one time or another, with policydevelopment. In countries where governmentofficials belong to a respected civil service theyare more likely to have the resources to analyzeand initiate actions with positive effects on in-dustry. The dominance of political appointeesin the United States, and their rapid turnover,works against the kind of consistency seen innations like Japan. So too does the laCk of un-derstanding of technology characteristic ofboth bureaucrats and politicians in this coun-try. An elite civil service does not ensure suc-cess, as the mixed record of French industrialpolicy shows. But especially in Japanwhereconsultation and cooperation between industryand government have been closer than in manyother countries, if not so close as sometimespictured in the Westconsensus is easier toachieve than in nations where adversarial rela-tions are the norm. Easier too is carryingthrough the actions that have been agreed on.Such factors have enhanced the effectivenessof industrial policy in Japan, the one nation thatliar; so far managed to catch upin at leastsome respectswith the United States. Ofcourse, Japanese electronics firms have beenfavored by other circumstances as wellskilled

labor supplied by a long-established educa-tional system is only one example. Structuralfeatures of the political and economic systemin Japan natural resource endowments, ex-isting capital markets, political stability, estab-lished mechanisms for policymaking, charac-teristic systems of labor-management rela-tionshave .tended to shape and limit indus-trial policy decisions, rather than the other wayaround.

A9 japan and other countries seem likely todiscover, it may be easier to develop policiesaimed at catching up than to devise strategiesfor keeping up or jumping ahead. For onething, as internationalization of industrial andmarket structures proceeds, the influence ofnational governments will diminish. But thefundamental point is that in any 'industry ortechnology, creating a new model is harderthan following a recognized leader. Goli-ernment aid has helped electronics firms inother countries improve relative to Americancompetitors; the situation for the United Stateshas beenand remainsdifferent. The leaders,be they American or Japanese, have to breaknew grounda commitment that industrialpolicymakers in Japan have long since -ffiade.Japan's publicly voiced determination to im-prove the technological base for the country'selectronics industry stems from a recognitionthat past successes have been built on the adap-tation and commercialization of technologiesoriginating elsewhere, mostly within Americanfirms. Now, Japan is a leader along with theUnited States. The public as well as the privatesectors in each face the need to devel-op appropriate strategies for the years ahead.

Industrial policies for the 1980's and beyondwill be most successful where policymakersgrasp the dynamics of ongoing shifts In domes-tic and international markets and industries.To the degree that public policies ignore or at-tempt to counteract, such forces, they will beless likely to reach their objectives. Policiesdesigned to complement and reinforce ongo-ing trends will be more likely to have positiveeffects. This is not to say that-public polic scannot help shape these trends. If it is true atthe industries which fueled postwar eco mic

2 d


growthsteel, petrochemicals, automobileshave attained a stage of relative maturity, thenemerging technologies are indeed an appropri-ate focus of government policy. Technologiesbased on genetic engineering, advanced ma-terials, computer-integrated manufacturingas

99-111 0 - 83 28

well as electronics and information procesSingwill contribute to growth in existing as wellas new industries. It follows that the appropri-ate emphasis of public policy may not be elec-tronics alone, but economic adjustment andtechnological development more broadly.

423/

CHAPTER 11

U.S. Trade Policies and.Their Effects

Contents

PageOverview 429

Tariffs; the Multilateral Trade Negotiations 430Tariff Effects 431Other Tokyo Round Agreements 436Dumping 438The Law and Its Administration 438The Color Television CaseProspects for Dumping Actions Elsewhere in Electronics

439441

Subsidies and Countervailing Duties 442Countervailing Duty Law and Its Administration 443The Question of Indirect Taxes 444Other Unfair Practices 445

Quantitative Restrictions and the Escape Clause 446Orderly Marketing Agreements for Color Television Imports 446Escape anise Proceedings in Color Television 447Effects of Quotas and Other Nontariff Restrictions 449The Escape Clause 449

1Prospective Effects of U.S. Trade Policy on the Electronics Industry 450Consumer Electronics 45'1Semiconductors 45;4Computers 457'


Table

Tdbit X. Pagei 1. Itypothetical Example Illustrating Tariff Effects on a Product With Nominal

Nlanufui:Inring Costs of 51,000 ; 431

CHAPTER 11

U.S. Trade Policies and Their Effects

OverviewAll trading nations develop policies dealing

with imports and exports. On, the import side,'such policies are usually intended to controlflows of incoming goods judged harmful to thedomestic economy. Formalized export policies,as a general rule, are less numerous and typ-ically intended to encourage overseas sales. Tothe extent that international commerce is re-stricted to trade and its financing, countriesmust export to be able to import, and vice ver-sa. Over time, exports will therefore approxi-mately equal imports. For such reasons, tradepolicies seldom have first-order effects in deter-mining overall levels of imports and exports,but tend to guide and regulate tradeinfluenc-ing, for example, the composition of a nation'simports. Policies can also be adopted to en-courage exports so that needed importso.g.,oilcan be paid for. Most common remain im-port controls serving to limit threats faced bydomestic industries.

In recent years, the governments of indus-trialized nations have, as matters of officialpolicy, generally taken the position that unre-stricted tradeor at least trade with minimumimpediments in the form of tariffs or similarrestrictionsbenefits all countries. Althougha principle often honored in the breach, nationsusually assume that relatively open trade is intheir self-interest. Countries import goodswhich they themselves cannot produce as ef-ficiently, and export products in which theyhave a comparative advantage ich. 5). Intheory, everyone is better off.

But while the benefits of open internationaltrade are spread widely across society, thecosts against which they are arrayed _tend tobe concentrated. .Individual companies, theiremployees, the cities and regions in which theyare located, bear the brunt of shifting patternsof trade and competition. When imports rise,the injured parties are more vocal than the

beneficiariesmany of whom do not realizethey are paying less for some of the goods andservices they purchase. Because of this imbal-ance, governments often raise barriers for pp-litical reasons, sometimes creating seriousdisruptions. The familiar example is the Smoot-Hawley Trade Bill, adopted by the UnitedStates in 1930, which raised the average U.S.tariff to more than 50 percent and was onecause of a steep decline in world trade. Morerecently, 'Japan has utilized a wide variety oftariff and nontariff barriers to protect develop-ing industries, including electronics.

Near the end of World War H and after-wards, the United States took the lead in effortsto establish a liberal world trade order. Thiscommitment has continued uninterrupted tothe present day. American leadership has beena major force in negotiations among tradingnations aimed at moderating tariff and, morerecently, nontariff barriers to trade. These ef-forts have taken place largely within the struc-ture of the General Agreement on Tariffs andTrade (GATT), an organization now compris-ing some 80 nations. GATT provides a forumfor negotiations together with mechanisms forresolving conflicts.

While trade negotiators have made :consid-erable 70f; 1'C% in redthcing tariffs, nontariffmeasuiu,, -,Are proving less tractablewithinGATT or on a bilateral basis. As more nationsdevelop industrial policies nominally for do-mestic reasons, the trade arena has taken ona new complexion: indirect and nontariff barriers have risen. as tariff walls have declined.The result has sometimes been termed "thenew prOtectionism." In essence, negotiatorsare struggling to fit the policy framework froman earlier eraGATT mechanism's have rootsin the 1940'sto a radically different setting.International corporations now compete insome parts of the world, cooperate in others,

432429


ship goods between subsidiaries located indozens of countries, and take advantage of na-tional industrial policies where they can. Gov-ernments design policies to attract foreign in-vestment and technology under some circum-stances, to keep it out under others. Traderelated complaints by U.S. firms embrace notonly the old-style unfair practicesdumping orexport subsidization to boost trade balances,predatory practices aimed at building monop-olies or cartelsbut asymmetries in the "rulesof tt.a game." The claim is that the industrialand trade policies of other nations tilt the rulesin their favor. Trade.negotiators will be faced,for years to come, With adapting rulemakingand adjudicating procedures to these newrealities.

This chapter briefly reviews the environmentfor international trade in electrOnics underGATT, then disCusses the trade policies of theUnited States, particularly as these relate to theelectronics industry. Only limited attentiongoes to other countries. The'chapter illustratesimpacts of trade policies and discusses policydirections that may be important in the future.

On the whole, the U.S. electronics industryhas been helped by the Federal Government'strade initiatives during the postwar period.Semiconductor and computer firms, in par-ticular, have benefited from the opening of in-

ternational markets. Much of their success hasbeen due to a global perspective and worldwideoperationsneither of which would have beenpossible without the open environment fortrade and investment created since World WarII. To be sure, foreign countries have oftenadopted policies intended to restrict inflows ofAmerican-made electronics products. But inmost though not all cases, such restrictionshave had effects that were marginal or indirector both. While trade barriers have sometimesencouraged U.S. firms' to establish overseasmanufacturing facilities, for many years Amer-ican electronics companies had such advan-tages in technology and cost that they wouldhave been potent competitors virtually regard-less of the trade policies adopted by other na-tions (the principal exception has been Japan).Still, these advantages have gradually dimin-ished over time.

Where technological change is less rapid andlabor costs more significant, trade policiescarry more, weight. In such products as televi-sion receivers, CB radios, and passive com-ponents, U.S. firms have-not been able to main-tain advantages in technology or manufactur-ing cost. Here, liberal U.S. trade policies havemade it more difficult for American firms tocompete effectively most notably in thedomestic consumer electronics market,

Tariffs; the Multilateral Trade NegotiationsThe General Agreement on Tariffs and Trade

provides the basic context for negotiationsamong nations concerning trade, and, whereneeded, for adjudicating disputes. Otherbodies, including the Organization for Eco-nomic Cooperation and Development (OECD)and the United Nations, play more limited rolese.g., collecting statistics. GATT is the pri-mary vehicle for multilateral trade negotiations(MTNs), the latest of whichthe so-calledTokyo Round, concluded in 1979resulted inan agreement which will be the principalframework for international trade over at least

the rest of the decade (another round of multi-lateral trade negotiations before the end of the1980's is unlikely). This Multilateral TradeAgreement was implemented in the UnitedStates by the Trade Agreements Act of 1979.'

Earlier negotiations under GATT had fo-cused on tariffs; although the Tokyo RoundMTN resulted in further cuts, negotiators con-centrated on such matters as quotas, customsprocedures, product standards, and public sec-tor procurement practices. Examples of Tokyo

'Public Law 96-39, July 26, 1979.

Round topic!: qf special relevance for trade inelectronics include:

A revised subsidies code, intended to, pro-vide a framework for dealing with nationalindustrial policies having the indirect ef-fect of subsidizing exports or otherwise af-fecting trade flows (as by giving domesticproducts advantages over imports).

(---\. Staging of tariff reductions for semicon-ductors accelerated by Japan in 1981 after.extensive bilateral negotiations with theUnited States, a similar acceleration ofJapanese tariff reductions on computersfollowing a year later.An agreement on government procure-Ment, where again negotiations betweenthe United States and Japan concluded, atthe end of 1980, in a bilateral accord moreliberal than that arrived at under the MTNframework.

In the United States, passage of the TradeAgreements Act of 1979 was accompanied bya reorganization of trade-related activities car-ried out by executive order. As discussed be-low, responsibility for dumping and counter-vailing duty investigations moved from theDepartment of Treasury to Corn , fee, whilea new Foreign Commercial Service was estab-lished in the Department of Commerce in placeof the commercial officers attached to theDepartment of State. At the same time, the Of-fice of the U.S. .Trade Representative was giventhe job of coordinating international trade ne-gotiations on a continuing basis. This reorga-nization followed mounting criticism of thefragmentation and diffusion of responsibilityfor trade matters within the executive branch.

Ch. 11U.S. Trade Policies and Their Effects' 431

Tariff Effects

As taxes on imported goods, tariffs directlyaffect price competitiveness. From the view-point of the country imposing them, tariffs c'inNerve multiple purposes. One effect is normallyto raise domestic prices; tariffs permit localfirms to manufacture at higher costs while re-maining competitive in the marketplace, pro-tecting domestic industries from foreign rivals.Alternatively, governments impose tariffs tocounter unfair trade practices such as dump-ing or export subsidies, or to retaliate againstrestrictions by other nations.

The impacts of tariffs on trade patterns arenot always so straightforward as the nominalpercentage rate would indicate; "real" rates ofprotection may exceed nominal rates by signifi-cant amounts. Table 81 gives a hypothetical butnot unrealistic examplea product (whichmight be something like a computer Aerminal)with a nominal produaior` cost of $1,000, pur-chased components constwating 80 percent ofthis, final assembly the remainder. The tablecompares two cases: 1' final assembly overseas,with the complete ,tc in imported and sub-ject to i tari, of dercent; and, 2) finalassembly in the United States, with compo-nents imported at a tariff rate of 5 percent. Inboth cases, the components are assumed to bepurchased abroad at the same cost. (Transpor-tation costs are ignored.) As shown, assemblyin the United States gives a cost advantage of$60. The real protective effect with respect tothe operations carried out domesticallythe"effective rate"would then be $60/$200, or30 percent. This percentage can be interpretedas the amount by which domestic costs of as-

Table 81.Hypothetical Example Illustrating Tariff Effects on a Product WithNominal Manufacturing Costs of $1,000

Foreign assembly U.S. assembly

Cost of components $ 800 $ 800Tariff on imported components (5%) 40Cost of assembly 200 200

$1,000 $1,040Tariff on imported system (10%) 100

Total cost in the United States $1,100 $1,040SOURCE: Office of Technology Assessment.

432 International Comootitivenes in Electronics

semhly could exceed foreign costs beforeAmerican firms would begin to lose competi-tiveness. As a result, even where nominal tariffrates are identical, protective effects can dif-fer; each case must be considered individual-ly. The example in table 81 is not atypical inthat tariffs or, parts and components gen-erally 'owe' Wm a tariffs on final pfuducts;wbrve this is kilo case, effective tariffs arealiv,!ys higher than nominal tariffs.

Tariff Changes in the Tokyo Round MTNand After

Nominal tariff levels on electronics productsvary a good deal, with the Tokyo Round re-sulting in significant changes for microelec-tronic devices and computers. A:, :.%entionedabove, tariffs on both semiconductors ,,,1. com-puters weir 111i! subject of bilateral negotiationsbetween the L itited States and Japan subse-quent to the multilatern= agn eruortt ftl

1981, at =reed to Ince tiffstegrated .rcA.4i (ICs) to 4.2 percent as of thebeginning of 1983. Originally, they were toLave dropped in stages, reaching the 4.2 per-cent level only in 1987. U.S. tariffs on ICs wentfrom 6 to 4.2 percent in 1982. Somewhat later,as part of a larger package of trade concessions,the Japanese Government announced a parallelreduction in tariffs on computers. The cuts,film 7 to 4.9 percentthe U.S. levelwent intoeffect at the beginning of 1983, rather than in1987 as again originally schediiled.2

As part of the Tokyc Round, the UnitedStates granted a variety of tariff concessionson imports of electronic products, but thesecuts will not have much impact because mostU.S. tariffs were already low. The reductionsseldom amounting to more than a few percent-age pointswill make little difference inlanded costs of imports. For example, theaverage level of tariffs on components (in-cluding passiv6 devices such as resistors andcapacitors, as well as semiconductors) andtelecommunications equipment will decline

1J. Robertson, "Japan Offers To Speed Up Tariff Cuts." Elec-tronic News. May 31, 1982. p. 1.

from 6.6 to 5 percent., Stagingthe sequenceof stepwise reductionsvaries by product; themost common pattern is yearly cuts over theper:od 1979-87 of about one-eighth the totalnegotiated concession. Likewise, duties on of-fice and computing equipment will fall fromar average of 5.4 to 3.6 percent. In certain ,

cases, the United States did not grant reduc-tions. `Jot surprisingly, these were generallyproducts where imports have caused problemsfor domestic manufacturers. Tariffs on colorTVs, for example, will remain at the currentlevel of 5 percent. Indeed, for items subject tosection 201 escape clause findings discussedbelow), of which this was one, U.S. negotiatorshad no authority to offer concessions.

Tariff reductions agreed to by countries:11 have been important export markets for

American electronics firms were generallysomewhat largerthough with important ex-ceptions. Many nations have maintained con-siderably higher tariffs than the United States;shipments of ICs into the European Communi-ty (EC), for instance, have been taxed at 17percenta duty that the Europeans declinedto reduce.4 The tariff wall has been steepenough that both American and Japanese firms

'MTN Studies, Vol. 6, Part 5. Agreements Being Negotiatedat the Multilateral Trade Negotiations in GenevaU. S. Inter-national Trade Commission Investigation No. 332-101, Subcom-mittee on International Trade, Committee on Finance, U.S.Senate, August 1979, p. 251. Computer parts, as well as periph-eral equipment, can be imported duty-free from some countriesas a result of the Generalized System of Preferences (GSP), underwhich the United States, the European Economic Community.and Japan have agreed to give preferential tariff treatment toproducts manufactured in developing nations. However, importsof such products into the United States under the GSP are ex-pected to remain small. Of those that do enter this country, manyoriginate in Americaff-owned facilities such as Texas Instru-ments' plant in El Salvador.

'A group of nations that did not join the European Communi-tyincluding Austria. Switzerland. Portugal, and several of theScandinavian countrieshave formed the European Free TradeAssociation, EFTA. In contrast to the ECwhich has commontariffs on importseach EFTA member sets its own duty levels.Once inside an EFTA country, however, goods can move free-ly within either EFTA or the EC without further tariffs. To keepexporters from channeling all goods through the EFTA memberwith the lowest duties, the Association has adopted a complexset of rules of origin. U.S. firms have sometimes charged thatthese rules are significant trade barriers. See Consumer Electron-ics Market in Europe (London: Frost & Sullivan. Inc.. 1978), p. 95.

3,F

Ch. 11U.S. Trade Policies and Their Effects 433

have built plants within the EC to avoid it.European countries did cut tariffs on a vari-ety of other electronic components and oncommunications equipmentbut for commu-nications especially, nontariff barriers remaina strong impediment to trade. Average EC tar-iffs on office and computing equipment willdrop from 6.9 to 4.9 percent. Overall, the Com-munity's reductions will have little effect oncompetitiveness because American electronicsproducts generally had significant price (ortechnology) advantages in the European mar-ket even at the-old tariff levels. The Europeancase Is a general one: reductions in tariffs byother countries will seldom have large net ef-fects on U.S. exports of electronics, if onlybecause nontariff barriers have usually beenmore significant (nontariff measures and theirimpacts are discussed in more detail in a latersection).

Reductions in Japan's tariffs must also bekept in perspective. The protective barriers thatshielded the Japanese computer industry dur-ing its earlier years have been coming downfor some time. In 1978, duties on mainframecomputers were cut from 13.5 to 10.5 percent,tariffs on peripherals from 22.5 to 17.5 percent.The further reductions to which Japan agreedare no surprise given that Japanese computermanufacturers are now highly competitive intheir home market. Likewise, accelerated stag-ing for ICs is evidence of the domestic in-dustry's strength; Japan's Government wastherefore willing to grant concessions in orderto reduce trade frictions with the United States.EC countries did: not feel they had this option.

Although both the EC and Japan have low-ered some of their tariffs on consumer electron-icsbut not on color TVsthis will have littleeffect on U.S. exports, which have not beenlarge. In Japan, prospective importers of col-or TVs face, in addition to tariffs, a commodi-ty tax levied on 17 categories of consumergoodsincluding automobiles, home appli-ances, and camerasthat adds 15 to 20 percentto the cost of imported as well as domesticallyproduced TV receivers.

Secondary Effects of TariffsIn addition to raising the costs of imports

compared with domestic goods, tariffs canhave a variety of less direct impacts on tradeand production; for instance, they may stim-ulate local investment by foreign manufactur-ers seeking to avoid the extra costs borne byimports. The complex patterns of U.S. directinvestment in electronics have been shaped bytariffs among many other factors. Foreign elec-tronics firms have also invested in the UnitedStates, particularly in the consumer sector;European and Japanese firms hold majority orpartial ownership positions in U.S. electronicscompanies ranging from producers of colorTVs (Magnavox, Quasar) to those designingand manufacturing sophisticated ICs (Ad-vanced Micro Devices, Fairchild) and com-puter systems (Amdahl).

Tariff barriers are seldom the sole cause offoreign investmentand may be minor factorscompared with the desire to locate R&D and/orproduction facilities closer to markets, or to ac-quire state-of-the-art technical knowledge. Still,tariffs can sometimes be a major consideration.In 1978, Nippon Electric Co. (NEC) opened asemiconductor plant in Ireland specifically tobe within the European Community.5 Produc-tion from this factory is not subject to the 17percent EC duty; semiconductors can alsobe sold in European Free Trade Association(EFTA) nations free of tariffs. NEC, like themany American firms th;it had made earlierEuropean investments, took advantage of whatis in essence a single market in Western Eu-rope. The opportunity to-reduce costs in sucha market, combined with the investment incen-tives provided by the Irish Governmentwhichwas seeking jobssufficed to attract NEC.Ironically, while both U.S. and Japanese firmshave been able to treat Europe as. one largemarket, local manufacturers have seldom beenable to manage this. The rather parochial at-titudes of both corporations and governments

SR. H. Silin, The Japanese Semiconductor Industry: An Over-view (Hong Kong: Bank of America Asia. Ltd.. January 1979).p. 161.

436


within the EC have hindered indigenous devel-opment. The Japanese case is quite different.There, relatively high tariffs on imports of elec-tronics were combined with restrictions on for-eign direct investment--imposed by the For-eign Investment Law of 1950 as well as strin-gent exchange controlsto protect the localindustry."

Secondary effects also arise when importssubject to tariffs are incorporated into finalproducts. While intended to shield domesticmanufacturers, say of components, these tariffsmay have the unintended consequence of rais-ing costs for firms making the final productperhaps harming' their competitiveness andeventually leading to demands for further pro-tection. Protection extended to the Americansteel industry, for instance, has increased costsfor U.S. automobile companies.

In the electronics industries of some coun-tries tariffs and other trade barriers havecreated incentives for internal production andvertical integration. When selecting vendors,companies weigh prices along with such fac-tors as quality and delivery schedules. Highproduct manufacturers to integrate backward,particularly where domestic suppliers havebeen protected because they were too weak tocompete effectively. Such factors have been atwork in both the EC and Japan; where manyfirms whose primary end products have beencomputerS or communications systems haveestablished internal semiconductor operations.The tendency has been especially pronouncedin Japan, where American semiconductorproducts were not as freely available as inEurope.

In the longer term, vertical integrationwhere semiconductor facilities produce for in-ternal as well as external salescould lead toscale economies that smaller U.S. merchantfirms may not be able to match. While Ameri-can firms have had the advantage in flexibili-ty compared with their integrated JapaneSecompetitors, and in products where innovativedesign has been critical for market success,they have not fared so well in mass-produced

',United StatesJapan Trade: Issues and Problems (Washing-ton. D.C.: General Accounting Office. September 1979). p. 27.

commodity-like products such as memory cir-cuits. To the extent that such patterns continue,they will imply that the tariff walls which pro,tected Japanese semiconductor thatiltlaCttiftiBfor so many years contributed to their eventualcompetitive success by making it expensive forthese companies to import for their own needs.

On the other hand, price competition fueledby imported components has probably bene-fited U.S. electronics firms that manufacturefinal products. Sectors like consumer elec-tronics and computers have gained from lowercost and better quality componentsthe con-sequences of heightened competition. Wide-spread foreign sourcing of components byAmerican manufacturers points to the poten-tial conflicts of interest with respect to importrestrictions that often arise between purchasersand suppliers.

Tariff Treatment of Offshore ManufacturingAmerican-made components incorporated in

imported goods have been exempted from tar-iffs for almost 200 years, The current versionof the law is embodied in items 806.30 and807.00 of the Tariff Schedules of the UnitedStates. Under specified conditions, shipmentsfrom overseas, plants benefit from duty-freetreatment of the value of materials or parts sentabroad for processing or assembly and then re-turned to the United States. Without this pro-vision, re-imports after offshore assemblywould be subject to tariffs on their full value.Because the tariff ex?mptions in items 806 and807 lower the cost of overseas production rel-ative to the no-exemption case, they Kritplicit-ly encourage American corporations to splitproduction between domestic and foreignplants. U.S. electronics firms began investingin production facilities in developing countriesas early as the 1950's. While central to costcompetition among TV and semiconductormanufacturers, offshore production has beena secondary element in the strategies of U.S.firms making computers and business ma-chines. Although labor unions have tended tooppose 806/807 On grounds that they encourage"exports" of jobs, the evidence concerning theactual extent to which this occurs remains am-biguous (see app. B).

43 /

'I he 806 and 807 provisions differ in scope.Item 806.30 is restricted to metallic articles,sent abroad for processing, which undergo stillmore processing after their return to the UnitedStates. Silicon wafers qualify under the typicalmanufacturing sequence outlined in chapter 6.Item 807.00, on the other hand, requires neitherthat the articles be metallic, nor that they befurther processed upon their return. However,there are three other conditions, not requiredunder 806.30: 1) the items. must have been ex-ported in a state ready for assembly; with noadditional fabrication needed; 2)they.niust notlose their physical identity; and 3) they mustnot have been advanced in value or improved.in condition except through the assembly.proc-ess.7 By value, the largest category of 807 im-ports consists of automobiles incorporatingparts originating here. Other major items in-clude clothing made from fabrics cut in theUnited States. Under both 806 and 807, tariffsare levied at rates.equal to those for equivalentarticles made wholly overseas but are basedonly on the value added abroad.'

Of the two statutes, 807 accounts for thegreater value of imports by far, in electronicsas in other product categories. Total value ofall 807.00 imports in 1980 was $13.8 billion,compared with $237 million for 806.30.8 Totalvalue of 806.30 electronics imports in 1980 wasonly $55 million, continuing a steep declinefrom more than $250 million 3 years earlier.9The major electronics imports under bothstatute items are semiconductors and parts,some of which qualify under either provision.Item 807.00 imports of semiconductor devicesincreased nearly threefold-- during the period1978-80, reaching $2.45 billionsomethingever three times the value of color TV ship-ments entering under 807.10 Imports under806.30 are being replaced by those under807.00 because of the 806 requirement for fur-

'Imports Under Items 806.30 and 80Z 00 of the Tariff Schedulesof the United States. 1977-80 (Washington. D.C.: U.S. Interna-tional Trade Commission Publication 1170. July 1981). p. 4.

°Ibid.. B-2.lhid B46. B48. The duty-free values run about two-thirds

of the total value."Ibid., pp. B-15. B-17. Color TV imports under item 807.00

can be found in ch. 4, table 14.


tiler processing. Offshore plants owned byAmerican semiconductor firms have been ex-tending their operations downstream, shippingcompleted rather than semifinished ICs backto the United States.

Semiconductor devices and TVs are nct theonly electronic products to enter under 806 and807. Modest volumes (in dollar terms) of ca-pacit-.;rs and vacuum tubes come in under806.30. Under 807.00 the list is much longer;it includes office machinery, communicationsapparatus, watches, stereo and high-fidelityequipment, and many types of components.

As the size of 806/807 flows indicates, thetariff exemptions have had significant impacton the global structuring of the American elec-tronics industry. Companies have rationalizedproduction by shifting manufacturing to partsof the world where costs are lower. In only afew cases have the tariff exemptions beendeciding factors, but they have certainly madeit easier for U.S. firms to move abroad. As dis-cussed in chapter 9 and appendix B, the effectson employment of such transfers are difficultto evaluate. Depending on the assumptions,they can be negative or positive. Even so, inat least some cases the choice may not be pro-duction here versus production there, but pro-duction there or no production at all.

Fri any event, much of the electronics indus-try today is globally integrateda trend towhich items 806.30 and 807.00 have contrib-uted. The consequences span a considerablerange. U.S. firms have retained competitive-ness in product lines where they would other-wise face marked cost disadvantages. Less-de-veloped countries have been helped to indus-trialize, while outward flows of American tech-nology have been accelerated. Some domesticemployment opportunities may have been sac-rificed. From a policy perspective, many of theimpacts by now appear irrelevant. The lawshave been on the books in one form or anotherfor decades, and are not likely to be rescinded.As tariff levels continue to come down, suchex imptions become more marginal to deci-sions on production locations; indeed, wagelevels rater than tariff exemptions have nearlyalways been the determining factor.

433


Other Tokyo Round Agreements

Ten distinct understandingscomprising theMultilateral Trade Agreement (MTA)camefrom the Tokyo Round negotiations. Some ofthese, each covering a particular subset of tradeissues, are irrelevant to electronicse.g., thaton dairy products. In other cases, little of sub-stance is changed under the new language; thisis the case for the antidumping and subsi-dies/countervailing duty provisions discussedin later sections.

Other MTA provisions pertinent to interna-tional tracde in electronics deal with:

government procurement,technical barriers to trade, andimport licensing procedures.

These agreements could yield dividends in theform of increased exports by U.S. electronicSmanufacturers, but are not likely to have mucheffect on imports of electronics products."Could" because the rather general nature ofthe MTA makes infractions difficult to pin-point. A series of test cases is likely, focusingat first on more blatant departures from the in-tentions of the codes.

The first of the three provisions listed above,that covering government procurement, callsin essence for nondiscriminatory treatment offoreign firms seeking access to governmentpurchases. That is, foreign and domestic bid-ders are to be treated the same. Exceptionsrelated to military sales and national securitywill doubtless be interpreted broadly. The stip-ulations -which cover purchases above about$200,000 -are rather far-reaching; they include,for example, state and local as well as nationalgovernments. On, the other hand, developingcountries are not bound by this part of theMTA, and virtually none have signed it.

The government procurement agreementalso addresses matters such as technical spec-ifications and notification of bidders, whichhave considerable impact in practice. Techni-cal specifications are, where possible, to behtlegrarl nn infnpnotinnol nearFnprnnninCs cfnnrinrrle

stating that invitations to bid should allow ade-quate time for foreign companies to respond.Obviously, considerable latitude remains forhindering foreign respondents, but grievancemachinery is to be established for handling thecomplaints of parties alleging discrimination.

The MTA procurement code could have far-reaching effects if it functions as written: Thegovernments of industrialized nations aremajor customers for many types of goods; ifthe provisions are fully implemented, thesemarkets would be opened to foreign suppliers.In actuality, this is not likely to happen veryrapidly. Imagine the repercussions in theUnited States if the General Services Admin-istration bought 5.000 Toyotas for the Federalmotor pool.

The second,of the listed agreementsthat re-lating to technical barriers to trade tackles,or presumes to tackle, the collages of PoliciPsused by governments in many countries to re-duce import volumes via diScriminatory tech-nical standards or regulations. This code is nottightly written, and leaves a number of loop-holes that could easily be employed to evaden:',--mingful compliance. For instance, govern-ments can promulgate regulations or productstandards different FT0171 international stand-ards for national security reasons, to preventdeceptive, practices, to protect health and safe-ty, to preserve the environment, and finally tohelp with "fundamental technological prob-lems." Such rationales liAve been marshaled inthe past to defend regulations that discriminateagainst foreign firms and, without much clues-

, tiOn, will continue to be so used in the future.This agreement, it is fair to say, is long onrhetoric but short on substance.

With the exceptions noted above, technicalregulations and standards are to be written soas not to discriminate among potential sup-pliers or be undue impediments to interna-tional trade. Where a country's regulations can-not be harmonized with international stand-ards, GATT and other interested parties are toreceive full notification of differences. Like-1AI; CO 1011 nrotnrn lintor foci-inn nrnnnt'-4nrCIC


parties to the MTA were at least in principlewilling to accept the notion of relatively free-access for foreign suppliers.

Whether or not the MTA code on technicalbarriers will have significant effects on com-mercial practices remains to be seen. In termsof U.S. exports, the extent tc which standardsand regulations elsewhere impede shipmentshas not always been clearleaving aside suchwell-known examples as the procurement prac-tices of NTT (Nippon Telegraph and Tele-phone) in Japan. U.S. exporters, in electronicsas in other iIclustries, have generally attemptedto sell goods abroad that are as close to theirdomestic production as pogsible. In somecases, exports have been stifled not by foreignstandards but by'the unwillingness of Ameri-can firms to cater to foreign-market conditions.

The third agreement relevant to trade in elec-tronics i!.; that on import, licensing. The text setsforth rather general stipulations intended tosimplify procedures associated with permitsand licenses, making it more difficult to, uselicensing procedures as nontariff barriersandespecially to single out and discriminateagainst particular countries. Because importquotas or Orderly Marketing Agreements fre-quently involve licensing requirements, com-panies attempting to gain, or hold market sharewhen such quotas are in effect have a specialinterest in equitable treatment. Perhaps themost important provision in the import licens-ing code states that any enterprise fulfilling theimporting country's legal requirements "shallbe equally eligible to apply and be consideredfor a license." The only exception relates to ap-plicants in developing countries, who are givenpreference. Governments signing the MTAalso agree, in awarding licenses, to take intoaccount: 1) economic order quantities or lotsizes, 2) past import performance of the appli-cant, and 3) "reasonable" distribution, of li-censes to new importers.

This brief review of MTA provisions pointsout the central difficulty now faced by inter-national trade negotiatorsnontariff barriers._ .

immense complexity and diversity of nontariffharriers in various parts of the worldandshould be regarded as no more than a first step.It represents an attempt to broaden the com-mon ground among participating nations, mov-ing beyond questions of tariffs and other directimpediments to trade while holding to thepremise that has guided negotiations since theoriginal Reciprocal Trade Agreements ACt,before World War IIthat free and open tradeis good for all concerned, with the distributionof benefits improved by concessions to less-de-veloped nations.

The ultimate impact of the Tokyo Round onnontariff barriers, and on future trading pat-terns, remains to be seen. As a statement of in-tentions, the agreementsincluding the newsubsidies codeare commendable. From anoperational perspective, the verdict is lessclear. Governments seeking politically accept-able reasons for eliminating some of their reg-ulatory clutter can begin; countries intent onmaintaining trade pintection will not findthemselves severely constrained. The courseof the world economy will also play a role;governments are loathe to reduce nontariff bar-riers during periods of stagnation.

In the context of electronics, the TokyoRound agreements have already had some ef-fect. For example, the U.S. Government hasbeen able to convince Japan to soften its standon exempting NTT from the provisions of thenew procurement code. NTT, a major pur-chaser of high-technology communications andswitching equipment, is notstrictly speak-ingan agent of the Japanese Government. Butits exclusion from the government procure-ment agreement created -a whirlwind of pro-test from spokesmen for tt.S.,industry; whobelieved the.exemption to be symbolic of con-tinuing efforts by Japan to evade the intent ofthe MTA while subscribing to its language.After prolonged discussions, the japaneseGovernment persuaded NTT to open its pro--curements to foreign bidders." Thus far, there

,,See, for example, T. J: Curran, "Politics and High Technology:


have been few foreign sale.; to the communica-tions giant.

In electronics and other industries, the even-tual consequences of the MTA for nontariffbarriers will depend on factors such as aware-ness among exporting farms of the possibilities

opened by the agreements. Without this aware-ness, and without the pressure on foreign gov-ernments that such awareness can generate,the agreements will have less effect. Equallyimportant will be attitudes of officials in im-porting countries who have responsibilities formonitoring and enforcement.

Dumping

The practice of dumpingselling goods inexport markets at less than their home-marketprice, or under some circumstances at less thancostis one of the unfair trade Practices re-stricted by GATT. In essence, dumping is aform of price discrimination; it is proscribedin export markets for_the same reasons as indomestic marketsbecause price discrimina-tion can be used to drive out competitors andconstruct monopolies. In recent years, asAmerican industries have faced stiffer com-petition from imports, the number of dumpingcomplaints has climbedfrom 11 in 1974 to 44in 1982.12

In electronIL:s, most of the dumping caseshave involved consumer products; there havebeen lengthy proceedings concerning TV re-ceivers, as well as products like CB radios. An-tidumping complaints were among the first at-tempts by American TV manufacturers to stemthe rising tide of imports in the late 1960's andearly 1970's. As other portions of the industryface increasing import competitionnot onlyfrom Japan, but in lower technology productsfrom developing countriesthe number of fil-ings may continue to grow. In recent years,American semiconductor firms have frequent-ly accused Japanese manufacturers of dump-ing, but have not filed formal charges.

The Law and Its AdministrationU.S. antidumping law is now contained in

two statutes: the Revenue Act of 1916 and theTrade Agreements Act of 1979. While the 1916

Act contains strong sanctions against pred-atory dumpingthat intended to eliminatecompetition and increase market poweritsapplication is narrowly circumscribed. An ac-tion filed in consumer electronics under thisstatute remained before the courts for someyears, but more generally the stipulation thatthe plaintiff demonstrate predatory intentmakes it unlikely that the Revenue Act of 1916will form the basis of future dumping find-ings.13 This leaves the Trade Agreements Actof 1979 as the primary mechanism for anti-dumpiroz enforcement. The 1979 Act modifiedU.S. law to conform to the revised GATT an-tidumping code negotiated during the TokyoRound." Although the Antidumping Act of1921 was repealed and the Tariff Act of 1930amended, with a few exceptions the substanceof the changes was minor.

According to U.S. law, dumping is the saleof foreign goods in the United States at lesstha.n "fair value." The 1979 Act transferredresponsibility for less than fair value deter-minations to the Department of Commerce;earlier, the Department of the Treasury had in-vestigated dumping complaints and made fairvalue determinations. The new act also short-ened the timetable for investigations, andchanged the definition and determination offair value somewhat; fair value had formerlybeen defined as foreign market valuebasically

U.S. Administration-of the Antidumping Act of 102i (Was/v.=_---ington, D.C.: General Accounting Office, Mar. 15, 1979).

""The Agreement on the Implementation of Article IV of theGeneral Agreement on Tariffs and Trade," 'Agreements

_ . _ .

the selling price in the country of origin, or,where such information was not available (thegoods might not be sold at home), the sellingprice in third countries. Prices formed the basisof comparison; the law allowed sales at lessthan cost provided the manufacturer also soldbelow cost elsewhere. if goods were sold onlyin the U.S. market, the old law specified thata "constructed value" based on estimated pro-duction costs be determined. In essence, cur-rent law extends the use of cost-based con-structed values to cover fair value determina-tions where goods are being sold below costeither at home or in third-country markets."Foreign firms th:. t, for whatever reasons, sellbelow cost at home cannot do so in the UnitedStates without risking dumping convictions,even under circumstances where this wouldnot otherwise be judged an unfair competitivetactice.g., when cash flows remain positiveeven though full- costs might not be covered.Earlier, sales at less than cost constituteddumping only in narrower circumstances. Thisprovision of U.S. trade law, which is not con-sistent With definitions of dumping in mostother countries, has meant that the Departmentof Commercenow responsible for antidump-ing enforcementloften finds itself estimatingoverseas production costs, an exercise f7aughtwith uncertainties and possible distortions."

iStatutory relief is available only when salesin the United States at less than fair value arefound to cause or threaten to cause "material"injury to a U.S. industry, or to materirlly retardthe establishment of a domestic ind,istry." Re-sponsibility for establishing injury or threat ofinjury rests with the U.S. International TradeCommission (ITC)an independent agency ofthe Governmentwhich weighs factors suchas actual or potential declines in output, sales,market share, profits, and employment. In theusual course of events, the ITC staff prepares

I 'Section -7 -7a, Trade- Agreements -Act of -1979, -Also -sec J.Sklaroff, "United States Antidumping Procedures Under theTrade Agreements Act of 1979: A crack in the Dam of NontariffBarriers," Boston College International and Compamtive Law


an analysis based on such considerations, afterwhich the six commissioners (appointed by ther)resident and confirrhed by the Senate for9 -yt.ar terms) voteeach making their ownjudgments as to injury or threat of injury. Com-missioners, singly or jointly, prepare writtenopinions that explain their reasoning. If a ma-jority of Commissioners find injury, the remedyis assessment of a special dumping duty intended to equalize prices between home coun-try sales and those in the United States. Theseantidumping duties are assessed and collectedby the Department of Commerce (formerlyTreasury).

The Color Television Case

The long and complex history of antidump-ing complaints in consumer electronicsstillnot fully resolvedwas no doubt cane of thereasons for provisions in the Trade Agree-ments Act of 1979 transferring responsibilityfor enforcement from the Department of Treas-ury to Commerce; advocates of stricter admin-istration of the law felt that Treasury officialshad been less than diligent, in part because ofthe Department's traditional commitment toopen trade.

Complaints that Japanese firms were dump-ing TVs in the United States began in 1968 witha filing by the Electronic Industries Associa-tion (EIA). This initiated what has perhapsbeen 1)- lengthiest case in the history of U.S.antidumping law." The EIA complaint allegedthat the Japanese were able to maintain lowprices in the United States for predatory pur-poses because prices in Japan were kept ar-tificially high by import barriers. The Japanesemanufacturers acknowledged that retail priceswere higher in Japan, but held that the dif-ferences were caused by higher taxes and bya complex and costly system of marketing anddistribution.

It took 3_years for the Department of Trees-ury to complete its investigation, findinginMarch, 1971that the Japanese had indeed

/ 440 international Co,/

dumped TVs in ththen went to ITCLater that year. ITconcluding that tthe U.S. industryassessment anddutiesat that titTreasury. Importrequired to post aduties.

Fixing the sizeintended to elevtlevel of TV pricelengthy process.tion provided byjudged inadequa

'mpetitiveness in Electronics

r

Color TVs undergoing long-term tests

te American market. The casefor determination of injury.

'C returned a positive finding,he dumped TVs had injuredand clearing the way for thecollection of antidumping

me still the responsibility ofters of TVs from Japan were9 percent bond toward these

of the antidumping dutiesate prices of imports ;o the

in Japanproved anotherThe wholesale price informa-failanese manure-auras-was--te and in some cases false,

\Photo credit: RCA

ing the duty rate. Eventually, the Departmentresorted to constructed value estimates basedon commodity taxes collected by the JapaneseGovernment.

An extraordinary number of claims andcounterclaims accompanied the efforts ofTreasury and the Customs Service to determineand collect these duties. Not only were Amer-ican manufacturers of TVs and components in-volved, but also the unions representing theiremployees. Arrayed un the other side were theJapanese manufacturers, their U.S. represent-atives, and the American firas which had beenimporting TVs s_from Japanmostly large retail-er s-s-u c h-at--Seetracted course of the disputes also mirrored

nrrucarnrrinnt


between the Customs Service and other partsof the Treasury Department.19 By 1980, onlyabout $13. million in dumping duties had beencollected. Moreover, assessment of duties for1975 and later :.,,fears has never been completed,pending final resolution of disputes coveringearlier periods. Not only have duties been atissue, but also civil penalties for alleged illegalrebates to importers as a means of circumvent-ing the added duties.

With the transfer of antidumping enforce-ment to the Department of Commerce in 1980,a new agreement was negotiated with import-ers. Commerce. agreed to accept a total of about$75 million, rather than pursuing in the courtsduties which the Department estimated at near-ly $130 million for the period 1971-79. TheEIAoriginal plaintiff in the dumping pro-ceedingsand its allies then claimed that theactual dumping liability was $700 million ormore, and challenged the Commerce Depart-ment's proposed settlement in the courts; a1981 decision allowing the settlement to standwas appealed to the Supreme Court. Late in1982, the Supreme Court denied the appeal;evidently Commerce's negotiated settlementcan now proceed. 20

This 15-year historywhich has still notcome to an end, and during which the com-plexion of the American consumer electronicsindustry changed irreversiblydramatizes the

iequacies of U.S. antiduinping proceduresas a means of relief from "unfair" import com-petition. The lessons hold for other industriesas wellwitness the example of steel. Not onlyis enforcement slow, complex, and suscepti-ble to delay by various parties, but the legaldefinitions of dumpingwhich, in the UnitedStates as in many other countries, predateGAVI.seem remote from the realities of busi-ness competition. No one argues that preda-tory practices should not be outlawed, but whatrelevance, for example, does the relationshipof home market price to export price have to

Wighunan, "Charges U.S. Blocks $400M Duty on Japan

predatory pricing? Would Japanese firms fora dozen years or more willfully cut the pricesthey charge in the United States below thosethe market would otherwise set because insome still longer term they seek to monopolizethe market? Does selling imported goods at lessthan costnow effectively prohibited by the1979 Trade ActaksZays..,constitute an unfairbusiness practice? Siill, reg-drdless of how thesequestions are viewed, the fact is that Japanesefirms were found under U.S. law to havedumped TVs. Injury to the domestic industrywas established. American manufacturers ofTVs have been entitled to trade protection buthave not received it. The uncertainty and con-fusion created by the these long and convolutedproceedings has probably done more damageto the industry than the dumping itself.

The modifications to U.S. antidumping lawincorporated in the Trade Agreements Act of1979 address some of the procedural problemsillustrated by the TV case. Not only has respon-sibility for dumping determinations and theassessment of duties been transferred fromTreasury to Commerce (by Executive Reorga-nization Plan No. 3, effective Jan. 1, 1980), butthe ITC injury investigation now begins im-mediately, rather than awaiting a positive find-ing of dumping. The concurrent investigationsfor which the act sets relatively short timeschedulesare intended to speed the process.If future dumping investigations are shorterbecause of the 1979 Act, this will limit uncer-tainties and disruptions, reducing costs forboth defendents and plaintiffs. This would alsomake it more difficult for domestic firms to usedumping proceedings in "strategic" fashion todeter foreign competitors from entering U.S.markets; dumping complaints can discouragemarket entry through the threat of futurepenalties as well as by imposing legal costs ondefendants.

Prospects for Dumping ActionsElsewhere in Electronics


but rare in senile nductors or computers. Thesingle case involving semiconductors, in 1972,led to a finding of dumping but not injury. Im-ports of c ,mputers into the United States havebeen at luw levels, leaving little reason to ex-pect complaints. If computer imports were torise and dumping to be alleged, less than fairvalue pricing would be difficult to establish; atleast for large systems. The complexity of suchsystems, the difficulty of establishing com-parability in performance, and the high R&Dexpenses that must be borne, complicate pric-ing comparisons. Moreover, selling prices fordata processing systems often include serviceor software charges that are hard to isolate.Pricing structures in the computer marketparticularly the establishMent of "quality-adjusted" pricinghave already created for-midable difficulties in purely domestic antitrustactions where predatory pricing has been atissue.21 Less than fair value' determinationsbased on foreign market prices or constructedvalues would be still more troublesome, at leastfor mainframes. The problems are not so in-tractable for small systems and peripherals,where significant import penetration is in anyevent more likely, while personal computerssold at retail could be treated much like otherconsumer products.

The characteristics of the semiconductor in-dustry also work against antidumping proceed-ings. Large-scale ICs--including computermemory chips, where import sales have in-,creased rapidlyexperience relatively shortproduct lifetimes. Coupled with the large econ-omies of scale in IC production, and the im-pOrtance of yields, deep market penetration-

21See, for example, R. Michads..'lledonic I'ricris and the Strutlure Of the Digital Computer Industry." Journal of Inthistriai Eco-nomics. vol. 28, March 1979, p. 263.

P17010 credit: Perkin-Elmer

Plasma etching system used in semiconductor fabrication

with resulting cost advantages from the learn-ing curvemight well occur before dumpingpro,:eedings could be resolved, even under theaccelerated timetable of the 1979 Act;moreover, the same factors lead to advancepricingwhich is not in general illegal. There-fore, while antidumping actions may continueto be filed in more mature sectors of the elec-tronics industrye.g., consumer ,Iroducts,where the technology is relatively stable andprice competition based on low productioncosts intensedumping allegations /in high-technology sectors seem less likely to escalatefrom verbal attacks on imports to formal com-plaints. In high- technoicgy industries, productscan be obsolete by the time dumping actionshave been resolved./

Subsidies and Countervailing Duties-inno thr, criortrum Frnm privortising in the form of credits or guarantees extended


agreements limit interest ratesto levels thatcan be below the market rate but not too farbelow.22 In such forms, export subsidies havebecome one of the most common nontariffmeasures affecting international trade; sub-sidies with domestic objectives also have con-sequences for exports.

Although important in capital goods in-dustries, neither export financing nor exportpromotion have played major roles in elec-tronics outside of telecommunications.23 Incontrast, subsidies with ostensibly domestic ob-jectives have become a major tool by whichgovernments promote their electrolVr:,s indus-tries; these have less direct and visible effectsthan export credits, making them difficult tocountervail or to negotiate over. While revi-sions to the GAIT subsidies code were a cen-tral item on the agenda for the Tokyo Roundnegotiations, little progress was made; thechanges were basically matters of procedure.Distinguishing export from domestic or inter--nal subsidiesthe latter of many forms butuniversally employedis central to a workablecode but fraught with practical difficulties.Measures adopted by governments that havethe effect of subsidizing domestic electronicsindustries range from grants for basic researchto regional development incentives. Becauseany such policy, even relocation assistance fordisplaced workers, could in principle helpfirms to exportby cutting costs, raising prof-its, or improving technical capabilitythe.dividing line between measures that most peo-ple would agree are domestic subsidies (e.g.,

"When Canada proposed a financing package for New YorkCity subway cars at an interest rate of 9.7 percentwell belowthe agreed international minimum of 11.4 percentthe Cana-dian Government defended this as necessary to meet France'soffer. See "Reagan Decides U.S. Should Not Match Financingfin NOW York Subway Car Sale," U.S. Import Weekly, July 14,1982, p. 448...

nror a discussion of export financing, see R. E. Shields andR. C. Sonksen, Government Financial Institutions in Support.of 1 -9. Exports (Washington, D.C.: Center for Strategic and In-ternational Studies, Georgetown University, September 1982).A nifire general review of U.S. export policies is Report of thePresident on Export Promotion Functions and Potential ExportDisincentives (Washington, D.C.: Department of Commerce,

R&D support) and what are clearly export sub-sidies (e.g., low-interest loans to foreign pur-chasers) will always be ambiguous. As nationspursue increasingly sophisticated industrialpolicies, it becomes still more difficult to drawthat line.

Countervailing Duty Law andIts Administration

GATT and U.S. law provide remedies paral-leling those for dumping where Americanfirms and industries are injured by export sub-sidies. In dumping cases, private firms set theprices at issue, while prices are distorted bydirect or indirect government action in the caseof subsidies.- Importing nations then imposecountervailing duties for essentially the samepurpose: to eliminate price differentials createdby the unfair trade practice. In principle, thecountervailing duty is set at a level that bal-ances the effect of the subsidy. In practice, theadministration of countervailing duties in theUnited States is even more problematic thanfor antidumping dutil.!s.

U.S. countervailing duty legislation is foundin two statutesthe Tariff Act of 1930 and theTrade Agreements Act of 1979. As in the caseof antidumping law, responsibilities in counter-vailing duty cases are splitthe Department ofCommerce investiaates foreign export subsi-dies (this responsibility was again lodged withTreasury until 1980); ITC determines injury.If ITC votes any of three findingsinjury to aU.S. industry, threat of injury, or impedimentsto the establishment of a new U.S. industrythen a countervailing duty equal to the netvalue of the subsidy is to be imposed on theimports.

Under the 1979 Act, the test turns on "mate-rial" injuryincluding actual or potentialdeclines in output, sales, market share, cashflow, profits, productivity, capacity utilization,employment, wage levels, or the ability to raisecapital." It: earlier years, U.S. law did not re-quire that injury be found before countervail-


subsidies was enough. Although differing fromGATT language, this provision had precededthe establishment of GA,TT and been retainedunder a grandfather clause. The original lawhad been passed before the turn of the century,but no countervailing duty was imposed by theUnited States until 1967.

The Question of Indirect Taxes

What then is a "subsidy" under GATT and/orU.S. law? As might be expected, the definitionshave been controversial. There has been a ma-jor legal action in consumer electronics, thecase hinging on whether exemptions or rebatesof indirect tax( :s on exported goods constitutea subsidy. bath the old and new GATT sub-sidies codes permit indirect taxese.g., val-ue-added taxesto be rebated, but not directtaxes. (Direct taxessuch as corporate incometaxesare levies based on factors of produc-tion like capital or labor, indir,..1ct taxes arethose levied on the product itsail.) The assump-tion underlying this rule is that indirect taxescan be easily included in prices and passedalong to consumers, while direct taxes cannot(they depend, for instance, on annualized profitlevels). If the full indirect tax is passed throughto the purchaser, profits to the seller are unaf-fected. Under these circumstances, a rebate orexemption of such taxes on export sales wouldnot constitute a subsidy under the usual defi-nitions.

Compared with its trading partners, theUnited States relies less heavily on indirectleviessales, excise, and value-added taxesand more heavily on direct taxation, primari-ly of cort,oraie and personal income. ManyEuropean nations impose a value-added tax(VAT) at each stage of the production process.In Japan, consumption taxes of 5 to 30 percentapply to items such as automobiles, electricalappliances (including TVs), and a variety ofluxury goods, while excise taxes apply to otherdasses,25 Under GATT rules, countries that

2 5 Flint) ri .Stimulation Prrigrams in the Maior Industrial Coun-

levy such taxes can exempt or rebate them asthey wish. The United States has less latitudethan nations with extensive arrays of indirecttaxes.

After the Trade Act of 1974 had been passed,Zenith challenged rebates of Japan's commodi-ty tax on exported TVs under the act's provi-sions. The American manufacturer sued theDepartment of Treasury, claiming that rebatedindirect taxes in Japan constituted subsidiesand that Treasury had failed to properly inter-pret the new law.26 Treasury countersued,claiming that decades of acceptance by all par-ties of its past practices had effectively ratifiedthese practices. Four years later, in 1978, theSupreme Court upheld Treasury's position)rui-ing that rebated commodity taxes do not con-stitute subsidies under U.S. law.

Countries with commodity or value-addedtaxes generally levy them on imports as wellas domestic production. Thu, within a coun-try having indirect taxes the impacts are, atleast in principle, neutral: both imports anddomestic goods are subject to a tax based ontheir value. However, matters are not really thissimple. Exports from a nation like the UnitedStates that relies on direct taxation may beburdened with higher selling prices reflectinghigher corporate taxes, thus at a disadvantagein maf.kets where indirect taxes are the rule(countries with substantial revenues from in-direct levies normally tax personal and cor-porate income at correspondingly lower rates).Furthermore, foreign manufacturers shippingto the United States may reap benefits: afterreceiving rebates on indirect taxes at home,such firms face no compensating border taxadjustments when their goods eater the UnitedStatesthough they generally must pay tariffs.They are free to sell in 'a market where theprices charged by domestic firms may wellhave to cover higher corporate taxes. As aresult, nations that rely heavily on indirecttaxes can be presumed to have advantages ininternational tradealthough the size and


significance of the advantages can be difficultto judge.

VAT systems have sometimes been suggestedfor the United States, in part --cause of theirpotential for stimulating exports (assuming cor-porate taxes were reduced at the same time).The effects of such a shift in U.S. tax policyon specific firms and industries would dependon factors such as:27

compensating reductions in income taxes,as well as the overall tax liabilities (andprofitabilities) of the firms in question;the extent of vertical integration charac-teristic of the industry, along with theplace of particular firms in the chain ofproduction;fractions of revenues stemming from ex-ports;,price elasticity of demand for each productaffected; anddesign and implementation of the systemfor collecting the VAT or other indirect taxand (optionally) rebating it for exp'irtedgoods.

While the merits of VATs have thus far notbeen seriously debated in this country, since1971, U.S. law has provided a mechanismthe Domestic International Sales Corp. (DISC)intended to put American exporters on amore even footing with manufacturers in coun-tries having indirect taxes. DISCssubsidiarycorporations whose activities are confined toselling goods in export marketspermit U.S.firms to defer a portion of tax liabilities fromprofits on overseas sales. Several thousandDISCs have been established, primarily bylarger American corporations with substantialvolumes of export business. In recent years,more than half of all U.S. exports have beenchanneled through DISCs.28 For exports of

"See A Value-Added Tax for the U.S.? Selected Viewpoints(New York: The Tax Foundation, Inc.. 1979).

"Export Stimulation Programs in tlw Major Industrial COUll-tries, op. cit., p. 319.

electronics, fv::vvever, the proportion is muchlowerperhaps in the range of 10 percent."

Other countries have registered complaintswith GATT against the DISC mechanism, argu-ing that it functions as an export subsidy butdoes not qualify as an exemption from indirecttaxes." Despite a finding by GATT that DISCsdo constitute subsidies, no country has yet im-posed countervailing duties on U.S. exports,nor has the United States offered to repeal thelegislation that permits DISCs. (Recently, theReagan administration has proposed an alter-nat've to DISCs, as pointed out in the nextchapter.)

Other Unfair Practices

Section 337 of the Tariff Act of 1930whichwas amended in 1979 but has seldom been useddeals with unfair competition in interna-tional trade not already covered by antidump-ing or countervailing duty laws. Most of thecomplaints filed under section 337 have con-cerned patent infringements, as in Apple'scomplaint to ITC over counterfeit computers,but in yet another case concerning importedTV receivers, two American manufacturers ac-cused Japalese firms of illegal predatory pric-ing practicesspecifically, of cutting prices inthe United States below costs in an effort todrive American firms from the market.3' Whenimports are involved, price-cutting complaintsare usually filed under antidumping or counter-vailing duty statutes, but section 337 actionscan also be brought if conspiracy or intent tomonopolize is alleged. In this instance, ITCproceeded with a section 337 investigationeven though the Department of Treasury

2,This estimate is based on a survoy of 325 member firms bythe American Electronics Associatior,. Because most of the mem-bers of the Association are smaller companies, it probably un-derstates the actual fraction for electronics. See Capital Forma-tion, Part 1, hearing, Senate Select. Committee on Small Business.Feb. 8 and 10, 1978, p. 53.

3°See J. M. McGuire, "The GATT Panel Report on DomesticInternational Sales Corporations: Illegal Subsidy Under GATT "Interrational Trade Law Journal, v11. 3, 1978, p. 387.

"ITC Investigation 337TA23. filed jam 15. 1976.


claimed exclusive jurisdiction under counter-vailing duty law. The theory behind the com-plaint was that assistance given Japanese TVmanufacturers by their governmentthoughnot necessarily bounties or grants within thedefinitions of countervailing duty statutes-might still constitute a conspiracy to restricttrade, an unfair practice under section 337. The

case was terminated when ITC issued consentorders prohibiting predatory pricing and spe-cial purchase inducements . 'color TVs. Fu-ture section 337 complaints by American elec-tronics firms are perhaps most likely as at-tempts to expedite relief, given the slow paceof past antidumping and countervailing dutyinvestigations:

Quantitative Restrictions and the Escape Clause

Over the 9ast two decades, tariff levels havebeen reducud by international agreement to thepoint that, for many goods and in many ad-vanced economies, they are no longer a majorfactor in market outcomes. Nowhere is thismovement plainer than in electronics. Withtariffs largely closed off as a legitimate vehi-cle for protection, industries exposed to therigors of international competitiontogetherwith their employees and political supportershave sought other forms of relief. Along withmany other nations, the United States has in-creasingly fallen back on import quotas. Bywhatever nameOrderly Marketing Agree-ment, Voluntary Restraint Agreementquotaslimit shipments originating in particular'coun-tries. Under GATT, unilaterally imposed quo-tas are explicitly disallowed except to correctpersistent balance of payments deficits, andthen are to be Jmpo ra r y . Nonetheless, quotashave proliferatedtypically on a negotiatedbilateral basisz--with the path often :;leared by"escape clause" actions permitted underGATT. An outline of the escape clause mech-anism in U.S. lawsection 201 of the TradeAct of 1974follows the discussion below ofquotas on color TVs.

Orderly Marketing Agreements forColor Television Imports

The only direct quotas on U.S. electronicsimeec-tg have heen termed Orderly Marketing

the result of unilateral action by Japan ratherthan negotiations between two governmentsexporting nations have entered into OMAs oftheir own volition.

The United States negotiated its first OMAcovering imports of color TV receivers in 1977with Japan. Under the conditions, Japan agreedto limit shipments of color TVs to this coun-try for a 3-year period; no more than 1,560,000complete sets and 190,000 incomplete setswere allowed each year. Except for being theoutcome of bilateral negotiations, the color TVOMA was equivalent to a quota of the type out-lawed under GATT.

The stop-gap nature of this first OMAcov-ering a single troublesome exporterwas il-luminated when Taiwan and South Korea tookup the slack (ch. 4). It quickly became necessaryto extend quotas to these two countries if theU.S. industry was to be effectively shielded.OMAs were negotiated with Taiwan and SouthKorea late in 1978, to expire at the same timeas the Japanese quotaJune 30, 1980. Importsfrom Taiwan were limited to roughly half a mil-lion units, plus twice as many incomplete sets(without picture tubes), over the year-longperiod beginning Jul- 1, 1979. Korean ship-ments were restricted to about 300,000 TVs.32This extension to other countries illustrates acommon failing: when initially directed againsta single exporter, quotas must often be wid-ened as now competitors step inthe series of


tifiber Agreement being the classic case. Notethat table 14 in chapter 4 shows imports fromMexico more than doubling over the period1976-80during which time the OMAs with Ja-pan, Korea, and Taiwan took effectwhileshipments from Singapore increased morethan seven times. Virtually all the imports fromMexico enter under item 807.00 of the TariffSchedulesmeaning they are shipments byArrLA `can -owned firmswhile both 807 andnon -807 imports from Singapore have gone upsharply; Singapore now ships more TVs to theUnited States than Korea did at the time theON1A with that nation was negotiated (table14). Might there be pressure for quotas withSingapore at some future time? Or other Asian.countries? If so, could unrestricted Mexicanshipments be justified simply because they areintracorporate transfers of U.S.-based multi-nationals?

As expected by the American negotiators,Japanese manufacturers responded strategical-ly to the OMA. To avoid the new restrictions,they not only invested in Taiwanese and SouthKorean manufacturing facilities but openedassembly plants in the United Statesa desir-able consequence from the viewpoint of theFederal Government because these plantswould help maintain domestic employment,d4fusing some of the pressure from laborunions. As these U.S. plants came onstream,Japanese shipments of color TVs (but not ofsubassemblies) diminished. By 1980, Japan'sexports of completed and nearly completed setswere no longer considered a threat, and theOMA with Japan was allowed to expire onschedule.. Of course, the possibility of a newquota continues to shape business decisions byJapanese exporters.

OMAs with Taiwan and Korea, on the otherhand, were renegotiated to cover the periodthrough June 30, 1982 at new levels (Taiwan:400,000 sets in the first year, 425,000 in the sec-ond; Korea: 385,000 sets in the first year,575,000 in the second), .after which they too

iinwnri In nnri of fine rnncram cs rra ana.n

predictable, were decision; by Korean andTaiwanese firms to follow the Japanese had inestablishing assembly operations in the UnitedStates.

Escape Clause Proceedings inColor Television

Why have the United States and other na-tions resorted to quotas? Partly because quan-titative restrictions are administratively cleansimple to monitor. More important, for a har-ried government, quotas may seem the bestchoice among a set of generally unattractivealternatives. The color TV case illustrates thepolitical dilemmas that often foster such deci-sions.

The OMA with Japan followed a series oflegal actions initiated by the U.S. industry inattempts to stem rapid increases in imports. Asdiscussed earlier, dumping charges against theJapanese came first, but for a variety of reasonsduty collection was repeatedly postponed.American firms together with labor unions rep-resenting their employees continued to pressfor import relief via other avenuesone beingZenith's countervailing duty suit, metdionedearlier and destined ultimately to fail. Theavenue that finally proved successful beganwith an appeal filed in October 1976 by a groupof 'companies and unions for relief under theescape clause, section 201(b) of the Trade Actof 1974. This provision, following article XIXof GATT, permits trade restrictionsindepend-ent of questions concerning fairnessif im-ports are found to be causing serious injury orthreat of injury to domestic producers. The pur;pose is to allow a temporary respite or escapefrom import competition while industries ad-just to new conditions. The protective meas-ures adopted in such cases, termed safeguards,need not be quotashigher tariffs are one 31-ternative.

In terms of the color TV OMAs, two featuresof the escape clause_ mechanism are note- _

Arr,nitisl, r. ;..e+ ne,ird.v eicsr.s_


liberalization. Without this change, protectionfor the American industry via the escape clausewould almost certainly`have been precluded.

Another feature new to the 1974 Actcon-cerning the role of ITC in the investigation ofinjurybore on the ultimate outcome of thecolor TV case. Under earlier law, when injurywas found ITC recommended remedies to thePresident, who could either accept or rejectthem. The 1974 Act added a time limit, stipu-lating that the President respond within 60days to an ITC injury finding. Furtherandmost significantthe act provided that what-ever action the President took could be over-ridden by a simple majority of Congress.*Thus, the options available to the executivebranch had been narrowed, the hand of thoseadvocating import relief strengthened. Thethreat of reversal by Congress greatly increasespressures on the Executive, for whom the col-or TV case posed a dilemma. The ITC Commis-sioners determined that the U.S. industry hadsuffered injury, andwith only one 'dissentingvoterecommended a large tariff increase. Ifthe President took this course, an internationaltrade dispute of major proportions wouldalmost certainly havl been precipitated. On theother hand, rejecting the ITC recommendationwould bring with equal certainty the prospectof reversal by Congresseven more embarras-sing. Under these circumstances, the WhiteHouse finessed the entire problem by negotiat-ing with Japan for voluntary restrictions. Dis-cussions carried out between the (then) Officeof the Special Trade Representative and`theJapanese Ministry of International Trade andIndustry (MITI) led to the OMA.

What have been the consequences for theU.S. TV industry? That the political victory hadany very substantial impact on its competitivevitality can be questioned. Imports were cutback, and the frontal assault by Asian firms ar-rested. The specter of U.S. manufacturers be-ing totally overridden, which underlay the ap-peals by industry and labor (thoughihe indus,

try did not in fact stand together on this), re-ceded. But the OMAs also accelerated a proc-ess begun earlierthe establishment of U.S.operations by Japanese TV manufacturers, andlater Taiwanese and South Korean firms. Sonyhad initiated the trend in 1972; since then,many others have followedas described inchapter 4sometimes by taking over the plantsof ailing American rivals. Wholly owned Japanese subsidiaries now supply perhaps one-third of the U.S. market (table 10). If Americanmanufacturers expected to recapture the do-mestic market, or if they anticipated a slacken-ing in price competition, they were disap-pointed.

The hill range of consequences providesother causes for reflection. OMAs did not stopthe transfer of U.S.-owned production facilitiesto foreign countries, a movement,that had be-gun earlier. Zenith, for instance, continued toshift TV manufacture to offshore plants inMexico and Taiwan. Still, if the industry doesnot appear to have gained materially from thequotas, it is likely that further losses wereavoided.

That competition did not abate is shown byprice data collected by ITC over the period ofthe initial agreement with Japan; retail pricesfor color TVs (19 inch and smaller) remainedessentially constant during a period of severeinflation in the U.S. economy." Even for large-screen sets, where U.S.-owned firms continuedto dominate the market, prices increased onlyabobt 6 percent. While price stability also mir-rors cost-cutting improvements in both productand process technologies, it seems :clear thatcompetitive responses by Far Eastern Manufac-turers were the chief cause. During the sameperiod, many household appliances rose inprice by 50 percent and more.

Nor did profits recover. While OMAs re-duced import market sharesin the 18- and19-inch categories, penetration declined fromabout 30 to 10 percent during the first year


in terms of the competitive position of the U.S.industry, this apparent benefit was partly off-set by the output of Japanese firms assemblingTVs here. Capacity utilization rates of domesticfirms improved but profitability did not follow.The ratio of net operating profits before taxesto net sales, which had declined from E.7 per-cent in 1:i;72 to a loss in 1974 of 1.2 percent,has been rn:1M:i.g at less than 2 percent in re-cent years, as pointed in chapter 4. Whilea'tfAs helped preserve domestic employmentopportunities, they provided no more thanmodest relief from competitive pressures.

Effects or Quotas and OtherNontariff Restrictions

Many nations have utilized restrictions otherthan tariffs to regulate trade in electronics.Japana major beneficiary over the past threedecades of vigorous advocacy by the UnitedStates of open international tradehas em-Tiloyed nontariff restrictions frequently and ef-fectively as part of its economic developmentstrategy. Among the more blatant nontariff bar-riers created by the Japanese has been MITI'sdefinition of domestically produced comput-ers. These are confined to systems manufac-tured by firms in which majority ownership isJapanese.35 Machines built within Japan byAmerican-owned firms are "foreign"despitethe fact that IBM-Japan, for instance, employssome 13,000 Japanese and only a handful ofAmericans. MITI has preferred that purchas-ers of computers chose "domestic" equipment,using controls over foreign exchange to helpenforce its wishes; although exchange controlswere dismantled in 1975, MITI continues tomonitor the market, and reportedly advisescustomers to buy Japanese computers.36

That nontariff restrictions appear to havebeen more effective in achieving their osten-sible goals in Japan than in most nations il-lustrates once again that evaluating industrial_policy-meaStiriiS-kSeldoinstraightfOrward.One lesson of the Japanese experience appears

to be that restrictions may work better in pro-tecting what are essentially infant industries,at least if combined with other policies support-ing industrial development. in the UnitedStates, on the other hand, quotas intended toprotect mature industriesnot only color TV,but automobiles or steelhave had ambiguouscuZcomes.

Could quantitative restrictions effectivelyshield other portions of the U.S. electronics in-dustry should imports surge as they have, say,in semiconductor RAMs (random access mem-ory circuits]? Probably not. Early in 1982,amidst consternation created by heavy importpenetration figures for 64K RAMs, Hitachi, Fu-jitsu, and NEC all announced accelerated time-tables for assembly in the United States. Thesemoves were clearly aimed at heading off for-mal complaints. If dumping or escape clauseproceedings had been instituted, the parallelswith color TV would probably have been rep-licated still further. As for color TVs, Japanesefirms already have enough volume in the U.S.IC market to attain the scale economies neededfor standardized products. In general, quotasare not a promising route to improved competi-tiveness for high-technology American in-dustries like electronics.

The Escape Clause

As mentioned above, GATT permits govern-ments to come to the aid of domestic industriesthreatened by imports. But before protectioncan be extended under the escape clause pro-vision in section 201 of the 1974 Trade Act, ITCmust return a finding that "an article is beingimported into the United States in such in-creased quantities as to be a substantiai causeof serious injury, or the threat thereof, to thedomestic industry producing an article like ordirectly competitive with the imported arti-cle."37 Fairness or unfairness is not part of thetext. The rationale is to pi ovide a time-inter-vat during which the threatened, industry andits workers can adjust to the (new) competitive


Revisions to U.S. law in the 1974 Act made'it considerably easier for an industry todemonstrate injury and thus qualify for protec-tion. As noted above, relaxation of the provi-sionAhat relief be contingent on a rise in im-ports stemming from tariff concessions orother forms of trade liberalization by theUnited States was instrumental in the color TVaction, Furthermore, previous incarnations ofthe escape clause. required that increased im-ports be a major cause of injury. The 1974 ver-sion changed the adjective to substantial, de-fined as "important and not less than any othercause." This standard is considerably weaker,and all vise equal makes it easier for belea-guered industries to secure protection.38

Other than the color TV case, only one suc-cessful escape clause action involving elec-tronics products has been advanced since thepassage of the 1974 Trade Act. This was filed

An late 1977 after a fourfold increase in importsof CB radius. ITC worded its findings strong-ly: " . . . serious injury is clearly immineni andthreatens the domestic industry with extinctionunless remedial action is taken to enable U.S.producers to compete on more equal priceterms."39 The President responded by raisingimpOrt duties .frOm 6 to 21 percent. After thefirst year, the duties decreased in increments,reverting to their original level at the end ofthe t rc kThe impact of this period oftari protection on the CB radio industry is dif-fi ult to judge, largely because sales dropped

recipitouslyfrom around 5 million units in1978 to only 2 million the next yearas the CB

.1.v. R. Cline. N. Kawanahe. T. 0. M. Kronsio. and T. Williams,Trade Negotiations in the 7'okyo Round: A Quantitative Assess-ment (Washington. D.C.:. Brookings Institution, 1978), p. 203.

"'U.S. Import WeeklyReference File (Washington. D.C.:Bureau of National Affairs. 1979), p. 58:0106.

fad tapered off. Nonetheless, imports capturedthe vast majority of 1979 sales.4°.Despite questionable effectiveness in past

cases in electronics, the escape clause remainsa tempting vehicle for portions of the U.S. in-dustry that find themselves hardssed by ship-ments from overseas. First and foremost, itdoes not require that imports be linked to un-fair behaviora condition that has oftenproved difficult to satisfy.' in dumping orcountervailing duty actions. Furthermore, in-jury can be defined in terms of narrow productcategories. The law requires only that injurybe demonstrated in "that portion or subdivi-sion of the producer which produces the likeor directly competitive article;" the market inwhich such injury cccurs can be limited to "amajor geographic area of the United States."'"The implications can be appreciated by recall-ing the typical competitive strategies of Japa-nese exporters. In both consumer electronicsand 5e`miconductors, exporters selected spe-cialized market niches where American man-ufacturers seemed vulnerable, the intent beingto gain a substantial market share within thisniche and then diversify. Thus Japanese semi-conductor manufacturers concentrated on 16KRAM chips, taking advantage of a shortfall inU.S. production capacity to quickly gain some40 percent of the Ameriban market. Under theprovisions of the Trade Act of 1974, an exportstrategy of this type could be subject to traderestraints.

40Electronic Market Data Book 1980 (Washington, D.C.: Elec-tronic Industries Association, 1980). p. 49.

"19 U.S.C. sec. 2251(b). Dumping and countervailing dutystatutes invite cdrnplaints on a narrow product line hasis'as well:during 198 ?, more than 120 separate investigations in carbonsteel products alone were undertaken by ITC.

Prospective Effects,olf LLF- Trade Policy on theElecf(onics Industry

-

To what extent, then, might the panopoly of the course of international competition, but


result of transferring responsibility for an-tidumping and countervailing duty provisionsfrom Treasury to the Department of Commercehas be& a more sympathetic hearing for Amer-ican busineisand as a consequence, the fil-ing of more complaints. Much also depends onthe complexion of ITC, which shifts as Presi-dentially appointed commissioners come andgo. Changes in the definition of injury havemad'rotection at least in principle easier toobtain; these too, in the ordinary course ofevents, serve to encourage demands for traderestrictions. Thus far in electronics, trade ac-tions have centered on consumer products; asJapanese manufacturers ,step up their pricecompetition in semicondtictors and computerequipment, there may be filings in these prod-uct categories. Furthermore, complaints by, theU.S. industry are increasingly centered on sub-sidies and other tools of national industrialpolicye.g., Japan's R&D programs. By andlarge, trade negotiations and the GATT haveproved unable to deal with such issues.

Certainly protectionist sentiment has beenrising over the past half-dozen years. Is a turn-ing point in the American attitude toward tradea real possibility? For some 50 years, the UnitedStates has taken the lead in international ne-gotiations to lower barriers to trade. From thefirst Reciprocal Trade Agreements Act in the1930's to the Tokyo Round MTN, U.S. policieshave supported liberalization as being in-theNation's long-term interests. Both major polit-ical parties have for the most part accepted theunderlying preMise of these policies: opentrade leads. to an efficient allocation of globalresources, as a result of which the Americanpeople will, more often than not, find them-selves better off. Other nations have benefitedas well. The export-based economic growth ofWest Germany and Japan owes much to theopenness of the large and affluent U.S. market,while American consumers have gained accessto a greater variety of products, as well as lowerprices resulting from foreign competition. Onthe other side of the ledger are the costs of dis-location and adjustment that follow upsurgesof imnnrte Moe,. mete tont] to rinrIct 1-Itanly;11,

out sharply, whereas the benefits are visiblemostly in the aggregate. Is it possible that theU.S. economy is too open, given changes in theinternational marketplace? Does the absenceof effective import controls in electronics makethe domestic industry overly vulnerable to in-roads by foreign industries?

Conbumer Electronics

International competition has generated con-tinuing pressures for trade protection in con-sumer electronics, yet the current situation ofthe U.S. industry is partly a consequence ofdomestic competition. Today, two American-owned firms, Zenith and RCA, account forroughly 40'percent of U.S. color TV salesasthey have, rather consistently, for many years.Although import penetration increased dra-matically during the 1970-77 period, the bruntof the sales losses were borne by other, manu-facturers. In a single year: Magnavox andMotorola each saw their domestic TV salesdrop by more than 15 percent. These Marketdeclines led rather directly to the sale of theirTV operationsto North American Philips andMatsushita, respectively. While only four do-mestically owned producers remained in 1.983(compared with 17 in 1970), they have beenjoined by. more than 10 foreign companiesmanufacturing or assembling sets here. Con-centration has not increased significantly andno one firmAmerican or Japanesehas comeclose to dominating the market. Policy deci-sions by the U.S. Government stimulated theinflux of foreign capital, although OMAs probeably influenced timing more than decisions toinvest. Foreign-owned plants in the UnitedStatestogether with continuing imports fromJapan and other Far Eastern nationscreatedrelentless pressures on American TV manufac-turers, even while quotas were in force. U.S.firms shifted production abroad to reducecosts, at the expense of jobs herebut con-sumers have benefited via low prices and high-quality products. Still, only the largest and


For other consumer goods, Federal policieshave rarely come into playarid seem unlike-,ly to, at least in a manner similar to events inthe TV industry. Far Eastern firms are todaythe unchallenged leaders in most other con-sumer electronic products; there is virtually noU.S. manufacturing left to protect when itcomes to portable radios, monochrorric TVs,stereo/high fidelity equipment, the simplerpocket calculators, or electronic watches. Moreimportant for the future, the United States haslagged in the development of new productssuch as video cassette recorders (VCRs). Atevery step, Japanese manufacturers lead inproduct development or are working in parallelor in cooperation with American firms, theprimary exceptions being electronic toys andgames and home computers. If products like-video disk players achieve mass-market suc-cess, the Japanese will be early and formidablecompetitors.

Under such circumstances, protection forAmerican manufacturers would have to comethrough legal provisions that have not in thepast been exercised. When applying the injurystandard in antidumping proceedings, ITC ex-amines whether imports are harming a domes-tic industry, are likely to damage it, or arepreventing such an industry from being estab-lished. The last of the three possibilities hasseldom been relevant because existing indus-tries have normally sought dumping investiga-tions. But in principle, the clause could be abasis for reliefif imports were priced at lessthan fair valuefor products that are not evenbeing made in the United States, such as VCRs.

On the other hand, the escape clause injurystandard would have to be considerablystretched. Here, there are only two possibilities:an existing industry must be seriously injuredor threatened. Given product leadership overseas, with imports achieving a sizable marketshare from the outset, serious injury to an ex-isting industry probably could not be demon-strated (assuming new types of products) un-less the standard was applied in a novel andunintended way. Given the general ineffective-

isting U.S. trade policies could shield domesticfirms producing the consumer electronic prod-ucts of the foreseeable future. Indeed, costpressures -will probably- continue_to drive agood deal of production by American entrantsoffshore.

SemiconductorsUnlike consumer electronics, where the ef-

forts of U.S. manufacturers have been largelyconfined to the domestic market, productionand sale of semiconductors is carried out ona global basis by the major U.S. merchantfirms, as well as by Japanese andto a farlesser extentEuropean producers. U.S. lead-ership has meant that domestic manufacturershave not, as yet, sought direct Government as-sistance in combating imports. For many years,sales expanded rapidly; American supplierswere often hard pressed just to keep pace. Thenumber of domestic companies serving themarket tripled during the 1960's, while importswere until recently almost entirely the- ,inter-divisional, shipments of U.S. multinationals(ch. 4).

The picture began to change at the end of the1970's. Aggressive competitive tactics and mar-ket successes by Japanese firms have had im-pacts in many parts of the world, but from theperspective of U.S. policymakers the domesticmarket has been the focus. Japanese companiesare providing the first real competition fromabroad in the experience of most of the Amer-ican industrycompetition that has driventhem to seek the attention of the Federal Gov-ernment. The most publicized examples.of Jap-anese inroads have been the 16K and 64KRAMs sold in large numbers to manufacturersof computers and microprocessor-based sys-tems. In. 1980, Japanese manufacturers cap-tured about 40 percent of the U.S. andworldmarket for 16K chips, partly because of inade-quate capacity in American plants. By 1982,the Japanese share of next-generation 64KRAM sales was running at about 70 percent.More than any other event, the rapid inroads


The tempest over 641 RAMs in the springof 1982 may prefigure future trade disputes inthe high-technology products of this and other

industries. As publicity over inroadsby Japanese imports, the Departments of Com-merce and Defense began examining the im-plications for national security. At the time,only Texas Instruments and Motorola amongU.S. merchant firms were producing 64K chipsin quantity. Prices had been dropping rapid-ly, driven not only by declining manufactur-ing costs, but by recessionary pressures leadingto price cutting in the Japanese market as wellas here: Worldwide production capacity for64K RAMs may have exceeded demand for atime.

When the Semiconductor Industry Associa-tionand Motorola specificallyaccused theJapanese of dumping, though without filingcomplaints, the Japanese responded by an-nouncing plans to move some production tothe United States. Meanwhile, the Commerce/Defense study had begun, evidently at the in-stigation of the latter agency. Among the possi-ble outcomes of the Commerce/Defense study,three appeared at the time to be among themost likely:

1. A dumping complaint against the Jap-anese, self-ir; Hated by Commerce.

2. A section 232 investigation, based on thenational security implications of U.S. de-pendence on Japanese ICs.

3. A complaint through GATT, probably con-cerning issues of reciprocal market access.

The section 232 alternative is noteworthy forillustrating the variety of instruments thatgovernments can bring to bear in trade-relatedmaiters. Part of the*Trade Expansion Act of1962, this rarely used statute permits the Presi-dent to limit importse.g., by tariffs or quotaswhere such shipments "threaten to impairthe national security." No section 232 pro-ceeding was started in this case, but a recentinvestigation of ferroalloy imports by the Com-

vision by industries suffering from foreigncompetition.42 Nor were antidumping proceed-ings initiated for 64K RAMs. Later in the year,growing-demand caused-prices to firm,-defus-ing allegations of dumping and redirecting lob-bying by domestic semiconductor firms towardreciprocity legislation. Nonetheless, followingthis turn of events, Japan got another unpleas-ant surprise: the Justice Department began aninvestigation of six Japanese semiconductormanufacturers, premised not on price-cuttingbut on restricting shipments to the UnitedStates in order to raise prices.43 To con-siderable extent. such episodes illustrate the in-ability of the traditional tools of trade policyto deal with events in a fast-moving, techno-logically based industry like microelectronics;they also illustrate the multiplicity of actorspopulating U.S. trade policy and enforce-menta multiplicity that some would charac-terize as leading to confusion and disarray."

In any case, the U.S. merchant semiconduc-tor industry has not thus far sought direct pro-tection. Rather, American firms and their tradeassociation(s) have continued pointing to fea-tures of Japanese industrial policies and busi-ness practices they feel are unfair, urging theU.S. Government to exert pressures aimed atending them.45 In addition, industry executiveshave sought Federal actions that would im-prove their own ability to competelobbyingin favor of R&D tax incentives and reductionsin capital gains taxes, as well as calling atten-tion to engineering manpower shortages. Many

42See "Specialty Steel Industry Attacks Draft Report by Com-merce on Ferroalloy Study," U.S. Import Weekly, July 21, 1982,p. 478. Ferroalloys are used in making steels. The investigationwas requested by a trade association of U.S. suppliers. Past ap-plications of sec. 232 have been restricted to petroleum imports.

43A. Pollack, "Inquiry Puzzles Chip Makers," New York Times.July 7, 1982, p. D9.

""Two dozen or more Federal agencies exercise some degreeof responsibility over foreign trade and investment policies"Opening Statement of Senator Roth." Government Organiza-tion for Trade, hearing, Committee on Governmental Affairs,U.S. Senate, June 4, 1981, p. 2.


of the tax changes advocated by semiconduc-tor firms were in fact. implemented by the Eco-nomic Recovery Tax Act of 1981 (ch. 7). Indus-

-try-leaders have also-asked-the-Federal Govern-ment to negotiate for easier access to the Jap-anese market, claiming that sales in Japan arevirtually impossible except for products thatlocal companies do not make. The U.S. indus-try's position is that asymmetries vis a vis Japanresult in a competition in which the two sidesare playing by different rules, with the advan-tage to the Japanese.

As in earlier export thrusts, Japan's semicon-ductor shipments have been concentrated ina few product types. The choice has been mem-ory circuitsparticularly dynamic MOS RAMsin part because of the close coupling betweenJapanese efforts in microelectronics and com-puters. Worldwide, Japan's RAM sales have in-creased more rapidly than those of Americanfirms, with European manufacturers the biglosers. At the end of 1979, just four compa-niestwo Japanese, two Americanproducednearly two-thirds of the total world merchantoutput of 16K RAMs. Two years later, a pairof U.S. firms confronted six Japan producersin the baffle for worldwide market share in 64KRAMs; while other American, manufacturerswere ramping up production or preparing toenter with their own designs, Japanese com-panies were investing heavily in additional pro-duction capacity.

Still, except for Mostekwhich is rapidlydiversifying its product line--RAMs have notdominated the sales mix of any American com-pany. Thus, while Japanese incursions havehad drastic impacts on the RAM marketaf-fecting prices and profits, as well as marketsharessimilar shocks have not yet been feltin other products. A major concern of U.S. pro-ducers is that this experience will be repeatedelsewhere, denying them the learning and scaleeconomies so important for competitiveness,and cutting into the profits they need to gen-erate cash for expansion. Moreover, MITI-sponsored R&D efforts like the VLSI project are

faced in gaining access to patents and othertechnical resultsis viewed as further evidenceof asymmetries favoring the Japanese. In sum,USsemiconductor_manufacturers believe thata closed market shelters Japanese competitors,leading to economies of large-scale productionthat translate into low prices. Investments inproduction facilities within Japan are one waythe U.S. industry sees to counter the threat.

As the discussion above implies, trade out-comes depend on complex sets of competitiverelationships, Consider, as an example, thequestion of scale economiesa matter morecomplicated than sometimes implied. A por-tion of the cost savings, associated with produc-tion scale come via learning curve effects; hewho gains an early edge in market share en-joys lower costsperhaps permanently. TheJapanese, in the simplest view, "learn" by pro-ducing for the domestic market, then penetrateforeign markets based on low costs and lowprices. But learning economies are not quiteso straightforward. Some cost reductions arefunctions of cumulative production volume;others depend on time as wel1.46 In the lattercase, obtaining a large early market sharewould not confer the same cost advantages.The extent to which market penetration resultsin lower costs, therefore, may be product-specific; for some types of ICs, cumulativevolume might matter much more than forothers. The limited evidence available suggeststhat costs for logic circuits depend more heavi- -

ly on time, memory costs more on scale.47 Asa consequence, the advantages of access to theU.S. market could be considerably greater forproducts such as RAMs than for at least someother types of ICsperhaps one reason Japan'sexports have been so heavily weighted towardmemory devices (the comparatively straightfor-

4ea W. Webbink, The Semiconductor Industry: A Survey ofStructure. Conduct and Performance (Washington, D.C.: FederalTrade Commission. Bureau of Economics, January 1977), pp.49ff.

" "Management Committee Report to the Management ReviewCommittee C.n Assessment Rennrt. IBM. October 25. 1971 f PX


ward design of RAM and other memory cir-cuits is also a major factor).

Even so, patterns linking design and develop-ment, manufacturing, and marketing tend torepeat fairly consistently from product to prod-uct (see app. C on the 4K RAM). Technologydominates production and marketing strategiesin .the early phases, because the firSt to offera new chip design may-possess a near-monop-oly--whether the chip is an innovative productor simply an incremental advance. After intro-duction, prices decline slowly as firms recoupR&D costs by selling to customers willing topay premium prices for leading edge devices.Eventually, other manufacturers enter the mar-ket, forcing prices downsometimes to a pointbelow production costs. The final phase in theproduct cycle finds prices low, with many ofthe earlier participants unable or unwilling tocompete; as a consequence, prices may evenbegin to rise once more.

The competitive dynamics of the industry re-volve around such factors. Some companies at-tempt to be leaders, bringing innovative prod-ucts to market early and capitalizing on theliigher prices they command; in the UnitedStates, Intel has become known for this strat-egy, which depends on heavy expenditures fordesign and developmentas well as abandon-ing products when they, begin to mature andmargins fall. Other firms manufacture a diver-sified line of more mature devices,' concen-trating on process technologies as a ,route tolow costs. For such companiesNational Sem-iconductor has been an example---Lmarket pen-etration is vital; as a result, they are particularly.vulnerable to import strategies that also empha-size market position. Still other entrants par=ticipate largely as a byproduct of internal op-erations. Semiconductors may account for onlya small fraction of their business, but if theyuse substantial quantities in their other .endproductsas do Japanese manufacturers suchas Fujitsu or Hitachithey may be able to sell

The vulnerability of the U.S. semiconductorindustry to foreign competition is therefore afunction first and foremost of technology.American firms with the ability to be consist-ently early to market with 1,,,w products havegenerally had less to worry about. Thus far,most imports have been standard circuitsasituation that could certainly change if semi-conductor manufacturers in Japan or else-where begin to design more innovative devices.At the same time, there are real limits to a tech-nology-based strategy. Incremental payoffsfrom R&D may diminish over time. Althoughnew types of microelectronics products couldstill open new mass markets, signs of tech-nological slowdown have begun to appear. In-evitably, the industry will mature, with greatercompetition from imports a predictable conse-quence: slower rates of technological changemake it easier for foreign firms to catch up andkeep up.

Industry structure is also important. Thehigh-volume merchant manufacturers in theUnited States have coexisted with a fairlt1 largenumber of small firms for many years, the lat-ter typically specialists filling market niches ofless interest to bigger companies. Structuralchanges are underway in the domestic indus-try, partly in consequence of heightened inter-national competition, partly because the capitalrequirements of advanced circuits make pur-suit of VLSI difficult for small companies. Thechanges are of two types, as discussed in pre-vious chapters. First, diversified American cor-porations are purchasing or merging withformerly independent, semiconductor firms.Second, foreign enterprises are continuing totake ownership positions in U.S. manufactur-ers. Such marriages have occurred, on the onehand, because roreign electronics manufac-turers want quick access to evolving tech-nologies, and on the other hand, because theU.S. partners have needed infusions of capital.

How will these structural shifts affect com-petition? As American semiconductor manu-facturers become larger and more diversified

invirnr, .7,11 ,,,,Tre,r4111

/456 International Competitiveness in Electronics

1

lines of business. Such) companies have theflexibility to shift resources internally, to meetthe competition on a prirle basis, to invest moreheavily in R&D and in 0.oduction equipment.To, the extent that American semiconductorfirMs have had difficulty in financing expan-sion, with possibly harmful effects for U.S.

j6irnpetitiveness, consolidation should help./Multinationals also tend to have 'free trade"/Perspectives because they depend on doing

r ! business overseas, and are less likely to press/ trade complaints. On the other hand, some

observers believe large firms to be less en-trepreneurial and more cautious, and that con-solidation will reduce the prObability of innova-tion, making technical leadership more diffi-cult to maintain. While small firms will con-tinue to existmany new ones have beenstarted recently structural shifts of the typevisible in semiconductors have charaCterizedthe maturation of many industries; rather than

-'llifocusing on the pposed virtues and liabilitiesof small and lar e firms, it is perhaps more per-tinent simply to observe that they have dif-ferent strengths .and weaknesses in a givencompetitive context.

More to the point in terms of trade policy,what are the implications of greater numbersof foreign manufacturers with active semicon-ductor design and production facilities in theUnited States? This tendencyTarticularly evi-dent in recent investments by Japanese firmsto mitigate trade frictionscombined with uni-versal fOreign involvement by the larger Amer-ican p oducer§, has made semiconductor man-ufact re' one of the more international of theworl/rs industries. Attitudes toward trade and/investment shaped by the traditional concernsof domestic firms and their workers may notfit the realities of such an industry. In micro-eleictronics and computers, the notion of "our"firms versus "theirs" is an oversimplificationwhen so many companies operate on a globalscale. Corporations engaged in bitter trade dis-putes in one part of the world may establishjoint ventures elsewhere; cross-licensing is the

r;vrvIe r-Irsrsaofn virifb rmo nnnthar utrhorrs

tors have no doubt contributed to the reluc-tance of American semiconductor manufactur-ers to press formal trade complaints againstJapanese exporters..

The old wayserecting trade barriers toshield domestic industriescan damage U.S.-based companies quite aside from any possi-bility of retaliation by foreign governments.Texas Instruments produces 64K RAMs in Ja-pan for export to the United States. In con-sumer-electronics, the OMA on color TVs fromTaiwan restricted shipments by RCA and Ze-nith, both of which had substantial investmentsthere. Japanese semiconductor manufacturerswould quickly shift production to export plat-formsin many cases, the t,,..1mo offshore sitesfavored by American firmsin the event of re-strictions on shipments of ICs from Japan.They would also move more production here.When those affected include U.S.-based mul-tinationals along with foreign firms havingsignificant interests in the United States, tradi-tional protective measures become let-, prac-tical. Such dilemmas have arisen, or are likelyto, in other industries as wellautomobiles,computers, possibly aircraft, chemicals,energy, pharma. .uticals. Increasingly, firmsin such industries are tied by a multitude ofco-production and joint venture agreements,irrespective of, the locations of their head-quarters. As a consequence, the "inside-out-side" or "good guy-bad guy" distinction be-comes a difficult one for policymakers to draw;in such a world, trade policies directed to anolder order may simply be overrun by events.

In any event, the American semiconductorindustry has not attempted direct action tostaunch the flow of imports, much less with-drawn to a protected position in the UnitedStates. Instead, while continuing to lobby theFederal Government, U.S. merchant firms havemoved boldly to maintain their competitive-ness worldwide. This is one reason the in-dustry's leaders have been more vociferous

. over what they view as unfair domestic sub-sdies by the Japanese Governmentand overimnorlimorOc to their num AtiPMT1tq to qP11 nr

From the viewpc5int of American companiesthat purchase semiconductors, the outcomesof intensified competition have been beneficial: /a wide range of product offerings, low prices,',,high quality. Would. these benefits have fol-lowed even if U.S. producers had chosen strat-

i..egies or trade protection? This`must remain. anopen questionbut to the extent that other in-dustries offer parallels, the benefits would/have(heen smaller and slower to arrive

As the technological leads of U.S. mieroelec-/ironies firms narrow, and competition con-( tinues to mount, the industry's support for( open trading relationships may diminish. Yet( American semiconductor produCers cannot( back too far away from a free/trade stancewithout jeopardizing their own overseas in-terests. Formulating equitable/ trade policieswill continue to be difficult, with a wide rangeof interests to be balanced. A/large fraction of


U.S. exports and imports' of semiconductoirproducts will continue to be transfers betweendivisions of multinational corporations. Linksamong U.S. and foreign firms will certainlypersist and may well strengthen. Trade poliCiesdealing with dumping or countervailing dutiesare unlikely tu be very relevant, if only beCauseof the pace of technological changewhichcan render :he products in question obSoletebefore the proceedings have run their course.

Computers

The picture is similar in the computer in-dustry. U.S. firms have been undisputed lead-ers, with subsidiaries engaged in manufactur-ing and marketing around the world. Importshave been at low levels, even for personal com-puters and peripheralsalthough this could

.4.mst=.....P.41NeSSESSE121E1

458 " International Competitiveness in Electronics

certainly change. The consequence has been,until quite recently, a virtual absence of con-cern by executives of American computerfirms tvitir U.S. trade policy except as'it sup-ports an open international trading environ-ment or affects transfers of components andsubassemblies among subsidiaries. Americanmanufacturers wish to see items 806.30 and807.00 of the Tariff Schedules -preserved, butdumping or escape clause provisions have sel-dom attracted their attention..

Yet here too the competitive picture is chang-ing. Foreign governments, viewing informationprocessing as vital to national interests, havefound a variety of methods for subsidizing localfirms, as well as an array of carrots and sticksfor encouraging American corporations totransfer technology to local computer man-ufacturers.48 And again as in semiconductors,foreign manufacturers have taken equity posi-tions in American computer companies, in partto acquire technical knowledge. The technol-ogy..gap between U.S. and foreign firms hasdiminished in both hardware and software,with several Japanese manufacturers beginningto ship mainframe machines to the UnitedStates either directly or in partnership withAmerican firms.

What impact will such developments haveon trade flows and on U.S. trade policy? Giventhat competition depends on much more thanfast, reliable hardware. it is too early to makepredictions. Manufacturers must be closely at-tuned to user needs; foreign computer firmslag well behind American companies in theirability to seek out and satisfy customer applica-tions (ch. 5). This deficiency has not gone un-recognized; other nations are devoting substan-tial efforts to software, often aided by govern-ment subsidies. Furthermore, countries likeBritain. have always been good at software andmay provide a resource that firms elsewherecan bp. The U.S. lead in software seems bound

"Overseas governments have used investment inceniiycs toattract AllWriCall fitans more actively in electronics and elec.

to narrow, following that in hardware. The flu-idity of market structures emphasized in chap-ter 5 Will leave room for aggressive foreigncompetitors.

As a result of the subsidies for computer tech-nology that virtually all industrialize f7ountrieshave employed, it is not hard to envision a sce-nario in which these subsidiesas well as thepreferential treatment many governments haveextended to local producersbecome the tar-gets of countervailing duty complaints. If afOreign firm benefiting from government. lar-gess were to establish a significant market posi-tion in the United States, can there be muchdoubt that American manufacturers wouldseek remedies under U.S. law? After all, thesubsidies extended to foreign computer indus-triesalbeit often rationalized on nationalsecurity groundshave been even more visi-ble than in microelectronics. What, then, mightbe an appropriate response on the part of theU.S. Government?

The issue,raisedand repeated in semi-conductors, communications equipment, andother high-technology products--is that the ex-isting structure of national and internationaltrade laws andagreements evolved.in anotherera; it was not designed with current varietiesof national industrial policies and subsidies inmind. Countervailing duties were intended tooffset export subsidies,,such as rebates or otherpayments contingent on sales to overseas cus-tomers. Subsidized financing via export-importbanks has strained the system. Domestic sub -.sidies with indirect effects on exports scarce-ly fit it. Indeed, U.S.-based companies maybeamong those benefiting from industrial policies in other nations. Antidumping laws weredrafted to counter explicit price discriminationby foreign monopolists, often involving govern-ments and/or cartels that encouraged exportsby charging higher prices to domestic than toforeign customers. Until recently, antidump-ing legislation.was seldom called on whereprice-cutting was extended to all customers,domestic as well as foreign. Today, when


Boeing? Or, more subtly, is U.S. support of re-search into solid-state electronics as part ofmilitary and space programs unfair---researchthat, after privately funded follow-ons, even-tually results in commercial applications?Many other examples, in any number of coun-tries, could be cited.

For computers or communications, govern-ments`have seldom proffered financial assist-nce simply to foster exports, although this has

been one motivea strong one in Japan. Rath-er, governments have targeted.the informationindustry as vital to a multiplicity of nationalinterests. Subsidies have been generalized,directed at industrial development over the

longer term. The question for trade policybecomes: As governments increase their in-volvement in economic affairsand indeed inthe actual operation of business enterprises=what types of trade policies and agreementswill be needed so that participants can agreethat the terms of competition are reasonablyfair? This will remain a central matterk for in-ternational trade negotiations over the currentdecade and beyond. While the Tokyo Roundtrade negotiations addressed such questions,the substantive chang:.. in procedures embod-ied 2n the new subsidies code are small, andunlikely to have much effect.

Summary and ConclusionsU.S. trade policy has been rather consistently

oriented toward open international trade andinvestment over the last half-century. In thepostwar period especially, the United Statestook the lead in eliminating both tariff and non-tariff barriersby reducing its own restrictivemeasures and pressing its trading partners todo the same. Some parts of the electronics in-dustry, notably manufacturers of consumergoods like TVs and CB radios, have sufferedas a result. But if import competition has hurtthe manufacturers of such products, other sec-tors of the U.S. economy benefited from free-dorn to export and invest overseas. U.S. tradepolicy has helped American semiconductorand computer firms become leaders in marketsall over the world. The exception has beenJapan; barriers imposed by European nationshave proved far less substantial. On the whole,the open trading environment resulting fromsuccessive rounds of multilateral negotiationshas helped the competitiveness of the U.S. elec-tronics industry.

More narrowly, trade policy impacts in con-sumer electronics have centered on longstand-

by American firms against competitors in theFar East, primarily Japan. The response of theFederal Government has been marked by de-lays and interagency conflicts. Fifteen yearsafter the initial antidumping actions, the situa-tion remains unresolved, duties uncollected.The uncertainty created by this long and con-voluted history has made life difficult for bothdomestic firms and importers. To considerableextent, as the shape of the industry has altered,

complaints over trading practices have becomemoot. Orderly Marketing Agreementsnegoti-ated as an upshot of escape clause proceedingsunrelated to unfair trade practicesacceler-ated what would probably have been wide-spread eventual movement by foreign firms to-ward assembly here. Plants owned by foreigninterests have replaced failing domestic TVmanufacturers. Meanwhile, the remaining U.S..producers have moved some of their assemblyoperations to low-wage offshore locations,helped by provisions of items 806.30 and 807.00of the U.S. Tariff Schedules. These provisionswhich allow tariffs on re-imports to be com-puted only on foreign value addedhave beena target of labor interests. But if in some cases


lent to the export of jobs, in others transfersoffshore have been necessary to retain anydomestic jobs.

As tariff walls in many parts of the worldhave_ slowly come down, the attention of bothprivate firms and governments has turned toindirect and nontariff barriers. Ranging fromimplicit subsidies for domestic firms to un-cooperative customs inspectors, such barrierspose much more complex subjects for interna-tional negotiationsas the Tokyo Round dem-onstrated. Some progress. has been made, butit is too early to tell how trade in electronicsmay be affected over the longer term.

Among nontariff and indirect measures, sub-sidies for economic and industrial develop-ment ostensibly aimed at domestic objectivesare perhaps the most difficult case, along withgovernment pi ocurement. As discussed in theprevious chapter, many nations have used bothtypes of measures consistently and aggressivelyas elements of industrial policyespecially inelectronics. While progress has begun in open-ing up government procurements, subsidieswill remain thorny issues for years. More andmore governments, for example, are resortingto R&D incentives to support local electronicsmanufacturers; inevitably, these function tosome extent as export subsidies, even if this is

not the primary er avowed intent. While directimpacts on international trade tend to be small,the visibility of programs such as Japan's jointR&D efforts, or West Germany's spending oncomputer technology, draws frequent attacksby businessmen and political leaders in otherparts of the world. The nations mounting theseprograms consider them vital for economic de-velopment; they will not disappear. As hasbeen the case with complaints over unfair trade;practices in consumer electronics, negotiationsconcerning indirect supports and subsidies arebeing overtaken by events; subsidies may insome respects function like other nontariff andindirect barriers to trade, but governmentsseldom institute them for such purposes. Nordo they view them as elements of trade policy.They are seen as vital tools of industrial pol-icypolicies developed in response to an eco-nomic environment in which domestic and in-ternational dimensions can seldom be isolated.

Negotiations aimed at reducing such subsi-dies make slow progress at bestindeed,although necessary, they may finally be ratherbeside the point. So long as governments re-gard high-technology industries like electronicsas essential to industrial development and eco-nomic growth, supports and subsidies seemmore likely to, increase than decrease.

CHAPTER 12

Federal Policies AffectingElectronics: Options for

the United States

Contents

( hie i'Vii!W 463

The Current Policy Environment 464

Alternative Perspectives on Industrial Policies 4671.;usurt! dit llomestic Market Base for 11.S Industries 468Support for rritical Industries 475.lolr&ctructural irod Adjustment Policies 485Promoting the Clohal Competitiveness of -American Industries 491

Inthistrial Policy Centered in the Private Sector 496

SummarV and Conclusions 500

List of Tables

Nle,rwres Likely to be Emphasized roofer Alternative Apploaclms1,1 Industrial PolitN 468

CHAPTER 12

Federal Policies Affecting Eiectronics:

Options for the United States

OverviewIn 1982 electronics accounted for 10 percent

of U.S. merchandise exports; manufacturers ofelectronic uroducts contributed 4 percent ofU.S. gross domestic product and 9 percent oftotal goods output; the industry employed morethan a million and a half AmeriCans. Beyondthis, the electronics industry produces goodsand knowledge that are vital for much of therest of the economy and for society as a whole.Banks, insurance companies, and many otherservice sector enterprises could hardly conducttheir businesses without electronic data proc-essing. Entertainment industries like televisionbroadcasting developed in conjunction withthe manufacture of consumer electronic prod-ucts. Before too long, prosthetic devices for thehandicapped will routinely be built around"smart" microelectronic devices. Modern com-mercial aircraft depend as heavily on radar,computerized flight control systems, and-elec-tronic navigation aids as do military planes.Over the rest of the century, computer-aidedmanufacturingshop floor management sys-tems, automatic warehouses, smart robotswill helo increase productivity and redticecosts throughout the Nation's manufacturingsector. If there is any single industry whosetechnological progress and competitiveness arecritical to the economic growth and nationalsecurity of the Uni;_ed States, it is electronics.

Yet even the most dynamic portionS of theU.S.. electronics industrysemiconductor pro-duction, where innovation is a way of life, com-puters, where markets seem to expand nearlyas rapidly in bad times as good find them-selves increasingly challenged by foreignmanufacturers, both here and overseas.Although few European firms have managed

mined efforts, strongly supported by nationalgovernments, continue in countries like Franceand 'West Germany. Japan has moved swiftlyfrom a position as technological laggard to be-ing one of the leadersnot only in product de-velopment, but in the fundamentals of elec-tronics technology. The strengths of the'`,,,panese industry have been described inprevious chapters: an ample supply of skilledand motivated employees; managements thatapproach markets on a global scale and havelearned to do business effectively in countriesranging from Saudi Arabia to the United States;an economic system in which the tradeoffs be-tween competitive rivalry and cooperation aim-ed at advancing common goals are well-man-aged; a government whose industrial policiesconsistently support and encourage the private

`sector. In the language of sports, the Japaneseelectronics industry has momentum. Else-where in Asia, developing countries can al-ready make many consumer electronic prod.ucts and compopents at less cost than japanor the United States; these countries will con-tinue to move into more sophisticated goodsthough at first continuing tVocus on consumermarketssupported by export-oriented indus-trial policies. Hong Kong, for one, has alreadymade the transition froin discrete transistorsto integrated circuits (ICS")computer chips asweld as thos'e for consumer products; This isriot to say the Ainerican electronics industryrisks overnight decline. It does mean that com-petition will be difficult in the years ahead, andneither the technological leads that the UnitedStates still maintains nor the size and affluenceof our domestic thorket will suffice to guaran-tee American primacy.

464 International Cornpettlivoness in Electronics

might choose to move toward a more explicitindustrial policy of its own. If other countriesarea adopting policies in support of industrieslike electronicsand having some success, or,more tellingly,-mOving down learning curvesleading to more consistently productive ef-fortsthen it makes sense to ask whether suchan approach could also work here. Such ques-tions need to be addressed in terms of concretepolicy objectives and prospective mechanismsfor achieving -themmatters to whit it thischapter is devote 1.

The United States could afford to live with-out a consciously formulated industrial policyin years past, but the realities of internationaleconomic competition -changed--dramatically--during the 1970's. The "information society"is here; industries are no longer defined by na-tional new jobs are opening forAmericans who are computer-literate at a time

r when employment of other types is in decline.' If new approaches to industrial policy are to

be considered, what is the range of possi-bilities?

After a brief review of the ad hoc nature ofpast U.S. industrial policies, the chapter out-lines five options for a more focused approach:

protection of domestic markets;a "critical industries" policy; ,

an orientation toward infrastructural sunportprimarily for technological develop-mentand adjustment;promotion of competitive U.S. firms on aworldwide basis; andFederal withdrawal, where possible, in fa-vor. of the private secto7.

The five are intended to span the realistic alter-natives; they overlap Somewhatin severalcases specific policy measures would be simi-lar if not identical. Still, if the five optionsthemselves are not exclusive, they are distinct.Each is discussed in terms of prospective ef-fects on electronicsand, by extension, otherhigh-thchnology-sectors:- --

The options are discussed in terms of direc-tions and objectives; they are intended to of-fer a set of alternative signposts. All start fromthe same point: the patterns of worldwide com-petition outlined in the preceding chapters; theincreasing capital7intensity of critical sectorsof the electronics industry; its continued de-pendence on research and development (R&D);needs for skilled labor and imaginative man-agement. But each alternative implies a dif-ferent ro'ite to a different destination.

The Current Policy EnvironmentChapters 10 and 11 summarized U.S. indus-

trial and trade policies as they affect competi-tiveness in electronics, contrasting them withpolicy approaches - abroad. Federal policieshave been notably ad hoc, formulated and im-plemented by many different agencies, no oneof which has overriding authority. Trade poli-cies are neither very predictable nor closelylinked with domestic industrial and economicpolicies.

,Outside the trade arena, regulation of broad-casting and telecommunications has been acontinuing influence on consumer electronics.Among the recent issues with potential impactsan U.S. Manufacturers, either direr,t or in-

.direct, are: the rights of owners of video-tap-

ing equipment to copy off the air; the fate ofAM stereo broadcasting (where the FederalCommunications Commission has avoided de-cisions on a standard system); regulation ofhoir:e information services and data commu-nications (videotext and teletext may be slowin corning, but will eventually be integratedinto home entertainment and information sys-tems). As with regulation of cable TV andrights to satellite transmissions, such mattersmay have only indirect consequences for man-ufacturers of consumer electronics products,but impacts that are no less real for this.

The U.S. semiconductor and computer in-dustries benefited in their early years fromGovernment-funded R&D and procurements;

O

Ch. 12 Fedeial PohcieS Affecting Electronics: Options for the United States 465

military and space programs drove develop-ments in microelectronics until the 1960's,while both civilian and military sides of theGovernment have been heavy purchasers ofcomputers and related equipment. AlthoughFederal R&D is still significant for both micro-electronics and computers, particularly interms of more basic researchand in the caseof the Defense Department's Very High-SpeedIntegrated Circuit (VHSIC) program, processtechnology and applied research as wellit isnow ,,decidedly secondary to the industries'own spending. Furthermore, the decliningfraction of total markets t'or both semiconduc-tors and computers accounted for by Govern-ment has made Federal procurement a muchweaker force than 10 or. 1:3 ago.

Indirectly, a wide variety of policiesaffect U.S. semiconductor and,computer firms:support for education and-training, particularlyof technical professionals; regulation of datacommunications; the 1982 antitrust settlementswith IBM and AT&T; copyright and/or' patentprotection, for chip designs and software; taxpoliciu as they influence competition for fundswithin U.S. capital markets. On the other hand,Government policies directed specifically atthe semiconductor or computer industries-

-either_in_thedomestic context or in terms oftheil'intPr -.Itionarcompetitive positionsareremarkauid by their absence, the vacuum astriking contrast with ambitious, comprehen-sive, and supportive (if not necessarily verycost-effective) ptiblic policies in other coun-tries. A major exception to the absence of poli-cy is antitrust.

American electronics firms have often com-plained that antitrust enforcement hinderscooperative R &D and joint ventures in inter-national trade- which, if allowed, wouldstrengthen this country's competitive position.The Department of Justiceand, to the extentthat it is involved in antitrust enforcement, theFederal Trade Commission (FTC)reply thatsuch cooperative activities are, in fact, general-ly permitted. The dialog, which has been go-ing on for years, is part of the prohloi-7,1: Justiceand the FTC have helped create ..( psychologi-

cal climate in which industry is reluctant to testthe bounds of the permissible. The two agen-cies generally act as if this status quo is desirable, with their public pronouncements nice-ly straddling the relevant issues.' While theforcement attitudes of the Reagan admini5.tration differ somewhat from those of its im-mediate predecessors, the more relevant e;on-cernat least from the viewpoint of Congressmight be: Do the antitrust statutes need recon-sideration in light of changes in the characterof international trade and competition sincethe Clayton and Sherman acts were passed inthe early decades of the century? Where wouldantitrust fit within a more coherent U.S. in-dustrial policy?

Beyond the intangible effects of Federal an-titrust_enforcement--beginning with their-forcein restraining clearly undesirable forms of an-ticompetitive behaviorchapter 10 mentionedthe two recent and major antitrust actions di-rectly involving the electronics industrysuits

it IBM and AT&T, both recently settled.In both cases, critics of U.S. antitrust enforce-ment had claimed, with some validity, that theGovernment was trying in the name of compe-tition to break up enterprises that were main-gtays of U.S. competitiveness.

Regardless of possible rewrites of communi-cations legislation by Congress, the AT&T set-tlement will change the form and function ofBell Laboratories. A weakening of the basic re-search foundation that Bell Laboratories helped

'Consider this quotation from the "Statement of William F.Baxter, Assistant Attorney General, Antitrust Division, BeforeThe Subcommittee on Employment and Productivity, Commit-tee on Labor and Hunan Resources, United States Senate, Con-cerning Antitrust Poiicy and Productivity, Apr. 16, 1982," pp..12-13. "Joint ventures may foster efficiencies riot available toindividual firms, and may promote technological progress andenhance productiiity. Such joint ventures should not be deterred .

by rigid or overly-broad applications of the antitrust laws. TheDepartment's recently published Guide Concerning ResearchJoint Ventures is intended to assist businesses considering jointventures by clarifying our enforCement policies in this regard,For example, as the Guide indicates, an important factor iswhether a joint venture leaves a significant number .of non-participating firms free to engage independently in research. Ifthere are not a significant number of such non-participatingfirms, and the joint venture's research could be done individuallyby the participating firms, antitrust problems could arise:. Muchof the lore of antitrust resides in such pronouncement's.

463


lay for so much of the U.S.and world--elec-tronic- aiustry seems inevitable. The reasonis straiAinforward: Western Electricthe pa-rent and main customer of Bell Labswill beoperating in a more competitive environment.It will iose some of its "guaranteed" markets,as well as implicit subsidies paid for in the pastby all telephone subscribers. Western Electricand 13011 Laboratories will have to learn to dealwith a less predictable set of market conditions;basic research is likely to appear more in thenature of a luxuryand hence be deempha-sized. Unless alternative mechanisms for per-forming basic research and diffusing the re-sults evolve and thive, one of the great sourcesof strength for the entire U.S. electronics in-dustry will atrophy. Th3 potential void is notrestricted to inicroelect.onics, but extends tocommunications and computer Technologies aswell; Bell Labs spawned the Unix operatingsystem now so popular in small computers, aswell as the transistor.

Leaving aside the special case of antitrustand perhaps trade policies also, where activi-ty has remained at high levers if without muchsense of directionFederal policies affectingelectronics have been marginal, indirect, oftensimply absent. To those who regard 'Govern-ment involvement in the affairs of industry asa usually unnecessary evil, this may seem ablessing. But the reasons for the absence ofpolio ire not so much conscious decisions thatFee- initiatives would be counterproductiveas a nAck of agreement about what Governmentcan and should do. The subject has been dis-cussedat considerable length; and in contextsranging from trade reorganization to the Fed-eral role in prodti ivity improvement and the.quality of working'fe. But the various partiesdisagreebeginning with the question ofwhether the U.S. Government should developpolicies aimed at affecting the competitive posi-tion of American industries, and extending toquestions of how the electronics industry, inparticular, might fit into a more general frame-work for.industrial policy.

On the one hand, some argue that the UnitedStates needs to search for new engines ofgrowth to drive the economy into the 21st cen-

tury. Others focus on organization, some advo-cating that the Department of Commerce betransformed into a more powerful agencyeven a "Department of International Trade andIndustry" modeled after Japan's MITI and re-sponsible for coordinating and implementingpolicies on a sectoral basis. Another view,while agreeing that new mechanisms areneeded if the Nation's de facto industrial policyis .to be replaced by a more systematic ap-proach, takes dispersal of responsibility formaking and implementing policies to be a hall-mark of the U.S. system, and sees a siagle cen-tralized agency as impractical, if not danger-ous. Still others, focusing on financial issues,argue that the first priority of the Federal Gov-ernment should be to channel investment capi-tal to speed "reindustrialization"for exam-ple, through a publicly operatedinvestment ordevelopment bank. In any of these views, elec-tronics would plainly be an early subject ofattention.

The dominant attitude within the executivebranch since the election of President Reaganhas run counter to the more activist positions.The Reagan administration has held that theproper role of-Government is to stay far re-moved from the affairs of industry. Those ofsuch persuasion believe that businessmen,rather than Government officials, have both theright and the ability to make decisions affect-ing the futures of their firmsand thus the fu-tures of the industries of which they are mem-bers, and the competitiveness of the U.S. econ-omythat Govei nment is largely incompetentin these areas. The corollary is that decisionsmade by private interests will affect local andregional economies, as well as the interests andlivelihoods of the people,who work for or other--wise depend on private industry.

Those at the extreme end of the spectrumhold that Government should minimize its ef-forts at macroeconomic policymaking, claim-ing that the Keyii,Aan economistswho, manyyears ago, defined a goverrirr.:.Int role in man-aging the aggregate economy through fiscaland mono -v policies represent a bankrupt-tradition. Ta..; reductions and other measuresaimed at capital formation,. rather than de-

4

Ch. 12Federal Policies Affecting Electronics:: Options for the United States 467

mend management, have been in vogue amongthe proponents of this view. Still, even ad-vocates of supply-side econe-. ics or monetar-ism commonly maintain thatwhile Govern-ment should stay out of the :.`fairs of busi-.nessit can and should ensure a climate con-ducive to economic

There is a core of truth to the claims of thosewho advocate a Government pullback from mi-cro-level involvement in economic affairsthistroth found in the past history of public policiesthat have affected U.S. industries with some-times adverse consequences. to say nothing ofthe lack of relevant expertise and analyticalCapability in the executive branch. On the otherhand, it is clearly possible to devise a self-conscious industrial _policy that does not de-.pend on direct or extensive Federal interven-tion. Thus, even if one feels that the pOliticalsystem in the United States works in such ways

that Government involvements will always beriddled with mistakes arid policy failures, thisis more a counterargument to proposals for astrong, centralized industrial policy apparatusthan to the general notion of a more explicitand coherent indusTrial policy.

In essence, two of the attitudes sketchedabovecentralized industrial policy versusGovernment pullbackrepresent extremes inopinion concerning the form and character offuture industrial policies fox the United States.Industrial policies have existed for several hun-dred years in this country; Federal Governmentactions will continue to exert influence overprivate sector decisions. On quetstions of howthe process might be changed, as on specific

_policy issues, there are many shades of opin-ion, a variety of perspectives that: fall betweenthe extremes. These are illustrated in the re-mainder of the chapter.

Alternative Perspectives on Industrial PoliciesRegardless of one's attitude toward Govern-

ment involvement in economic affairs--whenand where appropriate, for what reasonspol-icy choices will flow in part from analyses ofthe position of American industry, and the in-terpretations placed on these analyses. Eachof OTA's five alternatives has as its founda-tion a somewhat different interpretation of thecompetitive situation of American electronicsfirms; while the discussion that follows takesits context froin this industry, the policy op-tions are not specific to electronics, or ever: tohigh-technology industries as a class. The fivealternativeswhich .overlap to some extentwhile representing fundamentally differentviewpointsare:

1. Policies intended to ensure a strong do-mestic rit,,Lket base for U.S. industrieswithout particular reference to the natureof the industriesalong with preservationof existing jobs and job opportunities.

2. Policy measures designed to protect and/orsupport a limited number of industries

judged critical on national security orother grounds.

3. Policies that will support the technologicalbase and institutional iFfrastructure forAmerican industries, particularly thoseundergoing structural change.

4. Policies designed to promote the globalcompetitiveness of U.S. firms and indus-tries.

5. A policy that defers if possible to the pri-vate sector when choices concerning in-dustrial development are to be made.

All of these, even the laSt, accept at least im-plicitly that Government involvement in indus-try and the c:-::onomy is inevitablethe ques-tions being vven, where, how, for what pur-pose. In each case, different sets of assump-tions and goals underlie the policy orientation.From the repertory of policy tools availableoutlined in chapter 10 and summarized in table82each of the five would call for a differentmix and emphasis. The table is schematic, butgives an idea of the types of measures that

4a1u


Table 82.Measures Likely To Be Emphasized Under Alternative Approaches to Industrial Policya

AlternativeCritical

Protection industriesInfrastructure& adjustment

Global Minimumpromotion Government

Trade, foreign investmentTat

Competition antitrust, merger)Human resources (education, retraining)Technology (R&D, innovation,

diffusion)Irvestment (capital)GovernMent procurement ,..-

arn,s table is :mended only to be suggestive. many POssdpilities exist under ei..ch alternative

50tinCE Of-,e of Technology Assessment

would be emphasized under each of the alter-natives; a "critical industries" orientation, forinstance. would entail a strong presence byGovernment compared to the other four.

Regardless of which of the five alternativeswiTe chosen, a new approach to industrial pol-icy for the United States would bring pitfallsas well as opportunities; experience in othercountries shows that there is no substitute forgood judgment in selecting and implementingindividual policy measures. Table d2 stressesthat it is not so much the individual policy toolsbut the way they are put togetherthe objec-tives pursuedthat matters most. The remain-der of the chapter treats the five alternativesin detail.

Ensure the Domestic Market Basefor U.S. Industries

Protectionism is a loaded word. Not onlydoes it imply reversal of the primary thrust ofpostwar U.S. foreign economic policy,. but thearguments in support of open internationaltrade are strong and widely accepted. Protect-ing domestic industries from import competi-tion via tariffs, quotas, or other barriers distort?.market mechanisms, decreases economy '3ffi-ciency, andby raising pricesresults in a netkiss in standard of living.2 Hardly anyone dis-putes these general tenets; the issues morecommonly raised concern the' specific cir-

21'or a brief review,1. Site U.S. Inchstrial Competitiveness:.Comparison of Steel, Electronics, and Automobiles (Washingto,.!,D.C.: Of ficr: of Technology Assessment, O'fA-1SC -135, July 198 :),pp. 181182.

cumstances under which trade restraints might_be justifie.! to prevent or ameliorate greaterharm to a iewvulnerable firms and their em-ployees, communities and regionsat the ex-pense of net benefits that, when spread overthe Nation as a whole, are small.

Leave aside for a moment the political ques-tions, as well as the use of protection tocountervail the industrial policies of other na-tions, or unfair trade practices by foreign enter-prises. The question is then an internal one:What are the impacts within the larger domes-tic economy of protection granted a particularindustry? This is not only a matter of present-day costs and benefitse.g., to consumers, toowners, managers, and other employees butof the future prospects of industries grantedprotection. Some such industries :nay be intemporary decline. with reversal possibleothers unequivocal victims of shifting com-parative advantage. For example, long-termprospects for specialty steel manufacturers inilia United States appear brighter than forwakers of carbon steel. Not only do the tech-

.-;al demands of specialty alloys favor Amer-ican firms, but the diversified, high-technologyindustries of the United States provide largeand varied markets for alloy steels. None-theless, both specialty and carbon .steel pro-ducers face short- as well as long-term prob-lems. It is possible, to argue on the one handthat trade protection will benefit specialtysteelmakers by permitting them to rebuild theircompetitiveness so as to take advantage oflonger term opportunities, while on the otherthat pibtection for carbon steel producers will

Ch. 12Federal Policies Affecting Electronics: Options for the l'-;ted States 469

_,,,,,.:._;j.-,-*.a.4%,2.

V'

..,. _.4....._......_t

.....q4Lik'.-. --

4

" ."-- '6°2' - .._._?,- rilV:: --

tgal

-----7-- .1-----x::

'"";-..;-4*.,"5.atil

''-''''.-'j ..41 .to--` -P.-------r '.--4 A. v.,10....-

4

-4&67°'`-----

--

American consumer electronic plant complex

only retard the inevitable contractions. At thesame. time, for domestic industries as large assteel or automobiles, the adjustment problems.stemming fromi long-term shifts in competi-tiveness can i so severe that strong argumentsfor temporal-, trade protection can be con-5:tructed on this basis alone. This is one of thereason%.. escape clause actions are sanctionedtinder Gi.TT (the Gene.-%! greement on Tar-

Biliite;alism in TradeProtective actions other than tariffs, involv-

ingas they generally dobilateral discussionswith exporting nations, represent somethingof a turn away from the postwar U.S. emphasison multilaterPi trade negotiations: Persistenttrade friction between the Utii;:ad States andJapan !"1,19 led to bilateral negotiations (:ever-ing ranging From cigarettes to tef,sz.om-

Fholo c 'ode!: RCA

rnunications equipment. Similar bargaininghas taken place with European nations export-.ing steel, among other products, to the UnitedStates, as well as between Japan and the Euro-pean Communitye.g., in the case of video cas-sette recorders (VCRs).3 To some, this revivalof a bilateral rather than multilateral approachto trade is a sign of possible return to theprewar era of widespread protectionism.

Bilateral negotiations between the UnitedStates and Japan first found a prominent placein U.S. trade policy during the early 1970's,when market penetration by Japanese textileimports became severe; in one of the more re-

Faced with dumping complaints and informal import rostric-lions in France, Japan has voluntarily agreed to limit VCRshipments to the European Community. The ceiling will be4.55million annually. E. J. Dionne, Jr., "Japan Video Accord LeavesEu:op. ans Wary but Hopeful," New York Times, Feb. 22, 1983,p. D5.

472

470 Intern3tional Competitiveness in Electronics

cent cases, after a push from the U.S. side, theJapanese agreed in 1981 to accelerate tariffreductions on semiconductors (ch. 11). Furthernegotiations led to similar concessions on dataprocessing equipment. Japan's continued reli-ance on tariff walls to protect domestic indus-trieswhether or nut they remain infantsandon an array of slowly crumbling and largelyinformal nontariff barriers, contrasts sharplywith duties on imports into the th,,ited Statesthat for many years have been low; as pointedout in the previous chapter, this country hasonly rarely imposed tariffs to protect domesticindustries, preferring to negotiate quotas.

Recently, those concerned with Japanesepenefration into U.S. markets for productsranging from semiconductors to machine toolshave beet urging a variety of essentially pro-tectionist responses. The mirror image of con-cern with imports lies in the persisting diffi-culties many American firms have faced inexporting to Japan, or investing thereevenwhen the exports are goods in which the Japenese economy is uncompetitive. The paramot.aexample has been agricultural products.5 Theperceived asymmetry has been a major forcebehind calls for reciprocity in trade.

While to some, trade reciprocity need notcart y the in.).:r:ation of sector-specific bilateralconcerns, to t,.rs a means just that: if nationssuch as Japan discriminate against U.S. exportsor investm .nt,'we should retaliate swiftly anddirectiy. 5tirinp 19F 20 or more bills relatedto um -pci!yand covering manyshades of r r introduced, withmany re-lnli :ed in the 98th Congress.6 InApril I ;183, S. 14 the Trade and InvestmentAct of :983, intended to strengthen the hand

`An nt exception has been tha levies of 45 per..:ent.declining o.:er a 5-year period. placed on large japanose ---cycles early, n. 1983 These wen: imposed as the rest''escape daily! action. See "President Ira tees Sharpcrease cm Motovc,,cies. ;span Criticizes ...c.tion." 71.S. ImportWeekly, Apr. 6. 1083. p. 5.

5See, for instancy, Report tin Trade Mission to F:..r Rast, SUb-4.011111'01..0 4,11 Teild,!, Committee on Ways and Mc as. U.S.House of Representatives. Dec. 21. Mt

"See A. Redman and R. Ahearn, "Reaii.rocity is t-oreign'Friuli.," Issue Brief No. 111820.13, Congressional Rg.seach Serv-U:e. N1i.:r. 3 1083. Most of these bills won' intended to give thePresident authority to impose restr:ctions 1,n imports:se..:iral were IF re( led largely at /rad,: in .,:er.,,icet

of the President in dealing with other nationspassed the Senate unanimously. House billshave made less progress.

Most broadly, reciprocity is a call for equaltreatment, hence would entail measures to re-strict imports originating in nations that them-selves block the entry of American firmspar-ticularly through indirect barriers. Spokesmenfor the U.S. semiconductor industry, to take anexample from electronics, object to unlimitedimports of Japanese ;Cs at a time when they7'ee themselves confronting a formidable arrayof obstacles to doing business in Javnobsta-cles ranging from uncooperative customs in-spt;...ors to hidden controls on foreign invest-ment '

Pros and Cons of a Protected MarketBase Strategy

Arguments for temporary as opposed to long-er.term or permanent trade protection turn onquite different points. Proponents of temporaryprGtection for troubled industries.often judgesharp upturns in import shipments, as for col-or televisions during the 1970's, to be par-ticularly serious. Once lost, whatever thereasor market share can be ;difficult to regain.Thus, a sudden penetration of U.S. marketsperhaps as a result of unfair trade practicesmight devastate an industry, leaving it withoutthe ability to recover. The remedy is to protectthe industry. Whether the causes are lowercosts for labor or other factors of productionabroad, unfair trade practices such as dump-ing, or problems internal to the U.S. indus-trywhich could range from outdated plant fa-cilities to misjudgments of the market the ob-jective of Federal policy, in this view, shouldbe to limit import penetration with the expec-tation that, after a limited period of relief,domestic firms will again be able to compete.

Preferred measures to achieve such goals de-pe. on the circumstances of the import-af-fectPd sector. Examples from the recent pastinclude tariffs, uni:uterally imposed quotas,

'(in the latter, see U. C. Lehner, "Iti:ian'i Av.- &al to SeningCornpanis May Be Ultirnatii E , :: rirr in rade," MI' StreetJournal, Mar. 23, 108Z, p;.38; also. S. Loh r. ulapa.)', CapitalMarket lias0.S, Ne's York Times. limo 1, 1082, p. D3.

Ch. 12Federal Policies Affecting Electronics: Options for the United States 471

negotiated Orderly. Marketing Agreements(ONIAs). and voluntary restraint on the part ofexporters. In industries where allegations ofclumping have been commonconsumer elec-tronics, and, though there have been no formalco,-nplaints in recent years. semiconductorsalternatives to antidumping proceedingT. mightbe sought because legal redress has proven\slow, notably in the case of television imports.`fhe Trigger Price .Nlechanism for steel illus-trates a more nova 1 mechanism.

1.teyond temporary protection, the notion thatthe\ Federal. Government should provide aha0.ri for t S. industry appeals to many in-teress. mi,,,ht_be.to help theAmerican economy grow, to protect jobs, tomaintain the prosperity of cities, States, andregions,\,What are the arguments for more corn-prehensiVe or longer term import restraints?

Policies designed to ensure a domestic mar-ket base for American industries might be jus-tified on the assumption that a certain level ofsales at home provide the necessary foundationfor international competitiveness. In essence,a variant of infant industry, senescent industry,and critical mass arguMents, at root this per-speetive views in-wort penetration as inherentlydangerous, hence' worthy of Government atten-tion. Keeping out imports permits domesticmanufw.:turers to achieve scale economies andto earn the profits necessary for investmentsin new production facilities and R&D. A proL.tected home market vuuli:. also insulate' hemfrom sudden and unexpected competitivethreats. originating not in profit-seekingoverseas firms, but in goVernment-controlledenterprises seeking to create jobs, earn foreignexchange, build industries, that can sup,iortmilitary adventures. In short, this strategywould insulate the Nation from the disarraysof a world econoiny that is simultaneouslymore open to ail corners and more susceptible tomanipulation by organizations seeking endsother than those of private co.orations inAmerican mold.

If indu,:trial pi:acy is too strong a term to(!r;crilie foreign tactics, it is n.)netheless.trueti s3 nptionalizun enterprises can with

7.mpunity goals q...ite different

from those of firms that must live off their ownprofits. And if not all countries have national-ized sectors the size of that in France, in manyeconomies the incidence of government sub-sidy and control is such that market signalsbecome disti.ictly secondary. The precedingchapter stressed the relative impotence of thetraditional resler of trade lawsand of inter-national re?,--.nations for countervailing thewide range o:*.s!tpports and subsidies that somegovernments nol.v resort to. Those who seeworld trade as moving toward a no-win sit :a-firm for the United E.tates sometimes urge that1.ve shut our own borders to imports, acceptingthe consequences in terms of reduced exportswhile relying on the size and diversity of theU.S. economy to keep productivitymore gen-erally, the gross domestic producthighenough to maintain living standards acceptableto most American:

A related justification for trade restraintsstarts with international differences in wageratesa 7)oint emphasized in chapter 4 (seetable 27). , .ow-wage countries, many with hugeand mounting labor surpluses, can now pro-duce many types of goods at costs below thosein advanced nations. Increasingly, this is trueover a range from primary metals to manufac-tures like automobiles or the simpler electron-ics products that were mainstay: of countriesindustrializing earlier. Although labor produc-tivity in developing countries is often very low,if wages are also low, costs of production canbe less than else,where. Eeen Japanwith payscales in manufacturing industries moreth:.ti half those here. and labor productivitiesin some cases better faces competit'.fe diffi-culties in sectors I1.ke consumer electt Allies orsteel.

How can the United States hope to competeunder - :ch circumstances? On answer is tooffer .,ducts that are beyond .ihe tec1,:no-logical capabilities of low-wage countries, Inrro,:e conventional products, it maY-6c7-possi-ble to improVe labor productivtly--ilncughautomation or other advanced manufacturingtechnologies enough to caset e7,:isting wageciaerentials.\. Advocatv,,, of trada prote,;tion

1

z72 Con-7,7erit' eness In Electronics

point out that these avenues may not guarantee-flough lobs to keel) the American labor forceemployed. The alternatives are then to let\iges in the t"nited States drop. helping tomaintain competitiveness across a broaderrange of production. or to keep out importsfrom low-wage countries. Given the levels towhich y'. ages !night have to falldepend-ing on lice..' swiftly developing countries caniniprove their own labor productivitiesthefirst alternative is far front acceptable. Theplight of unemplmed auto and steelworkers inCalifornia.,:s.vbere plants in both industries havebeen shut dowh, illustrates the difficulty. Someir.not..all of the workersmany ofwhom had 'been making $12 to S15 an hour ex-clusive of fringe benefitscould find employ-ment in Silicon Valley electronics firms. How-ever, unskilled or semiskilled electronics work.pays in the range of Sti per hour: and even atthis level. Atari, for one, is moving some 1,700jobs overseas. Given such a picture, trade pro-tei:lion and industrial self-sufficiency begin toseem tractive.

%Vino is the other side of this scenariotheargument against either temporary or longerterm trade restrictions? First, import restraintsalmost always result in higher prices for Amer-ican consumers. Witness the automobile priceincreases following; Japan's voluntary limits onexports in 1981. Indeed, under such circum-stances 'rice increase::, are generally intended:the cominon rationale is that import-affeciedt S. firms must he temporarily shielded so thatthey can raise ,.prices, generating increaseoprofits to he invested in restoring their coin-

-..!ss. Of course, rising prices often rip.:pie through the cconomycausing inflationarypressures. To the extent that protection fordomestic steelmakers has raised steel prices,costs gone up for automobiles, consumer

roads, bridges and buildings, militaryhardwi.ireeveryvvhere steel is used.

Should these higher price:-; be considered thenecessary costs for reviving import. ;Meted in-dustries? Where there is goc ' reason to expectrevival, the anoser might bc: yes. Unfortunate-ly, experi:,.nci.---e.g., in the Case of color tele-.

pro,fides little ecicf1..:nce it) support of

trade protection as a road to recovery for sec-tors that have lost_ competitiveness interna-tionally (which is not to say that protectionmight not serve other objectives, or be anecessary if not sufficient prelude to recovery).Industries ancl/er their employees may claimthat import penetration stems from unfair tradepractices, dubious _management decisions,adverse effects of Government regulations, orother transient problems. If so, the argumentruns. recovery is possible, given ime. The reali-ty is generally more tangled. Co. plaints of on-fai:- competition or adverse regulations may beIvel!-founded but nonetheless only secondaryfactors; decline may result more fundamentallyfrom long-term trends .n the world economyi.e., shifting comparative advantage. Where thisis thc case, trade protection will be ineffectiveif temporary, costly if permanent.

When a good argument can be made that re-vival is possiblethat longer trends favorthe United States or at least do not run toostrongly the other waythe question remains:How long will protection be necessary? Where_the Government has imposed or negotiated im-port quotas, these have typically been for 3- or4-year periodswith renewals not nliheard of.Fixed periods are desirable so that domestic.'es well as foreign producers face a relativelypredictable situation. Protection granted for anindefinite period risks de facto permanency,decreasing incentives for domestic firms tomake new investments or alter their businessstrategies.

To illustrate some of the factors involved indecisions on protective mechanisms, considerthe situation in early 1982 as concern mountedover imports of t.+4K RAM (random accessmemory) chips. Japanese penetration of the

marketrumThIg at about 70 percentwas the out.:;olne of a complex of factors: rapidcapaci'iy expansion ID Japan facilitated,by am-ple supplies of canitek fcr investment; produc-tion pro! '.:;ins at the plants of several prospec-tive U.S. suppliers; pHce7eutting by both Jap-anese and American firms striving to build,i,arkc,, share (acce ipanied by accusations of(Imp, oft !,-,,('led at the Japanese). At a time

rim el merchant firms--

;

Cr:. 12Federal Policies Affecting Electronics: Options for the United States 473

c>

Texas Instruments and :Motorola were ableto 1);'utiti: 64K RAMS in large quantities, asopposed to six Japanese manufacturers. thefirst question froin the standpoint of com-petitive dynamics became: How long would itbe before other suppliers entered thearena? Given the learning and scale effects,:haracteristic of RAM production, too great ahead start might be virtually impossible to over-come. On the other hand, if American com-names carne in later but with superior designs,would they he able to turn the tables? Theseare nontraditional kinds of questions for U.S..poli4:ymakers, indeed difficult for governmentsanywhere to deal with: as emphasized in theprevious chapter, the fast moving events char-acteristic of high-technology industries do notfit verN, comfortably into the existing frame-wOrk of international trade policy. But effec-ti::7 Government action depenth; on graspingsuch facets of competition.

To return to the question of the costs asso-ciateed with trade protection, note first thatregir(liess of rationaleimport restrictionsFunction as implicit subsidies for proteeter I in-dustri':s and their employees. The cost'- arepain by other sectors of the economy i.e., bythe public at large. Beyond direct costs in theform of higher prices, a protected industry mayhe able to attract re,.ources such as capitalaway from other parts of the economy; in at-tempting to help one industry, Governmentpolicies can harm others.

These are not the only indirect effects forpolicymakers to worry about. Foreign compet-itors often pursue inward investment as a wayaroudirade barriersthe pattern in color tele-vision, now also taking place in industries asdifferent as microelectronics and automobiles.In 'contrast, foreign investment in U.S. steel-making capacity has been smallno doubt be-cause overseas investors do not see long-terrtrends favoring the production of iron and steelhere. Direct investment is partk.ularly attrac-tive where companies feel they have competi-tive advantages that can be exploited regardlessof location. American semiconductor and com-puter manufacturers invest overseas in partbecause' their technological advantages areeasily transportable.

0-111 n - 71

about the Lne Texas Instruments an-nouncec.' it ',:as transferring all of its 64K RAMproduction to Japan, Hitachi. Nippon Electric,and Fujitsu revealed plans to moveor sneedup previous timetables for movingsome oftheir own 64K RAM assembly here.8 One mo-tive was to dampen trade frictions; despite theabsence of constraints or even formal com-plaints concerning RAM shipments, the colorTV case appears to prefigure that in semicon-ductors. Is this an outcome that U.S. policy-makerswhose actions accelerated onshore in-vestments in consumer electronics7-shouldwelcome? Certainly there are major differencesbetween the two industries. Competition intelevision manufacture is cost-driven, withtechnology playing a relatively minor role. Inmicroelectronics, moving closer to markets isone way a company can capitalize on its tech-nology to meet customer demands. Althoughdecisions by Japanese semiconductor manufac-turers were spurred by concern over trade,they see many other advantages to their pres-ence here. For instance, they car learn fromAmerican technical expertise more easilyoneway is to hire American engineersif theyhave bases in this country, especially now thatU.S. companies are guarding their own tech-nology more closely. Irithe same way, te,,hnol:ogy acquisition has been one of the motives be-hind efforts by U.S. firms to set up R&D andmanufacturing facilities in Japan.

When foreign firms invest in U.S. plants,they employ American workersunskilled aswell as skilledalthough a substantial fractionof value added tends to remain overseas. Butfrom the perspective (:f U.S. semiconductorfirms, sales by Japanese-owned competitors-regardless of where the products are manui'lc-turedrepresent a loss to the domestic in-'dustry. The numerous joint venture and tech-nology exchange agreements that U.S. and Jap-anese electronics firms have entered into corn-pP,;ate matters further. With Hewltt-Packardgetting RAM technology from Nation-al Semiconductor sharing with Oki, the com-puter firm Amdahl joined to Fujitsu, easy na-tional distinctions vanish. Such trends are still'

Kanaboyashi. "64K Ram Chips ilants in U.S.." WallStreet Journal, Mar. 2, 1982, p. 35.

470

Joj 0:1qop alll s.1,1r.,.1:,

01l p(i inn ):Iftp.tl!,.+A(Cif flPrItitftig :11/1.1/11

'V, TA{(iL 11:).11IN 'Ns'alt!s//1/ .11;1.1, IMP

J0).111r1r) ti1it10..1 'NUM; 11!sill!.1.1,1V110!111ill:P111 'S1) 11:1111

Tatialeam aq ppm.) 's! If se papadtuf apP.11 !VUOtiPti

JO walsAs aJtitto atp 'scunstlaw AIotuflui

aJ tiva.fls alnAon.1 Oj urgioa, sla)t.ww apsaump amasauf ,)1 paptraluf fupfsnpuf Ji um

'Sal.1101111.11...:1! 0111 .1.101.1,0 puu-a:lltu 'sfsurf 1:m1)wci-S(1-)anp1mi

U uo ,suaquu `duotuu lamp imp aq Audi

01110Pr.0 PUO :11.1011M 110 SUOIS5a0U0.9 Joj anurap -xa L11 sIalndwol.) uo suofssaatioaiiue;

'cum suopupaau fellow-m.1:mq sisaJoulf s.i UO specluif aufnuaii aAutf of snofAcio Emu 1:m.1

1p aq 1UU paau uo0rfitula21 -uofssaniaa ay Jo tp"i3nof aq1 pu.! 111(11: alp tooq of pammipplo:)

:Ramp 's,01;13 I '.3111 UI :samseaw A.mleftelai 01 pal umia ;Meg SA101.1 OPeJI U0 suafpufsaJ 'Afirmf.rof

-sfyl uru Ja2uol alp tit 1Spa1 1e ltiauaq pina.v, auoiclaAa letp Jaffa(' acp UI ape.0 uado Jot maul

-PAOW 0111 pat A.111.11-100 pa)furi atp 1OJ aauuls u6pDaloid imputu e aleacmpu pinoAl

otim °sot o1 asnud pinotis-saidwexa -.10111i103 Auew tofm puu SfOuplug Ifarlle-apu.I1 ppomio tuariAs t.rado in? f.mumn/ puu tLsfuotpat

-cud woJi Aume uou,t Atii:Ui .10.1

°Aug suopuu 2utpu.rt suosew a:11 'Al futtf,.1

-Spawa.1 algetA e uaAa JO isag .91p sr uopaa/0.1(1

imp ueaw )ou saop 2upsej-Suot Atippualoci pue anAas a.JP, suopeDoisfp 12111-Z861. Ui uong LLS

critIfta101 11;1(1E1 4TM -ofjap )StpUF?ti.:iipW p Jo surmsuawtp popfiaci all; .1,.)u-ape.11 UI slitqs Act

paleala swap-told )uatur-olpe Tunic] ol2uftpou swop am-rmd -)1uSa.180e ato alum Jattuni titic

scuotied peo.uf asacp sannsgo-sapciotuainu JO SJOlanptioap.ups Impatim -spapoici .11,111:111

-.led as wow tLa.Na--UtKitlf pUP safeis per!tii-I uaamiaq imp se Epos sxmlufugtuf fe.. pq

-ssa.fis 6.)tofFup astpuetp.r:,tu aqi Mifpoo:)xo .tit scflf:uas sn[cDos i861 all; 'uoffifq LIS pafulai saDfAJas pue, spoor uo snid.ins S 1 atp ti:rttim

tit Ipp.i e smotfoi sfql uoff imq SpAN z961 sa:mAa:5. pue spooii 1pfiat) lou :-Ipulf tio snit'

pp:lupip(i ,s:Taspaid 1S01111p SpA\ P.1081.1

Sp.11 tit tiO111tc1 cvtlip7)

uo p0fiap opal! -s.1-1 aqi optim ';uno..y.m! Jo too oaf ttaiv, ';1--tulfclu:)io smait futloqutwol;

cuouf--saafAaas tit optui..s:uni -;minutmu Sputaap.ted 'spo0 1;1111!1 ma s1(')0i

s1 ))1.1ew 1utiopfpu.11 Ui uofffso( sr- atp ozfsutfil

-um pplOhl 11!Ll1 AlIPIP.I!S 1! J0-.10(111 poZtti -PSJO UI asagf 1.1f pop. jsoi\:

'pup aip llt-Htlattip.:0A0f) Jo foil -110,, aptS1110 SaaJOJ 111J.10,%10(1 EWA\ pflitio-, .iucu autfaap put asl.J fu.lorms 10 stual)1

aa:ui 01 siciwapu Attafe.tis 10)1.1PUI misotnop Jr

imp tuaixo otp o,J. .inisSa:mns A.10:\ tuopfos alp sfip idtuapu Imp sapflad luulsoptif :o;M!1111',.,pu

a,Ntlu.luciula;) UI stifqs ValfAllaptin Artil Sp 0a JO sociAi atuos Jot sloiluto :mison;

-op ;nstio of 'itificittlaffu II11S Afpua.uf ;unix

atulsas :mu imp sualyitutv-,:su uo s)soa sial.tuut :)tisottic,f

.'11!..,1,)so.1(1 p.fumof paittapo .A:mfad ruiest10 -uf tie ..I.rwituutt0 (It feuofluu.lama aJatm nutty; -aq slu)1Jut11 PUP sapisaptif .srlucl.tuf poxfut

.fpappap tiffm (nom! opup, atil Lrt Alau o.utf se Lions) suofspop asu:)-Aq-asua io tumpowyi

Lump A:mfad p SSPlatii00011 Ker;fut linso.1 aq1 'ase;) ftuaua`ii atp tit ua11:)alo.1(1 11fln0.1111 .C.1stiput

sri JO alp o s;.priod appli 01 oFFssoci nq AIM! Ir `0011p1SilliTur) A10,1.011

pip 1! 1 ,;.11!0(1(IPSip 110A0 ',1111(1 Sr;c1.101111

Sat.11SilpLit utflaloi puu 'r1 1100A1F1(1 f-11.iPp

LInGg atp uatiAN asPq 131.1u11I 'Jtis1)tw.;1) P

3110 sPOp MOH St10110:l1saJ apP11 110 papt11101

feutstipuf UP ..:0; i,u1(0,,pp fultioutup -uni P aleada aAoqu 7)0(1u:is:)f) sulollud

46aleils uoRorAoid-la iJeykj e pinoo

s.loN,PtuA:-mfoil

otp O1 1-iiittyp 'spun-0 polo .ti;ialotitiPol-tptut III 1);11:)0(1X0 ;Hi 11 ('" .)11;,f Ell a0.11-i 1! ;),,,f11\ 7,,,11(f

-1-107-)0 :)!I5ati;01) .I0J sLRP ditl;i11S ft41(;r11.,:i

0(1 Alplett utr.)-,...to:mod at: all ..,:u10 poppp on afau,v, tpt-1,,; JE

salps pill! i101111(p..1r-qp p.-,..:0111..clupl.1

alp 11.1P110(1 ()11,11 "S:1 II

Ilutpeal 1,)tmin })lo,; oiu nut-h7f

:Tutu -.10.10 pa-mit-m.1cl

,nrituud s:mt10.(1:)010 7;:r,iP11-W ;i

Cr). r2 Federal Pcicies At fectirg E_-"riectroncs: Options for the United Stated 475

Fear u i.etniia.iion has been n very real fa;tor in the choices made by the United States

concerning trade in steel. particularly im-port from Europe. If the United States makesit ton difficult for European steel to enter theAmerican market. it risks restrictions on U.S.exports of electronics products, or f.inanciaiservices e'..en military goods. Trade wars sel-dom benefit those involved so much as thoseisI position to pick up the spoils.

Support for Critical IndustriesRather than attempting to preserve domestic

markets in generalwlyich would in largeliWitSlIre reward those able to build thestrongest political constituenciesthe FederalGovernment could decide to support and ifnecessary protect only those industries hidgedcritical to national security. Security might bebroadly or narrowly defined. Either way, sucha policy would find deep histor;eal roots. Gov-ernments support transportation technologiesand systemscanals, railroads, highways, avia-tionin part for reasons of national mobiliza-tion and defense; other examples range fromarmories ind shipbuilding to telecommunica-tions regnH'ions and space exploration. A na-tional sem: ,:y criterionrestricted to militarysecurity or extended to "economic security"would narroN the focus compared to the mar-ket preservation alternative, helping to controlthe political pressures that will always bedevilefforts at industrial policy in a country like theUoited States.

Manufacturing sectors suffering from import(..:ompet t ion frequently argue for Govern meatremedies on the basis that their productsortheir plant and equipmentcontribute to na-tional security. Some clearly have better casesthan others. The end products of some com-panies and some industries consist of militaryhardware: armaments, communications sys-terns. In other cases. end products may be usedonly indirectly for national defense, though noloss critical for that. This is true of supercom-'inters, needed in the design of some types ofmilitary systems. In still other cases, the goodsproduced by an industry may be vital, but onlysome fraction of the industry's production ca-pacity would ever be consumed in meeting mil

nary needs. Examples include the st; eland ma-chine tool industries.

The assumption underlying a critical or stra-tegic industries alternative is that only a subsetof economyperhaps relatively smallisindispensable for national defense; unless thelist is kept short, this approach would differ lit-tle from the first option discussed above- Grit-i(al industries would begin with. but not be re-stricted to, the traditional defense sector: aero-space, suppliers of armaments, military elec-tronics firms, R&D contractorsenterprisesthat, along with large numbers of suppliers andsubc(..anractors. sell to the Department of De-fense (DOD). Beyond this, other portions of theelectronics industry would be obvious candi-dates for any critical industries listcomputerhardware and software, integrated circuits,communications equipment. Indeed, numer-ous manufacturers of computer systems andsemiconductor products have divisions de-ited exc?usively to military sales. In contrast,

consuc..ar electronics, as a sector, would havea weak case despite the fact that firms like RCAand GE are major defense contractors, just asChrysler's tank business was largely divorcedfrom the automobile side of the corporation,so electronics suppliers that engage in militaryproduction generally do so through separatedivisions or subsidiaries.

What is Critical?

The difficult questions in identifying "crit-ica" firms and industries imolve those that donot engage directly in defense-related research,or production, but whose products or R&Dmight still have vital military applicationsunder some circumstances. Synthetic fiberslike nylon and Kevlar provide an example.Used in clothing, parachutes, body armor, andfiber-reinforced composite materials for struc-tures ranging from missile casings to stealthaircraft, these materials are obviously criticalto the defense base. But would this .have beenpredicted when synthetic fiber technology wasin its infancy? That is the nexus of the problem;if the Federal Government is to support criticalindustries--identifying those that will he vitalin the future.

4

47.5 E-lec:;rcrucs

the telterie are to extficti to cini)::iiceect,11()11,1t; strength. then the leaner o! nlli-fying critical industries--without being so in-citee'.'t, that support and protection go toevorvmle who asksbeeomes still more per-plieke ee One reason is simply that terms like-economic strength" are not very meaningful.In practice, virtualft industry threatened

fore42,n competit , ould attempt to de-clare itself cI itical. alit:tiny in recogniz-ing iedustries that will he critical in the futurewouli.i arise here as wellwhere it is a variantof the -sunnse industry" problem. Once thesun np, dad everyone knows it, Federal pol-icy may not be especially important; oppor-tunities will be evident, investors will be at-tracted. Alt Iv nigh Government might be ableto nurture the growing industry, its role couldwell be peripheralmore so in an eco nomy likethat of the United States than in japan orFrance. But when an industry is truly an infent, its prospects for the future uncertain, thenGovernment may be no better able to recognizeits potential than the private sector (sc:eewould say the primary difference will bethat GovO'rnment's time horizons need not beconstrained by the desire for quick returns oninvestments as for private suppliers of funds.

In any event. a strategic or critical industriesapproach implies that the Federal Governmentcan and should identify such industries, thenadopt 7eolicius to:

I. Ensure that the United States eittintains anindigenous prodteltion capaeility suffi-cient to meet direct military needs. par-ticularly in the event of national mobiliza-tion or crisi

2. Support industries and technologies thathave a substantial role in providing the un-derlying bte-.eeither in terms of R&D orproductionfol. U.S. military strength:

3. Optionally, support industries and tech-nologies that clearly and unambiguouslycontribute to economic strength.

In a context of growing East-West telUadvocates of such an approachparticularlythose who emphasize, direct military produe-

tione-conteed that the U.S. Government shouldtake a more active role in ensuring the well-being of strateg:e industries.16 A primary strandin the argument is that if the United Statescomes to depend too heavily on foreign prod-ucts or the -Nation's defensive

could be impairednet only in theevent of war, but even in a rapidly escalatingarms race."

As table 82 indicated, a wide variety of policyinstruments could be used to provide for thecontinued strength of critical industries, goingwell beycnel tariffs or quotas for protecting do-mestic manufacturers and beyond the well -established relationships that already link DODand the community of rnilitt.ry contractors andsuppliers. Multiyear procurements have beensuggested as a means to strengthen the defenseindustrial base. DOD is also paying a good dealof attention to manufacturing technologies asone way of getting more for our money, as wellas short.:ning procurement cycles. The atten-tion to manufacturing will have spillover ef-fects in the civilian economy that could be sig-nificant. Beyond such steps, secbral policiescould provide targeted supports and subsidiesin much the same way that the American farm-er has been given special consideration. Pro-curement could be steered to particular firms.DOD-sponsored R&D efforts like the VHSICprogram and the other research and engineer-ing activities of the services and the DefenseAdvanced Research Projects Agencymany ofthem related to electronicrmight be enlargedand broadened still further, with the aim ofstrengthe!!ing the U.S. technological hese andinfrastructure. Manpower and education pol-icies could channel institutional support to-ward engineering and relevant scion ,!s, fund

,°For a detailed presentation of this view, see "Statement ofGen. Alton 0. Slay. Former Commander, Air Force SystemsCommand,' Revitali7 tint: and the U.S. rr7onomy, hearings, Part

Subc.ommittel, on Economic Stabilizat, in, Committee on Bai;It.ing. Finance. and I Jrbari Affairs, House of Representatives, Feb.25; Mar. 25, 26, 1981, pp. 258-479.

"Stich arguments were athanced by opponenU of th-.: awardof contracts to Fujitsu for a Boston-Washington fiber-optic com-munications link After intense lobbying by the 1)00 and others.AT&T gave the contract to its own subsidiary, Western Elec-tric. See E. Meadows, "Japan Runs Into America, Inc.," For-tune, Mar. 22. 1982, p. sta.

7, j

CS. 12Federal Policies Affecting Electronics: Options for the United States 4-, 477

SUIdell!S minoring iii these reward peo-ple who choose to work in defense industries,provide incen :es for retrainilig and continu-ing education in advanced technical subjects.The services have recently argued that, in theyears ahead. they may be unable to meet re-quirements for skilled workerselectronicstechnicians. Lf'-ospace fabricators, aircraftmaintenance sp- 'a listsas well as ,olgtneers;a strategic industries policy would air to rec.,tify such probhiins.

Critical Industries for National Defensel'.:ations traditionally give special attention

to economic sectors on which military strengthand secur:tv depend. Shipyards and armoriesare obvious examples; for many years, histor-ians .and economists have probed the sNmbiosisbetw.-!en military and civilian production, ex-emplified by the evolution of precision manu-facturing and interchangeable parts. If 19thcentury productioa technologies were drivenin part by military nee.ds, certainly the relation-ship between military and civilian sidPs of theeconomy has altered greatly since. Theprva-siveness and complexity of modern technclogymakes identification and support of strategicindustries more problematicships, arms,even missiles and panes, hardly exhaust thefequiremeo.; of modern warfare. DuringWorld War II, automobile. plants could be re-tooled to make weapons, but in the past fourdecades. military and: civilian technologieshave diverged, As for commercial technologies,new demands and applications come in rapidsequencechemical and biological warfare,terrain-following cruise missiles, surveillancesatellites, war in space, cryptology, computer-ized translation of foreign latuages. One neednot stop here. Economic w.rfare, in variousforms, has a long history. Wheat, cobalt andchromium supplies, energyail can be weap-ons. Ultimately, a nation's military potent;.al isa function of the size and composition of itsof , the fraction of gross national product

f. Wing to spend on defense.

Sooner or later, then, any policy based on acritical industries approach will face -a seriesOf decisions on what is really essential. Lines

ij have to be et-awnin some cases fairly ar-bitrarilybecause in the most general sensenearly all industries and technologies con-tribute in some way to defense readiness. Cor-porations may produce the boots that soldierswear, the food they eat, or small computers forbattlefield command and control. When only

portion of an industry's output goes to themilitarywhether the industry be steel or semi-conductorshow might the Government allo-cate its support?

The struggles of DOD with the "militarilycritical technologies list" recommended by thewell-known Bucy report and endorsed by Con-gress in the late 1970's shed light on the prac-tical difficulties of a defense-centered industrialpolicy. The first list of 15 militarily criticaltechnology categories was published in 1980.12Included were computer networking, largecomputer systems, software, design and man-ufacture of very large-scale ICs, and a numberof others related to electronicsof the ,15 cat-egori..1..i. only 3 had little or no electronics con-tent. The thrust of the exercise was to developa systematic approach to export controls; as aresult, it was narrowly focused on military-ap-plications. Despite the well-defined purposein essence to update and s :ipplant the Com-m( .ty Control Listprogress has been pain-fthly slow. Oncc 'he 15 general areas had beendetermined, the eilort bogged down in'details.Critics doubt that it will ever be possible toagree on procedure , for reducing the case-by-case reviews of export licenses that are nownecessary. If nothing else, the continuingdebate over militarily critical teihnologieswhich in principle seem relatively straightfor-ward to defineindicates how difficult itwould be to devise criteria for entire industries.After all, these industries would be rewardednot with export licenses that might add a fewpercent to reverr..:eshut in Lit least some caseswith substantial subsidies and °filer Govern..ment favors.

'1Tech nology and East-West Trade: e tWalthingtoil,D.C.: Office of Technology Assep.sment, May 19:33), p. 37. A de-fined critical technologies list published in cla.sified farm isthe end of 1581 ran to 800 pages. Also FOC Technology and East.West Trade (Washington, Office of Technology Aase&s-:nerd, OTAISC-101, November 1475). pp. 92434.

4 u

473 International Competitiveness an Electronics

Looking more narrowly at the electronics in-dustry. consider. again the situation created byimports of 64K RAMs from Japan and the re-sulting flurry of activity in the Federalbureaucracy. In December 1981, the CabinetCouncil on Commerce anr:1 Trade authorizedan interagency study of high-technology indus-tries--carried out largely by the Department ofCommerce." Several months later, the Depart-ments of Defense and Commerce began theirjoint examination of the national, security con-sequences of 64K RAM imports, consideringthe advisibility of a more formal section 232proceeding.14 As pointed out in chapter 11, thissection of the Trade Expansion Act of 1962 em-powers the President to restrict imports in theevent . of harmful implications for nationalsecurity; the remedies available include quotasor higher tariffs. At the time, domestic man-ufacturers were accusing the Japanese ofdumping 64K chipsin fact, industry lobbyingappeared responsible for much of the concernover national security.15 Simultaneously, DODwas trying to convince the same group of Jap-anese firms to transfer some of their technol-ogy to the United States, as well as to producecomponents and equipment that would helpmeet American, military needs. This was alsothe period when Texas Instruments was 'mov-ing its 64K RAM productieri. ';',:ipan and Jap-anese firms were announcing plans to makethese parts in the United States. A little later,the Justice Department announced its price-fix-ing probe of Japanese importersinvestigatingprices that might be too high instead of too low(ch. 11). Meanwhile, Congress was floodedwith trade reciprocity bills, some motivated bytrade friction in semiconductors.

"An Assessment of U.S. Competitiveness in High TechnologyIndustries (Washington, D.C.: Department of Commerce. Feb-ruary 1983). The first paragraph of the summary states: "Thisstudy is being released as a Department of Commerce document.The methodology. findings and concluSions do not necessarilyrepresent the views of other Executive Branch agencies" (p. iii).

"C. li. Farnsworth, "Japanese Chip Sales Studied," New YorkTimes, Mar. 4. 1982, p. D1. Also see ch. 11.

"On lobbying efforts by the industry, see "Horror Story," Elec-tronic News, Feb. 8, 1982. p, 12. One of the reasons nothing cameof the Section 232 study was simply that 64K RAMs new prod-ucts in the marketplacehad not yet been incorporated into anyU.S., weapons systems. thus, the national security implica-tion's of a supply interruption remained matters of speculationconcerning future weapons needs and designs.

Such is the circus for which a critical in-dustries policy would provide the rings. De-spite the concern generated by 641:. RAM im-ports, the underlyin. national security questionremains unanswereda question which is infact much broader than that of RAM chips orsemiconductors in general. One way to framethe questionin the context of microelec-tronicsis as follows. As technologies becomemore complex and industries expand, oppor-tunitieS for different countries to specialize incertain kinds of products grow; Japan's semi-conductor manufacturers; at the moment, arespecializing in RAMs. As a consequence, U.S.production might decline, with the result thatthe Nation could find it difficult to meet futuredefense needs, particularly in a situation call-ing for rapid mobilization.18 Again, the pointis that the ongoing dynamics of internationalcompetition hold one of the keys to policychoice.

So long as questions such as these remainnarrowly definedconcerned with particularproducts or with classes of technologyitshould be possible for policymakers to agreeon priorities and make the necessary choices.In its recommendations for the fiscal 1984defense budget, for example, the Defense Sci-ence Board ranked the following technologiesin order of importance for future U.S. militarysystems:17

1: Very high-speed integrated circuits, ex-emplified by the DOD R&D program(VHSIC) mentioned elsewhere.

2. Stealth aircraft.3..Computer software.4. Microprocessor-based teaching aids.5. Fail-safe and fault-tolerant design meth-

ods for electronic systems.

"Part of the reason is simply that the military market does notattract that many manufacturers. The 20 percent of U.S. elec-tronics sales that go to the military are unevenly distributed; insome product categories, defense needs account for only a smallfraction of outputElectronics, Jan. 13, 1983. pp. 128.140: Insemiconductors, the military market is perhaps 10 percent Ofthe total (fig. 34, ch. 5), and heavily weighted toward less sophis-ticated devices: during 1982, any 64K RAMs going to DOD wouldhave been embodied in commercially available hardward for usein offices (.r laboratories. not weapons.

"See R. Connolly, The Big 17 Future Technologies." Elec-tronics. May 5, 1982, p. 98.

C71. 12Federal Policies Affecting Fectronics: Options for the United States 479

Photo credit: GCA Corp

Directstep-onwafer systein for lithographicfabrication of integrated circuits

6. Rapidly solidified materials--e.g., amor-phous metals with high strength and re-sistance to corrosion.

7. Computer programs for artificial intelli-gence.

8. Supercomputers for nuclear weapons de-sign and computational fluid dynamics.

9. Composite materials.10. High-density focal-plane arrays for in-

frared imaging.11. Radiation-hardening techniques for elec-

tronic systems.12. Space nuclear powerplants.13. High-power microwave generators.14. Technologies for erecting large structures

in space.15. Optoelectronics,16. Space-based radar.

17. Short-wavelength lasers.

As many as a dozen of these are electronicstechnologies, or systems for which electronicsis a vital element:

Given some agreement on prioritiesofwhich lists such as that above might form onestarting pointand recognizing that prioritieSwould have to be reexamined and updatedmore or less continuously, what policy mews-

, ures, beyond decisions on R&D funding levels,I might then be called for to ensure that military\needs were met? Almost certainly, such ques-

t; ns would have to be approached much asfor those dealing with research prioritiesi.e.,on a case-by-case basis; given past experiencein trying to define critical technologies for ex-port control, formulating general criteria foran industrial policy based on national securi-ty would seem a hopeless task, one con-f-pounded by uncertainties surrounding mool;i-zation scenarios... Furthermore, quite apartfrom debates over the needs of high-technol-ogy sectors like electronics versus bas::: in-dustries like steel or machine tools, in-dustrial policy that set defense priorities con-sistently above civilian needs would be politi-cally painful. Like all sectorally based policies,such an approach is susceptible to the criticismthat other industriesand economic welfareas a wholewould suffer relative to sectorschosen for support.

One of the underlying questionsfor this andother industrial policy alternativesbecomes:Given the policymaking environment in theUnited States, would this framework contrib-ute to good decisions at the level of individualpolicy instruments, or would it simply curdusematters further? The Nation's policymakingsystem is not likely to change very quickly orvery dramatically. As a result, one of the pri-mary objectives of a more focused industrialpolicy for the United States can be viewedsimply as a movement of the system towardbetter decisionmaking on the average. Fromsuch a perspective, it would seem more desir-able to regard national securityparticularlydirect military procurement and productionas one factor to be weighed when making in-dustrial policy decisions, but not the center-

482

480 Internationa! Competitiveness in Electro -CS

piece. Where military security isstake, DOD and the defense community ge,1 ,c.5r-ally prove more than capable of marshav, i;"gstrong and effective arguments.

Critical Industries for the U.S. EconomyCould the United States profitably adopt

broader interpretation of critical industry,taking a leaf from the books of Japan Or. Frarr...!=.countries that have consciously tried toindustries that will drive economic growth?This would be akin to an industrial policy builtaround support for "sunrise" industries, whfAeproducts or technologies will stimulate at.idsupport other sectors of the economy.

Such an alternative has its attractior,principle, the Government could steersources to sectors that would have aeffect on tae rest of the economyor sirup':!to those expected to grow rapidly, increasi;.fgemployment and exports. In essence, an in-dustrial policy that aimed at targeting such in-dustries would be based on the premise thatGovernment can do a reasonable job of predict-ing where the Nation's comparative advantageswill lie in the future. This is part of what Japanattempts.

To pursue such an industrial policy success-fully demands:

1. Prediction of the sectors that will be vitalfor future growth and competitiveness.

2. The design and implementation of Federalpolicies that will effectively support thesesectorsstrengthening their competitive-ness in ways that markets alone could notor would notbut without creating unac-ceptable distortions or misallocations ofresources elsewhere in the economy.

3. The political will to pursue such policiesin the ordinary circumstancewhen thepressures generated by declining firmsand industries and their employees out-weigh public perceptions of rewards forencouraging nascent industries.

The first of these is relatively easy, at least onthe gross le' '!n the ability of Government

to "pick winno7...,."' is c;.AaS5,:i4:YE!.d, secondand third poin'<!--i are ,i,;c.,ner:-;ly at it-islre.

lt is only a ti say that,eve:.-gone knows the. nne..1:;. will come-from.. For the U=3.1V.Ni States t_71.1-ftlx .qt-ivanced

industrial econt4-,--,Itz.s, the crj rren.: 17.s. *. includes,3Ct- take the mo'a

/2 compute; semicontinci(sr:;, along-with relate. -.information" tt.c.litologies;...-Ltr.ogrammi-l'i:.;.,, automated MO :,;,:cicturing;i.:Tiplicationt- biolckinology 411-3 genetica2gineering..naterials wh4:c.e :properties cap, be tailoredfor desired esr1.2!9:ally poly-mers and

haF 5. the top in nlaTly counT!or years. Roboti-c 3 and ot',.ier fyr.;n3 of pro-

givinmable auton-17:. ve.:41ng ;,,,overnment7...4.7...:,ntion in West,.. : Ev-(1-.)po ?rid .1a1an, as well

through the U .1-7feft-is: Department. Bio-%:4:::chnology is el- vcrite et-iir,ple ofan industry that slit: -,uppor'4ed toreap -dividends lat,;4r.

ith origins in ap-plicationsarc. Eik.tik.'iy ic ploduc-tion volume. Such . '4. or am-plified upon almost t.. Juldbe found for medical technol.,..igt con-version devices, 4..-iculture.

Once past the gros3 selection of winning in-dustries, good policy decisions require careful

\analysisbut defining candidates for supportis not, in principle, an intractable problem. Ifthe chief objective is to stimulate economicgrowth, comprehensive support would not beneeded (as it might be for militarily criticaltechnologies). In electronics, good cases couldbe made for examples like the following:

continued development of device technol-ogies for high-speed, high-density ICs (gal-lium arsenide circuits and Josephson junc-tions as well as silicon-based devices);processes for submicron lithography;computer-aided circuit design methods;automated inspection of ICs and printedcircuits based on computerized pat-

. tern-recognition;automated generation of computer pro-grams .. togetlie.r with other method 3 for en-


hancing productivity in software genera-tion;natural language programing and relatedtopics in artificial intelligence;fiber-optics and, more broadly, integratedoptics.

Such a list. still quite general, would eventual-ly have to be refined further, a, normal whenplanning R&D. At finer levels, uncertaintieswill mount, technical judgments diverge. Fromthe Government perspective, this means pri-marily that payoffs in a number of areas mightbe possible, leading to strong arguments forsupporting competing technologies. Several ap-proaches to submicron lithography look prom-isingion beams, X-rays, electron beams; itwould be foolish for Government to "pick" oneof these.

One of the .:;plications of the discussionabove is that the Federal role might be primari-ly a matter of technology development. Howmight the Government design and implementprograms in support of commercial tech-nologies? The United States has extensive ex-perience in funding military research and en-gineering, but little background in civilian sec-tors; the principal exceptions are agricultureand energy, and the record in the latter is hard-ly flawless. One possibility is Simply tr findcompanies with expertise and good tra rec-.ords, then give them Government aid. Thiscould be research funding (ineluding furtherinitiatives such as the Defense Department'sinternal R&D program, which sets aside moneyfor industry-performed R&D on a "no-strings"basis to encourage innovation), a protectedmarket, Federal procurements, loan guaran-tees, direct grants of investment capitalthelist of possibilities comes froth chapter 10 (seealso table. 82).

From time to time, a number of European na-tions, as well as Japan, have followed policiesthat select companies for support. One exam-ple is CII-Honeywell Bull in France. TheFrench have also built their integrated circuitprogram around chosen firms rather than corn-

. petitive grants, although West Germany has,taken the latter. approach. Great Britain has

channeled funds and procurements to ICL,capitalized the semiconductor firm Inmos. InJapan, a good deal of political jostling goes intothe selection of participants for joint researchprojects such as the VLSI program or the fifth-generation computer effort. Experience in allthese countries illustrates the pitfalls ofcompany-centered support schemes. The Euro-pean record, in particular, has been poor.Siemens has garnered the lion's share of fund-ing in West Germany, with little evidence ofsignificant returns in the form of enhancedcompetitiveness to the German electronics in-dustry as a whole. Britain has recently beenforced to bail out ICL. Although Le Plan Cir-cuits Integres seems to be faring better,France's earlier Plan Calcul must be judged afailure.

Still, as in most of the countries experiment-ing with industrial policies, France appears tobe learning from its experience: Le Plan Calculsupported a single company, while the micro-electronics program has been structured to in-clude an element of competition among several-participants. In Japan, the record is rather dif-ferent. MITI excluded Oki Electric from theVLSI project, believing that the company couldnot compete in advanced integrated circuits.Oki prevailed on Nippon Telegraph and Tele-phone (NTT) for help, and managed to enterthe 64K RAM market. Given the multiplicityof competitive semiconductor manufacturersin Japan, thik, can hardly be judged a policyfailurebut might have been in a country witha thinner array of prospective entrants.

In any event, direct aid for selected firmswould not be an attractive option for theUnited States, going as it does against so manyof our traditional attitudes. It is a big step fromdropping the Government's long-running anti-trust suit against IBM to making that com-panyor any otherthe Nation's annointedchampion. Precompetitive support, the ap-proach taken by the European Community'sEsprit programwhich falls more naturallyunder the next alternative for a U.S. industrialpolicy=would fit the American system better.

482 internati.on3 Corn,:ii:iveness in Electronics

Nonetheless, an industrial vilicy that fo-c:used on R&Dand perhaps technology dem-onstration, for mature industries as well asgrowth sectorswould begin with technicalquestions that can in principle be evaluated inrelatively straightforward fashion. DOD experi-ence with R&D contracts and procurements of-fers a model, at least for the case in which Gov-ernment is the ultimate customer. This last isa major difference between supporting a de-fense industry and a commercial industryand one of the chief reasons for sticking totechnology; product development for civilianmarkets is a much riskier and less certain un-dertaking. Nevertheless, DOD's recordif lit-tered with failures or partial failures in in-dividual programsdoes show that in an over-all way the Federal Government can supportand develop industries, particularly given pro-curement authority. On its results, the U.S.space program must also be judged a clear-cutsuccess. Of course, that the Nation is militari-ly strong, or that the space shuttle flies, doesnot mean that the processes involved in reach-ing these objectives have been efficient. For an'industrial policy aiming at economic develop-ment, however, efficiency is more urgent. If theultimate goals include raising the standard oflivingand this will always be one of the prin-cipal arguments in favor of an explicit in-dustrial policy for a country like the UnitedStatesthen improving productivity, economicefficiency, and international competitivenessbecome vital. Asa spur to efficiency, competi-tion for Government largess is a poor substitutefor the marketplace. This does not imply thattargeted R&D support for growth industrieswhere judgments can be made largely on tech-nical groundsmight be counterproductive somuch as that supports and subsidies goingbeyond technology could be.

Capital for InvestmentOf the variants of supports and subsidies,

channeling investment capital to selected in-dustries has attracted a good deal of attentionin the United States. The goal would be to en-hance the. competitiveness of industries thatmight be.either growing or in decline. Advo-

cates of such an approachin essence, urgingprograms that would function as developmentbanks or a Reconstruction Finance Corp.focus on the cost and supply of funds as a bot-tleneck for critical or growing Industries In-vestments in ironmaking or integrated steel-making, for example, have not been attractivein recent years; prospective investors can ex-pect higher returns elsewhere. If the steel in-dustry were judged critical, the Governmentcould step inas indeed it lnzs in a very limitedwaywith loan guaranty ,..3 or ot:er forms ofsubsidized capital: .7.onw-,.:sely. firms in someindustries might be expanding so rapidly thatthey have difficulty in financing expansionthe case described in chapter 7 for portions ofelectronics. Venture capital markets tend to bespotty; at some stages in their development, en-trants in high-risk sunrise industries may findthemselves starved for capital because invest-ors jtidge returns to be uncertain or-U.10 far inthe future. ,z

If capital constraints pose-genuine problemsfor industries judged vital-to U.S. interests, theFederal Governmentyrnight indeed choose torespond with mech-anisms such as a Recon-struction Finance Corp. or a publicly backedinstitutional supplier of risk capital. But arecritical industries starved for funds? In coun-tries where capital markets are less developedthan in the United States, they may bepar-ticularly where venture financing is hard tocome by or simply unavailable. West Ger-manyeven Japanhas experimented withgovernment-financed venture capital pro-grams, as has Great Britain with its NationalEnterprise Board.

That governments in some countries inter-vene in capital markets does not imply thatpublic sector decisionmakers can do a betterjob of balancing risks and rewards than thosein the private sector, but that the governmenthas different criterianamely, that the publicwelfare is paramount for government, ratherthan private returns to capital. Indeed, govern-ments can set priorities ranging from main-tainence of the defense base to employmentstalaility or calming political turmoil. National

tTh 12Federal Policies Affecting Electronics: Options for the United States 4th'S

defense and employment have been particular-ly strong motives in Europe.

in the United States,-too, Federal assistanceto troubled firms has occasionally taken theform of loan guarantees or similar forms ofcapital subsidy. An overextended corporationin competitive difficulty will sooner or laterfind its access to financing cut off: examplesinclude Lockheed, Chrysler, Mc Louth Steel,Braniff, international Harvester. The FederalGovernment aided the first two; Mc Louth hasbeen rescued, at least temporarily, by a privateinv, stor; part of Braniff s aircraft fleet has been/repossessed: International Harvester's fate re-mains in its own hands. If the Federal Govern-ment decideson whatever groundsthat es-tablished enterprises which have fallen on hardtimes are indeed essential, precedents exist forbail-outs on a case-by-case basis. Should itregularize procedures for bail-outs? What aboutthe more general situation? Should the Govern-ment take steps to make capital cheaper or

\ more easily available?\ If the Government were to channel invest-rnent . funds to growth sectors, one reasonwould surely be to interject criteria other thanthose. applied in capital marketse.g.., somebroader notion of the public interest., ratherthan simply financial returns; where this is tbncase, capital subsidies are but one among manypolicy tools that Government might choose. Onthe othe.r hand, it might he that the market doesnot do a good job of evaluating long -term op-portunities where rewards are far in the fjture,risks high.19 In many instances of new technol-ogies or entire new industries, social returnshave exceeded private returns, creating a par-ticularly potent argument for Government ac-tion where the developinents in question wouldhave a multiplier effect on productivity or com-petitiveness elsewhere in the economye.g.,microelectronics or biotechnology. Innovating

11Tbe market failure r q.11 mtn: 'or government interventionin capital (and other) rear,e4r;:sq.../11-'ee.d in US. industrial Com-petitiveness: A Cocv,.. =rel. Electronics, andAufu:nobilw,, researchwhere vir-tually by definit:-.Jo the c&,:not be captured by theperforming en'. ,:r.rhapc the plainest case.

firms in growth technologies or growth indusa-tries may not, for a variety of reasons, be ableto capture all the rewards of their work. In theextreme, they may go out of business; econom-ic history is littered with examples of early in-novators %,-.,ho have failed, 13L: whoselater been picked up by others. This .1,,

the ways in which markets worksome inno-vators are a few years ahead of their time. Inother cases, a business failure may be quite un-related to new technology. Many of the pion-eering semiconductor firms have disappeared;while typically absorbed by other companies,the circumstances have occasionally been suchthat financial rewards were slim.

The case for an industrial policy that chan-nels capital toward long-range technologydevelopment or growth industriesespeciallyto sectors wher? the effects will spill broadlyover into the economy at large, giving socialreturns in excess of private returnsis thenquite different from that for subsidizing sec-tors having trouble competing for investmentfunds because of stagnation or apparentdecline. In the end, this second classdeci-sions on bail-outshinges, not on questions ofcapital markets, but on the justification forGovernment aid of any sort. If troubled indus-tries are judged critical, and deserving of sup-port, then capital preferences are one of severaltools Government can chose from. The high-risk, growth industry case depends largely onthe ability of Government decisionmakersevaluate social returns and spillever effer-and to determine when innovators are Ii'to go unrewarded because the nature of toe,:activities makes full capture of returns unlikely.Such analyses must be made on a case-by-casebasis. Given recent examples of venture fund-ing in microelectronics, computer software,robotics, and biotechnology, it is hard to arguein general that money for new and promisingstartups is not available; as pointed out inchapter 7, some observers have c,risk capital has bee:.with rather slim p:ost, cyclicalIty ofventure capital rni:;" :ts ther nestien, asis that of gaps at stages such as pre-startup.

48i-)

484 in:eri73!,onai Cornpetiti.'eness

In electronics. costs of capitalmore fun-darrintaliy. sources of capital to sustaingrowth at high rates 111 the face of rising capitalintensityare matters of concern primarily Iforestablished. companies that have already dem-onstrated their competitive ability. It would bedifficult for Government to justify subsidies inOle foriri of cap4a I allocations, low-interestloansnor loan guarantees for such firms.

In essence, that brings back One of the pointsraised initiallythe political context for selec-ting critical industries. Even for an industrialpolicy that devolved into a support scheme re-stricted to P.&D and technology development,politics will perturb decisionmakingdeci-sions made by business as well as Government.As the experience of the Defense Departmentshows, company-funded R&D will tilt towardareas where eventual Federal support is moreprobable.

So long as the goal is relatively clear to alli.e., military securitythe political dimensionscan be managed. National defense as an objec-tive of public policy generates little controver-sy; disagreements tenter on the means. On theother hand, if the objective is competitivenessor economic efficienCyparticularly in somenebulous futurethen the less-than-concretenature of this goal, and the :ntrinsic complex-ity of the supporting analysis, can easily con-tribute to obfuscation, confusion, and conflict.If the stakes are high, not only may politicallypowerful industries, if in decline, oppose pro-grams that reward industries judged critical,but the try to ..n.v that they are criticaltoo. Grov.ni industries might find themselvesfighting among one other for the biggest slicesof the pie. Cana board of experts inside theGovernmentor an advisory body includingrepresentatives of industry, labor, the financialcommunity, the public at largemake. deci-sions that will stick in such an environmnnt?An -in& trial policy advisory board" or"reconstruction finance board"retired in-dustry executives, leaders of Government andlabor, well-known academicsmight have theability to make good decisions, particularly if

bncked by a co inpetent sl tan' They would haveto he politic:0y sensitive simply to keep the ef-fort alive and pointed in the right direction. Butis it realistic to expect that, even if good deci-sinins lAreie made, they could be implementedgiven the political pressureswith any consist-ency? !If not, such a process would be a poorsubstitute or supplernent for U.S. capita/ mar-kets. On the other hand, ilGovernment sup-port is modest enough to avoid conflicts, willnot ,any positive 'J mpacts be equally modest?

Alternatively, ',nie,Government might choose.simply to protect critical industries from tradepressuresadopting an essentially passive pol-icy, rather than active support; one result mightbe to shift the riskireward expectations elfprivate investors. Protection for infant in-dustries is a common element in the industrial!policies of many countries, some of whomnotably Japanhave been accused of overdo-ing it. But in the end this is simply a varianton the more active approach, with most of thesame pitfalls. It assumes, first, that protectionwill baif rot essentialat least a positive fac-tor. Others would argue that exposure to for-eign competition stimulates a nation's own in-dustries, at least over the longer term. A .:our: -try attempting to develop an industry whereforeign enterprises are. already strong may havea good case for protection. Even the UnitedStateswhich is in a position to enter new in-dustries at the same time as its competitors ifnot ahead of themmight choose to protect in-fants if competing nations try to protect theirown. Absent this motive Land granting ti, ,, itis counterproductive, if not impossible 'toshield all industriespicking sectors to be pre-tented would create much the same set of prob=\,,lems as picking some to receive capital pref-erences.

'For a typical suggestion, see L. C. Thurcw, "Solving the Pro-ductivity Problem," Strengthening the Economy: Studies fn Pro-ductivity (Washington, D.C.: Center for Democratic Policy, 1981).p. 18.

4 8 /

72Federal Polic.ies Affecting Electronics: Options or the United States 485

intrastructural aild Adjustment PoliciesThe thrust of this policy alternativeof the

perhaps the hardest to summarize con-ciselywould be to create an environment thatwould aid private firms in strengthening theirown competitive ability. As table 82 indicaled,it would do so by relying preferentially onmeasures that support the infrastructure forindustrytechnology development (includingR&D. incentives for innovation, diffusion oftechnology within the domestic economy),human resources (education and training, par-ticularly in technical fields and including con-tinuing education and retraining), and struc-tural adjustment (measures that encourage mo-bility DI capital and labor, investments ingrowth industries, competition domesticallyand internationally).

By designing policy instruments that targetparticular industrial sectors only under specialcircumstances, instead relying preferentiallyon measures that affect,the economy in moreaggregate fashion--often policies that fall in thecategory of market promotionthe UnitedStates might avoid the pitfalls of an industrialpolicy with a strong sectoral thrust.20 Aimingto build future competitiveness, the role of theGovernment under this alternative would beto encourage beneficial change, while smooth-ing the negative ,impacts of adjustment.

Central to such an im'l!irial policy is a senset: aicsthe reality of change over time

in national economies, in the world economy.GovernMent policies that run counter to ongo-ing shifts in patterns of trade, competition, ortechnology are seldom effective in more thana Ina sejilsi:'; they rarely succeed in rover-

,..,ngoing transformations, although perhaps slowing them. They can aggravate the as-sociated dislocations. In contrast, policies thatwork in parallel witheven reinforceproc-

2.Market promotion policies are defined and discussed in U.S.Industrial Compfilitiveneg.9: A Comparison .......... Electronics.and Automobiles. op. cit., pp. 155fh.also pp. 175-182. Examplesinclude antitrust, support for R&D and innovation, plus policiesdirected at labor and capital marketse.g., for enbancin,_2 thei::obility of capital and labor in response to changing ecouunta.conditions. The latter are commonly referred to as adjustment

esses of economic and technological change.or that aim at smoothing adjustment and eas-ing dislocations, are more likely to have posi-tive effects. This third alternative for a U.S. in-dustrial policy flows from recognition thatcomparative advantages shift over time, withthe result that some industries in some coun-tries will thrive while others decline. Often thearc of growth or contraction is obscured byshort-term fluctuations; sometimes declinesprove temporary, expansion resumes. The U.S.textile industry is a case in point; the emer-gence of synthetic fibers provided an oppor-tunity or revitalization through new invest-ments that greatly increased productivity.

As the textile example illustrates, new tech-nologies are one of the forces that can sparkrenewal. Rather than trying either to anticipateor counter them, governments can accept thereality of such shifts and work toward max-imizing their positive impacts, minimizing thenegative. Public policies that function in thisfashion include:

Aid and stimulus for the development ofnew technologies, which might range frommoney for R&D to improvement of the pa-tent system.Better mechanisms for the diffusion oftechnology to industry, pr,rtir-,0 1,smaller comi.11 :Js; one possibility is awork of federally supported centers withthis mission.Tax incentives or other aid for firm's,install manufacturing tnchnolo- airat irnpri's prOdU ity an.. con,,tiveness .vhether new production proc-esses or those that are well-proven; ex-amples range from microprocessor-con-trolled heat treating furnaces to robots.Support for training and retraining of em-ployees displaced by economic .changethose in blue- and grey-collar ranks, as wellas professionals; this might entail encour-agement of company-sponsored continu-ing education programs. as will as policiesthat would support training ai. I retrain-ing irrespective of the boundaric of par-ticular companies Or industrial sedors.Improvements iii vocational-technical ed-

483

456 Iniernational Competitiveness in Fecrronics

ucation at the post-high school level, withparticular attention to skills that will beneeded as a result of predictable changesin the composition of U.S. industry--e.g.,computer-aided drafting and manufactur-ing, service and repair of electronic sys-tems.Continued emphasis on high-quality engi-neering eduction backed by renewedFederal resource commitmentsin fieldssuch as_ electrical and. computer engineer-ing, materials science and engineering, de-sign for automated production, and thewide range of other specialties that will beneeded for continued growth in high-tech-nology industries.In particular, renewed emphasis in univer-sities. supported by Federal funds, on en-gineering design, and on manufacturingengineeringaimed at upgrading the qual-ity of the work force in theSe professionsand bringing them more fully into themainstream of the engineering sciences.Tax and other policies aimed at increas-ing the rate of capital formation, morespecially at encouraging investments inemerging or rapidly expanding industries,as well as investments in R&D and : ri man-ufacturing technologies that 1,vir, increase,productivity in industries already well es-tablished.

Depending on the el sign of the instruments,such a list could al,m fit quite comfortablyunder several of the Other policy orientationsdiscussed in this chapter.

Infrastructural SupportHuman resourcesdefined broadly to in-

clude management styles and techniques thatmaximize the contributions of individual ern-nloyeesa7e crucial for ompetitiveness. Anyindustry depends on the skills and abilities ofthe people it employs; chapter 8 outlined thecurrent problems in technical education in theUnited States, as well as the general declinein technical lite: icy among the public at large.Education and training are traditional domainsof public policy. Declining emphasis on tech-nical and scientific training in American

schoolsas well as high unemployment along-side unmet demand for those with skillspointdirectly to problems calling for Federal action.Among the questions to be faced are: Whatshould people be trained in, beyond the obviousneeds for at least minimal competence in read-ing, writing, and mathematics? How can re-training best be accomplished? Within indus-try? Through community colleges and voca-tional educational programs? Whatever the re-spons::, it must incorporate a foundation forcontinuing learningon the job and offif peo-ple are to keep pace with advancing technol-ogy. Widespread public attention focused onsuch matters over the past year or two, togetherwith new initiatives emerging from Congressand the executive branch, are positive signs;the danger remains of a response that willprove too little and too late.

Tax policies can create incentives for privateindustry to train or retrain workers, engage inPAD, invest in new producti-m facilities. Still,incentives alone do not alwa sufficeone ex-

ni being long-term basic research of theAint that undergirds industries like electronics.Only the larger firms find it in their self-interestto support much basic research; the foundationfor the semiconductor industry, for instance,cam, in considerable measure from Bell Lab-oratories. However, the unique circumstancesthat caused Bell Labs, first, to perform a gooddeal of basic research, and, second, to help dif-fuse the results, seem hound to change asAT&T restructures and adapts to its new cir-cumstRnces. Other large electronics companiesIBM, Texas Instruments, C 3neral Electricalso perform substantial amounts of basicwork, although it has been less accessible tc.the R&D community at large. At various times,Governincnt laboratories and Government-funded research have made significant con-tributionscurrently, the more basic elementsin the Defense Department's VHSIC program,as well as the $20 million per year that the De-fense Advanced Research Projects Agency isfunneling into gallium arsenide.

Despite these examples, the level of-researchthat supports the U.S. electronics industry isless than adequate. This is shown most graph-

48,,

Cn. 12 Fecera! Po!icies Affecting Electronics: Options for the United States 487

ically by plans. originating vithin the industryitself, for joint R&D. Both the SemiconductorResearch .Cooperative, organized by the Semi-conductor Industry Association to fund univer-sity projects. and Nlicroelectronics & ComputerTechnology Corp. (NICC). an independentprofit-seeking venture. are aimed at similarneedstechnology development that will bene-fit a range of firms. The aims are to avoid ex-cessive duplication, help :-"use research re-sults, and undertake projects with longer timehorizons than individual companies feel theycan afford. Still. it is not at all clear that suchefforts will fill the researchas opposed to ad-vanced developmentvacuum. For example,MCC will concentrate initially on four areas:computer-aided integrated circuit design; com-puter architectures, especially itiose designedwith artificial intelligence in mind; productivi-ty improvement techniques for software gen-eration; and interconnections and packagingfor microelectronics devices.21 Three of these,if not all four, are bell removed from the basicend of the sprectrum. Likewise, the Semicon-ductor Research Cooperative has announcedMans to develop prototype large-scale RAMSan effort quite divorced from basic research.In any event, as part of its industrial policy theFederal Government could find positive waysto aid such joint research efforts; if direct as-sistance were not forthcoming, at least the Gov-ernment could takes steps to see that publicpoliciese.g., antitrust enforcementdo nothinder R&D that could be vital for the compet-itiveness of U.S. industry.

Antitrust, one of the fundamental varietiesof market promotion policy, is indeed show-ing signs of strain in the United States. As goodan example as any is the seemingly pervasiveconcern of business executives that behaviorthey regard as innocuousfor instance, mul-tifirm R&D efforts such as MCC proposes toundertakewill be subject to antitrust corn-.plaints. More fundamentall:, when U.S. an-titrust laws were drafted, most economic com-petition was a purely national affair; now inmany industries it is worldwide. When Amer-

"C. Barney. "R&D Co-op Gets Set 'Co Open Up Shop," Elec-tronics, Mar. 24. 1983. p. 89.

ican firms seek to cooperate in R&D, IyhatIveight should he placed on cooperation as aresponse to foreign joint R&D activitiessanctioned by governments and often fundedby them as well? The case for trying to reducethe duplication of effort accompanying simul-taneous pursuit of similar R&D objectives is.of course, strongest at the basic research end,fading as development is approached. In indus-tries like semiconductors and computers, com-panies typically want to compete at thedevelopment end of the spectrum, and Govern-ment in the United States has encouraged this;in these highly competitive fields, Americancompanies find it difficult to cooperate andprobably always will (Japanese firms are notdissimilar). Nonetheless, antitrust enforcementseems to be a constant in business complaintsover Government regulation, and a real bar-rier-- although perhaps as much psychologicalas legaleven when firms desire to cooperateonly in basic resrfarch. The guidelines on jointR&D published by the Justice Department in198J have done little to lower this barrier.22Moreover, the point at winch cooperation inR&D moves from being efficient and produc-tive to inefficient and counterproductive willbe industry- and technology-specific. Neitherthe Department of Justice. nor the FederalTrade Commission seems very well preparedto deal with such questions.

The Federal Government can i-210 play an im-portant role in stimulating industrial develop-ment by helping ensure an open trading envi-ronmentsomething individual firms are ill-equipped to do on their own. Open trade wouldcomplement this policy alternative as well asthe last two to be discussed. Indeed, for thenext alternativesupport for U.S. firms export-ing or operating on a worldwide basisitwould be the centerpiece. In contrast, for thepolicy approach under discussion here, export

22See U.S. Industrial Competitiveness: A Comparison of Siert!,Electronics, and Automobiles. op. cit., pp. 184.185. for a reviewo' antitrust law and enforcement, including the joint researchguidelines. As pointed out earlier in this chapter, insight intoexecutive branch intentions concerning antitrust enforcomentmust often be gleaned from sources which go unpublishede.j;;.. speeches at bar association MI' 'sings.

4U

4a6 (.7:1i=r,137,0Pal Con7pe...;:r.i:-...:ness in Elestmnics

success would be ranked as but one among anumber of goals.

Government effort to reduce trade barriersdirect and indirectcontribute in immediatefashion to structural adjustment. An industrialpolicy intending to promote competitivenessshould press for fair treatment of U.S. firmsthat export or invest overseas, as well as forvigorous competition within domestic markets.Thus, trade policies could take their placealong with adjustment measures aimed at fa-cilitating flows of capital and labor from staticor contracting industries to those with goodprospects for expansion and future competi-tiveness.

AdjustmentIn many ways, facilitating structural adjust-

ment lies, together with technology develop-ment, at the heart of this alternative. Adjust-ment policies are those that encourage move-ment of resources within the economy in re-sponse to market signals, as well as mitigatingnegative impactson sectors in decline, groups.of workers affected by shifting competitivenessor technological change, particular communi-ties or regions. While the United States has experimented with a variety of such measures inthe pastranging from Trade Adjustment As-sistance (TAA) for employees who lose theirjobs because of import competition to the manylocal and State development programs aimedat attracting new industryfew of these havefunctioned well. In particular, measures in-tended to aid workers or communities suffer-ing from adjustment woesTAA, administeredby the Department of Labor, the CommerceDepartment's Economic Development Agen-cyhave come to be widely regarded as fail-ures.23 This is one reason the current adminis-tration has turned away from Federal effortsat adjustment, arguing that marketsand thoseaffected by themshould be left to their owndevices.

--"Economic adjustment programs in the United States are brief-

ly reviewed in U.S. Industrial Competitiveness: A Comparisonof Steel, Electronics. and Automobiles. op. cit.. pp. 155-156.

There is no question that many Federal in-itiatives aimed at easing adjustmentincludingTAA, which functioned largely as a form ofsupplemental unemployment insurance ratherthan a positive aid to those seeking new skillsand new jobshave been less than successful.But the argument for falling back on the mar-ket, leaving those affected to shift for them-selves, is weak; the people involved have littlecontrol over economic events or impacts. Theplight of the individual is far different from thatof the corporation. Rationales for adjustmentassistance are well-accepted; they are groundedboth in improved economic efficiency and insocial equity.24 It is true that market inecha-nisms will suffice for economic adjustmentin the long run and in an overall sense. How-ever, the problems that adjustment policies areintended to remedy exist on a micro-levelrather than in the aggregate. While U.S. ex-perience with job training and retraining hasnot always been positive, the experiences ofother countries (ch. 8) demonstrate that man-power policies can function effectively. Ifoverall employment levels are a major objec-tive, adjustment policies can play a mediatingrole between growing and declining sectors.

Consider the situation of an assembly workerin a color TV plant. As figures 57 and 59 inchapter 9 indicated, while employment levelshave been declining in color TV, they have con-tinued to rise in semiconductor product5But while the consumer electronics indi..,iryis concentrated in States like Illinois and In-diana, semiconductor firms have tended to lo-cate in California. Since assembly labor in bothindustries is essentially unskilled, employersdraw on local labor pools. It would make littlesense for someone in Chicago who has beenput out of vt ork because of automation or for-eign competition to move to Silicon Valley

"Ibid.. pp. 177-179. The efficiency argument is based largelyon barriers to mobility that keep people from moving to seekwork, also on the friction that retards wage declines in responseto changing market conditions. The equity argument, in simplestform, holds that those who bear the brunt of adjustment sufferfrom causes outside their control while others prosperalso forreasons quite independent of their own decisions; under suchcircumstances, society as a whole has good reasons for easingthe burden.

491

Ch. 72Feceral Pcbcies Affe.cting Electron:'cs: 0;;; Ions icr tre United .9..a-:es 459

Photo crdlt CIrcInat, M,:acro!

Robot transferring refrigerator compressorsfrom one assembly conveyor to another

unless that person has specific skills that arein demand in the California labor market. Eventhen, relocation is a major hurdle. What sortsof assistance might Federal policies provide inthis case? The obvious possibility is trainingin skills .,or which there is current need; evenin the absence of relocation assistance, peoplewould then have greater incentives to move tolocations where jobs were available. The Gov-ernment might operate the training programor simply provide financial assistance to thosewho could not finance schooling on their own.

Such programs can aid adjustment without .

introducing economic distortions...Much thesame is true of supports for technolov devel-opmInt and diffusion. Targeting the base andinfrastructure for competitive industries-rather than targeting industries themsclvesCa 11 contribute to economic efficiency withoutexplicitly favoring some sectors of the econ-omy. In this view, technology development anddiffusion encompasses much more than simplyR&D support. Indeed, diffusionencouragingfirms to utilize available technologies, par-ticularly manufar:turing processes that improveproductivitymay, for many parts of the econ-

1 1 0 - 3 - 32

omy, be more vital than support for the devel-opment of Pri!IV technologies.

Driving forces for technology diffusion andutilization vary dramatically across industries.Sectors at the forefront of technological changeand international competitionsemiconduc-tors or commercial aircraft rather than con-sumer durablesmust and do take advantageof the latest technical knowledge. There is lessimpetus in industries that are growing slowlyor contracting; in the steel industry, Americanfirms have often failed to install the latest pro-duction equipment, although this would saveenergy, improve lab& productivity, and cutcosts. One reason is simply that alternative in-vestments promise higher returns. Yet it maynot be vise, from the viewpoint of the economyas a whole, to wait until the need for more ef-ficient production equipment mounts to veryhigh levelsi.e., until payback periods areshort. Manufacturing firms that lag in movingtoward programmable automation or comput-er-aided designperhaps because pressures toimprove productivity and competitivenessbuild slowly at firstmay at some point findthemselves overwhelmed before they are ableto react.

From a Government perspective, then, theprimary objective Of structural adjustmentpolicies is to encourage resource flowstech-nological, human, material, capitalto themore productive and dynamic sectors of theeconomy, while providing assistance to work-ers and regions suffering from deterioratingcompetitiveness. For the United States, marketpromotion policies seem best suited to fillingthis role, but other countries have sometimesemphasized sectora'; measurespicking win-ners and promoting them, for exampleas amean:, to "positive adjustment." This is mucheasie; in simple economies, such as those ofthe newly industrializing countries.

Design and Implementation

What would be the likely effects on the U.S.electronics iuiustry of an industrial policyoriented toward adjustment and infrastruc:uralsupport? If bne intent of such a policy is to en-

49 ;-2


courage growing industries, the more dynamicsectors of electronicscomputers (particular-ly smaller systems and software), semiconduc-tors, instrumentation, roboticswould be log-ical beneficiaries. Electronics typifies ongoing

ructural shifts in the U.S. economy:growth in services, ranging from elec-tronic banking and electronic mail to theproduction of computer software itself,many of these services made possible bycheap computing power;increasing relative demand for skilledworkersthose with manual skills, as re-quired for building, maintaining, and re-pairing advanced production equipment,and those with mental skills, as requiredof integrated circuit designers; andgreater capital and R&D intensity associ-ated with high technology.

Other growth sectors share similar character-istics.

The major assumption underlying an indus-trial policy oriented toward adjustment and in-frastructural strengthening is that Governmentis capable in a general way of identifying thesources arid impacts of economic change anddesigning policy measures that will speed thepositive consequences while ameliorating thenegatives. Government need not depend onsector-specific policies to accomplish this;much can be done with market promotionmeasures and other policies with aggregateobjectives.

Industrial policies that call on Governmentto pick and choose among the sectors of theeconomy risk political defeat or deflection; fur-thermore, they depend on the ability of policy-makers to devise programs tailored to par-ticular sectors without gross sacrifices inoverall economic efficiency. There are strongreasons for relying on market mechanismswhere possible. Nonetheless, in many circum-stances market forces alone are inadequate forachieving legitimate goals of public policy.Several of the cases were mentioned above: na-tional defense, long-terin basic research. Othertimes, Government actions may interfere withthe operations of markets. Indeed, one of the

fundamental tasks for industrial policymakersis to determine when markets are working welland when they are not.

The task of devising policy measures appro-priate to this, third alternative for a U.S. in-dustrial policywhat OTA has elsewheretermed macroindustrial policymust thereforestart with a strengthening of the Federal Gov-'ernment's analytical capability. Nowhere inCongress or the executive branch is there nowthe expertise to grapple with the evolving dy-namics of industries or markets domestically,much less internationally. As in the case of acritical industries approach, the Governmentwould need to begin by improving its abilitiesfor identifying patterns of change, understand-ing the forces driving them, and formulatingpolicy responses that would lead to desiredpolicy outcomes. This is not an easy task, butit is certainly not impossible. Such a capabili-ty will be esseni ial if U.S. industrial policiesare to be redirected to support growth sectors,which almost by definition evolve in unex-.

pected directions. As many examples in theshort history of computers and microelec-tronics show, such industries follow paths thatare full of detours and surprises. The reactiveapproach of the past, with Government policiesare mostly responses to short-term economicand political pressures, is far from optimal.

Indeed, even if the goal were to defer asmany decisions as possible to the private sec-torthe last of the five alternatives to be dis-cusSedthe Federal Government would stillneed a basis for deciding which responsibilitiesto retain. In any economy as complex as thatof the United States, Government decisions in-fluence business activities in many waysoften indirectly, and sometimes inadvertently.At the minimum, they do this through taxation,plus monetary and fiscal policies. To the ex-tent that policymakers'grasp the probable im-pacts of alternative courses of action, they canprovide an environment that encourages inter-national competitiveness. Any of the five policyperspectives outlined in this chapter thereforeimplies an improvement in the Federal Govern-meht's capability for analyzing industrial coin-

49j

Ch. 12Federal Policies Affecting Electronics: Options for the Unitedtates 491

petitiveness and the effects of public policy onthe activities of the private sector. Otherwise,industrial policy will be made in the future asit has in the pastby defaultand other con-siderations will take precedence over compet-itiveness, productivity, and economic efficien-cy. In the absence of such analysis, successfulimplementation of a coherent and consistentindustrial policy of any stripe would have tobe judged something of an accident.

Promoting the Global Competitivenessof American industries

An industrial policy directed at building theworldwide competitiveness .of U.S. industries-might tie regarded as an extension of the long-standing thrust by this country toward opentradea policy that would entail, not only con-tinuing pressure to reduce tariff and nontariff

barriers in all countries, but also active en-couragement of exporting and foreign invest-ment by AmeriCan firms. Such an industrialpolicy would differ from the others discussedin this chapter first in its outward rather thandomestic orientation. Drawing on past ex-amples of industries that have expanded rapid-ly while marketing aggressively on a worldscaleAmerican manufacturers of computersor aircraft rather than steel or consumer elec-tronicsa globally oriented approach to indus-trial and trade policy would be based on thepresumption that active participation in mar-kets all over the world is a primary route tomaintaining competitiveness. Some advocatesof such a1policy would contend that if the U.S.consumpr electronics and steel industries had,in fact,/moved more decisively to export andinvest Overseas during the 1950's and 1960's,they would have been better positioned tomaintain their competitiveness during the1970's. Worldwide marketing and Sales, alongwithmultinational production, are then viewedas central elements of this policy alternativewhich is based on the premise that the mostcompetitive industries and firms are those thatprepare themselves to compete in the &be]markeiplace.

The United States has been a leader in themovement toward an open world trading sys-tem since the later years of the depression.After the passage of the Smoot - Hawley Act in1930, tariffs steadily decreasedfrom levelsnear 50 percent, to the range of 5 percent (ch.11). Following the war, as the Marshall Planhelped to rebuild the Western European andJapanese economies, U.S. international eco-nomic policy was directed at promoting "freetrade" through multilateral agreements suchas GATT. This country provided much of theimpetus for the establishment of GATT, andhas almost always supported its efforts to lowerbarriers to international commerce; open mar-kets have been viewed as an important objec-tive of U.S. foreign policy, a vibrant worldeconomy as central to the postwar political sys-tem.

Product Cycies and Structural AdjustmentThis approach to industrial policy would take

as a starting point the fact that some sectorsof the economy, and some firms, will be bet-ter able to compete than others. Implicit arenotions of product cycles and trade restructur-ing. The constant' pressure of internationalcompetition, along with other forces acting onthe world economyparticularly technologicalchangecreates a dynamic of shifting compar-ative advantage. Manufacturers in countries atthe leading edge of a technology introduce newclasses of products first. In electronics, the ob-vidus examples include digital computers, col-or television, dynamic random access memory'chips, video cassette recordersthe first threecommercialized by American firms, the last inJapan. As such-products move through theirlifecycles, the technologies they embody be-come better understood, easier for competingfirms in other nations to duplicate. As a result,production costs grow more importantandmanufacture spreads to economies that' are notnecessarily at the forefront of the technologyThus, terminals and small processors for com-puter systems are now made in many coun-triesalthough often by subsidiaries of Amer-

4 9 4


ican or Japanese firms; but while a nation likeBrazil may have a burgeoning minicomputerindustry, this does not mean it will manufac-ture larger mainframes. A few years after dy-namic RAMs were introduced by Americansemiconductor firms, production was under-way in Europe and Japanby foreign manufac-turers, as well as the overseas subsidiaries ofU.S. multinationals; eventually, RAMs will beproduced in countries like Hong Kong, SouthKorea, and Taiwan. The spread of color TVproduction has followed similar patterns; here,the comparative advantage of the United Stateshas slipped further than for RAMs, and Amer-ican firms have been able to maintain theircompetitiveness only by transferring manufac-turing operations overseas. VCRs for consumeruse were developed by Japanese firms, but asthe technologies involved diffuse, productionwill begin in other parts of the Far East; it hasalready started in Korea.

Product cycles in most industries follow sim-ilar patterns; the common feature is specializa-tion of production in parts of the world fa-voredat a given timeby comparative advan-tage. Thus the United States emphasizes agri-culture and technology-intensive manufacturedgoods among its exportsalong with services.Where wages are low, labor-intensive productsare among the more competitive; to exploithigh technology, countries need a well-trainedwork forcewhich normally will be well paidby world standards. An open system of inter-national trade and investment is intended toallow product cycles to follow their naturalcourse, with nations specializing in what theydo best. Adjustment problems represent thedarker side of the picture.

One rationale frit. an avowedly global U.S. in-dustrial policy is simply the persistent concernthat strains in the international trading systemwill undermine that system's openness. Themost visible sign of strain is the proliferationof national industrial policies that, among otherthings, tend to protect local industries whilediscriminating against efficient producers inother countries. Another is the frequency ofrecourse to bilateral trade negotiations and

agreements, rather than the multilateral ap-proach of GATT; prominent examples in theUnited States have included OMAs for colorTVs. Western European nations have seldombeen as committed to open trade as the UnitedStates, and disputes over steel, textiles,automobiles, and consumer electronicsailingsectors in Europe as in this countryhave ledsome observers to voice concern over revivalsof protectionism, even trade war.

Slow and painful structural adjustments liebehind many of these pressures. Industries inadvanced nations with large and complexeconomies seldom respond very quickly tochangeincreasing wages, escalating raw ma-terial and energy costs, technological advance,challenges from abroad. As living standardsrise and social welfare programs proliferate,countries facing the need for rapid adjustmentfind that sudden and sharp dislocations bringequally swift political reactions, rather than themore or less resigned acceptance of earlieryears and more primitive economies. Tradebarriers are an easy response.

The Relation Between Open Trade andIndustrial Policy

A global approach to industrial policy by theUnited States would find a natural anchor inthe GATT system of multinational agreements.Absent special circumstances such as indus-tries calling for protection, nations have tendedto prefer the multilateral approach over bi-lateral negotiationsfor consistency and tominimize discriminatory impacts on some na-tions. Advocates of a global approach stress thegains that producers in all countries can makeif free to develop their own strategies, combin-ing domestic and foreign resources in an openmarket system. A common corollary is to min-.imize restrictions on flows of technology, withbarriers limited to those motivated by nationalsecurity and arms control. Antitrust policiesalso fit naturally into on industrial toolicyoriented toward open trade and competition.Cartels -and monopoliesinternational or do-mesticareamong the classic examples ofmarket distortions. Because an industrial pol-icy centered on open trade is motivated ulti-

36


mately by faith in market mechanisms, anti-trust would be an essential element. Domesticantitrust policies, along with multilateralagreements fostering competition, would C3M-plenient reductions in trade barriers.

What would this fourth alternative for a U.S.industrial policy then look like? It mightembody:

international trade agreements, on a mul-tilateral basis, aimed at further opening ofworld markets and at keepiu them open;measures intended to ensure equal treat-ment of firms from all nations seeking toexport or to invest beyond their borders;standardization of customs and other na-tional regulatory proceduresprOductstandards, as well as those those dealingwith exports and imports; andcompetition policies aimed at preventingmonopolization and cartelization in bothdomestic and world markets.

As the list implies, nontariff and indirect bar-riers to trade would need a good deal of atten-tionas indeed they will regardless of the di-rection of U.S. industrial policy. Measures suchas those listed would have generally favorableimpacts on the U.S. electronics industrypar-ticularly if genuine success were achieved indismantling nontariff barriers. Portions of theindustry that are already highly competitivewould be helped the most.

Promoting U.S. Trade Competitiveness

What elsebeyond essentially passive meas-ures aimed at opening marketswould beneeded for an industrial policy that encouragedthe gobal competitiveness of American in-dustry? Compared to the early postwar yearswhen GATT was organized, the environmentfor international trade has changed markedly.At that time, the economic and politicalstrength of the United States was literally over-whelming. The United States was able to pushits alliessome, such as Jap^n, rather reluc-tantlyinto the international system. Japan'sreaction was to establish a new set of govern-ment supPorts for domestic industry in an-ticipation of trade liberalization, but at least

that countryand many othersmade a com-mitment to membership in the internationaltrading community.

Now, over 30 years later, political and eco-nomic power are more widely dispersed: theUnited States is still first among trading na-tions, but without the preeminence it once pos-sessed. Forging international agreements ismore difficult in a multipolar world. The elec-tronics industry is no longer the province ofa handful of technologically advanced Westernnations, but the battleground for increasinglyintense competition involving industrializingcountries as well. With the traditional leadersexhibiting quite understandable concern, rap-idly expanding economies in the developingworld look both to invade the markets of ad-vanced countries and to protect themselvesfrom those a rung or two down on the ladderof economic advance. As nations at all levelsadopt government policies in support of theirown industries, severe trade frictions can easilydevelopparticularly when overall growthslows. In essence, the current system of inter-national trade is suffering its own adjustmentproblemsit was conceived in a different era,and is showing unmistakable signs of age.

More concretely, negotiations of past.yearscovered matters on which it was easier to reachagreementprimarily tariffsthan those of to-day. In the Tokyo Round, still lower dutieswere achieved. In a few instances, renegotia-tions on a bilateral basis have hastened reduc-tionswitness Japanese concessions on tariffsfor semiconductors and computers. While thisprocess could certainly be pursued furtherand might be expanded to include the Euro-pean Communitymany tariffs are already atlow levels. As parity is approached, attentionshifts to areas less amenable to internationalagreement: government procurement policies,R&D subsidies, indirect barriers. Here, discus-sions between the United States and nationslike Japan have borne less fruit.

The protracted discussions over the procure-ment practices of NTT illustrate some of thesecomplexities. After months of negotiation anddebate NTT agreed, in 1981, to open biddingto foreign firms, but American companies have

496

494 International Corri.'litiveness in Electronics

not had much success in selling to the corpora-tion. While the Amei.icans tend to ascribe thisto informal barriers, the Japanese say U.S.firms are not trying hard enoUgh.25 Even roughsymmetry in public sector procurement poli-cies can be difficult to achieve when both cor-porate and government practices differ amongcountries. AT&T's decision, mentioned earlier,to give the Boston-Washington fiber-optics con-tract to its own subsidiary, Western Electric,is a case in point. Fujitsu, which entered thelow bid, may (or may not) have offered a quota-tion below its reasonably expected costs in or-der to gain access to a rapidly expanding mar-ket. If it did so, the tactic is hardly unknownto firms outside Japan. AT&T's action wastaken after intense lobbying efforts within the

Government centering on claims that giv-ing the work to a foreign enterprise wouldjeopardize national security.2° After each suchoccurrence, it becomes more difficult for theUnited States to convince other governmentsthat open trade is intended as a two-way street.Direct military procurement is, needless to say,an even more sensitive subjectone where na-tional interests will necessarily remain para-mount.

Given such considerations, a logical first stepmight be disCussions on product standards andcustoms prodedures, where differences tend tobe visible and political controversy less intense.This is .not to say that agreements would beeasy or reach' very far; the nations of the worldhave never been able to agree on standards fortelevision broadcasting, &metric power, or-which side of the road to drive on. Interna-tional discussions extending over many yearsaimed at settling on common designs for elec-trical outletS were abandoned in 1982 when itbecame clear that agreement would be impossi-ble. Still, continuing progress in reducing non-tariff barriers can be expectedalbeit slow andpainful. Many of thesee.g., government pur-chasing policiesare perceived as largelydomestic issues; after all, people often feel thattheir own industries should be favored.

25See R. Neff, "NTT's Open Door Draws No Crowds," Elec-tronics, Dec. 29, 1981, p. 58.

'26"Japan Runs Into America, Inc.." op. cit.

Export promotiona recurrent theme in de-bates over U.S. trade policyis another facetof the global approach to industrial competi-tiveness. Export incentives offered by theUnited States have often been criticized asweak and ineffective compared to those ofother countries.27 All trading countries employexport promotion measures of one form or an-other, even though these have generally beenviewed as detrimental to a free and open trad-ing systemparticularly when they involvesubsidies, as opposed to activities that functionas advertising or related marketing aids. Sub-sidized export credits have been particularlycontroversiale.g., the low-interest financingthat Canada's Government offered to NewYork City for the purchase of subway cars (chi.11).

The United States has recently taken a num-ber of positive steps to help exporting firms.The Export Trading Company Acteasing re-strictions on bank participation as well as pro-viding protection against antitrust suits forfirms that enter export joint ventureswhichbecame law at the end of 1982 is one example.Estimates of the extent to which this act willhelp American exports and create new jobsvary considerably.28 Consideration has alsobeen given to finding replacements for theDISC (Domestic International Trade Corpora-

"See, for example, Export Policy, hearings, Subcommittee onInternational Finance. Committee on Banking. Housing, and Ur-ban Affairs. U.S. Senate, especially Part 3, Foreign GovernmentPolicies and Programs to Support Exports, Mar. 9, 1978, Part6. U.S. Programs and Facilities Designed To Support Exports.Apr. 5, 1978. and Part 8, Oversight on Foreign Barriers to U.S.Exports, May 17. 1978. Also Export Stimulation Programs inthe Major Industrial Countries: The United States and Eight Ma-jor Competitors, prepared for the Committee on InternationalRelations, House of Representatives, by the Foreign Affairs andNational Defense and Economics Divisions, CongressionalResearch Service. Library of Congress. Oct. 6. 1978: H. L.Weisberg and C. Rauch, "A /Comparative Study of Export In-centives in the United States,/ France, the United Kingdom. Ger-many and Japan," Internatimial Division, Chamber of Commerceof the United States, Washington. D.C., 1979: and R. A. Flam-mang, "U.S. Programs That Impede U.S.. Export Competi-tiveness: The Regulatory Environment." Center foi Strategic andInternational Studies, Georgetown University, Wathington, D.C..1980.

2°C. H. Farnsworth, "Measure Expected To Spur Exports,"New York Times. Oct. 5. 1982. p. D5; R. E. Taylor. "Law To En.-courage Joint Export Ventures Is Expected To Be Signed by -,Reagan Today," Wall Street Journal, Oct. 8, 1982, p. 12. A par-ticular aim is to help smaller companies wishing to export.

Ch. 12Federal Polic,:s At doting Electronics: Options for the United States 495

lion) mechanism discussed in the proceedingchapter. DISCs have been determined to violatetJ.S. obligations under GATT; and tax incen-tives that !night have comparable impacts onexport competitiveness have been proposed.2"In the United States, as concern over apparent-ly slackening competitiveness has mounted,many in Congressas well as the businesscommuttityhave also called for changes suchas modifica lion of the Foreign Corrupt Prac--tices Act. The resistance during 1981 to pro-posals for scaling back the Export-Import Bankillustrates the importance that many place ona more active 'approach to promoting U.S.exports.

Still, export promotion is a limited tool. Theroots of international competitiveness lie inslomestic industryin the efforts of privatefirms to design, manufacture, and marketgoods. How these firms adapt to the realitiesof shifting comparative advantage and chang-ing competitive circumstance outweighs gov-ernment policies aimed at encouraging exportsunless these policies function as subsidies ofsubsta,ntial magnitude relative to the costs ofthe goock in question. Even then, no govern-:nont can promote all exports all the time. Inthe longer term, therefore, export promotionseldoinc'has major effects on trade competitive-ness. Of course, in a given case it may makeall the difference: promotional measures canhelp firths and industies in temporary difficul-iy they can be useful as a means of equalizingef.n,-IpeC:ion by matching the efforts of othergovernicit:ot:z; they can help private industry

a foo!hohi ii rooA, markets. But export pro-raotott e,.,rse. the tides of competitive

I. prec.;:s- ia point that an industrialainled at cimoting the global compe-

titiveness of U.S. industries would have toconfrontand on whichit might founder. Anation can certainly promote its industries; butno matter how extensively it does so, all its in-dustrieS cannot export at once. There will al-ways be winners and losers in world trade. Nstrategy aimed at promoting fair and open

""Administration's DISC Substitut Bill Introduced in I3othHouse, Senate, U.S. Export Wee*, August 9. 1983, p. 685.

global competition implies that the mix ofAmerican firms able to take advantage of op-portunities in the world marketplace wouldchange over timeperhaps rather swiftly. Italso implies involvement of foreign firms inU.S. marketsthrough direct investment aswell as exports. More so than the other fouralternatiVesand especially a domestic marketpreservation strategy, which would take pene-tration by foreign firms to be, in and of itself,cause for concerna global approach placinghigh priority on market access for entrantsfrom all nations could be politically difficultto implement. As pointed out earlier, whenfirms and their employees in declining indus-tries combine, their influence can outweighthat mustered by the friends of open trade. Thenegative implications for some sectors of theAmerican economy might be difficult for anavowedly global U.S. industrial policy to dealwithparticularly given the poorly developedadjustment mechanisms the Nation has inplace.

The United States is already experiencing theconsiderable hardships that cities, regions, andoccupational groups face when industries losecompetitiveness slowly, as happened with theAmerican steel industryor, even worse, rap-idly, as in the automobile industry. Whether ornot these declines are, permanent or transitory,the hardships are debilitating, and an industrialpolicy encouraging open world trade couldbring such changes more quickly. The primaryargument against a global promotional strategythen lies with these short-term negative im-pacts; extensive promotion of U.S.Industrieswithout better methods for dealing with ques-tions of adjustmentcould place a heavy bur-den on those sectors unable, for whatever rea-sons, to compete effectively. In the long term,a global strategy might increase economic op-portunities at the aggregate level,. but, in themeantime the price could well be judged toohigh. This will be particularly true to the ex-tent that economic growth is slow; rapid ex-pansion gives companies, employees, and com-munities adversely affected by rising foreignCompetitiveness a broader array of alternatives.

493


Progress toward an open environment forworld trade has never come easily; today,, thepace of change may have picked up, interde-pendence risen, but the basic arguments infavor of trade between nations have not altered.The fundamental assumption underlying thisfourth alternative for a U.S. industrial policyis that an open world trading system is in thelong run interests of all nations. In the UnitedStates, despite periodic bouts of protectionistrhetoric, both parties have generally supportedthe propositionflowing directly from notionsof comparative advantagethat if each coun-try devotes its efforts to goods for which it is,relatively speaking, an efficient producer, neteconomic welfare will be maximized. Providedthat world trade is not greatly impaired by tariffand nontariff barriers, the exchange of prod-ucts and services among nations will permitpeople everywhere to attain standards of liv-ing that are as high as their resource endow-ments and state of development. permit. An in-dustrial policy based on this premisea prem-ise as true today as a hundred years agocouldbe viewed as an extension and reinvigorationof traditional U.S. attitudes.

An Industrial Policy Centeredin the Private Sector

A fundamental reason why there has beenno coherent or consistent industrial policy inthe United States has been the widespread be-lief that corporate executives rather than Gov-ernment officials have not only the ability butthe right to make decisions that affect businessactivities. While many disagree with this view.,the political power of organized labor, consum-er groups, and others who advocate a strong-er Government role has had more impact onrelatively narrow questions such as rules forcollective bargaining or environmental protec-tion than on matters of trade and competitive-ness.

One of the more pointed indications of thestate of Government-business relations in theUnited States is the attitude of the businesscommunity toward the Department of Com-merce. Nominally the center of advocacy for

business interests within the Federal Govern-ment, the Commerce Depaitment is a weaksister among Cabinet agenciesnot becausecorporations in America are weak, but becausebusiness and industry do not take the Depart-ment very seriously, and often bypass it.

At a time when some Federal officials joinwith spokesmen for industry in anti-Govern-ment rhetoric, the feeling that public agenciescan do nothing right naturally grows. A morepositive' view might acknowledge that per-formance varies in both private and publicsectorsthat the ups and downs of an Inter-national Harvester or a Chrysler Corp. may notbe all that different from the ups and downsof a Government agency. Nonetheless, thereare political and institutional realitiesmanyreviewed in earlier sections of this chapterthat must be altered if the Federal Governmentis to design and implement a more coherentindustrial policy. The most pressing need is fora better developed understanding among Fed-eral agencies of how industries actually func-tion. Advocates of this fifth and concludingpolicy alternative believeat least implicitlythat Goveri,inent cannot hope to succeed atthis, and should not try; they want to "get Gov-ernment off the backs of industry," and leavethe private sector free to compete witl. mini-intim interferelce.'°

Of cuurse, some Federal involvement in theaffairs of industry will always exista mini-mal level is necessary, indeed is one of the rea-sons governments exist. But advocates of anapproach maximizing private sector responsi-bility for industrial policymaking argue that the "-narrower and more limited the Government'srole the better. Beyond the posturing that af-flicts such questions, the argument becomes:Government involvement in economic affairs

."When polled, corporate managers in the United States andJapan respond very differently on questions dealing with govern-ment "planning." When asked whether their economy wouldbenefit from: 1) more Government planning; 2) about the sameamount; or. 3) less planning, 90 percent of American managersresponded that less 'Government planning is called for: in Japan,

response was evenly divided among the three alternatives."Perspectives on Productivity: A Global View," American

and Foreign Attitudes on Productivity, hearing, Committee onthe Budget. U.S. Senate. June 3. 1981, p. 64.

49 j

Ch. 12Fedrr,ll Policies Affecting Electronics: Options for the United States '497

is counterproductive because it distorts marketmechanisms; governments too often subsumeor override the economic: ratio.. sales for privatechoices, on both supply and demand sides.'"I.Juder these circumstances, economic efficien-cy decreasesto the presumed detriment of all.From this perspective, an industrial policywhether intended to encourage economicgrowth and development or, at the otherextreme, emphasizing regulations and con-straints on business activitymust seem boundto weaken U.S. competitiveness. A relatedargument holds that Federal regulations costindustry and the public treasury more than thesocial gains set against them. The most extremeview is held by those who argue that any Gov-ernment action impairs market mechanismsand hinders efficiency; amore moderate atti-tude grants the Government a place where mar-ket imperfections can be unequivocally dem-.onstrated. A still more centrist perspectivetheone to which the rest of this section is di-rectedhOlds that the;; appropriate role for Gov-eminent lies in creating a climate conduciveto economic growth', giving industry access tothe tools for its own development. In many re-spects, this attitude is a traditional one in theUnited Stater:, vinwing macroeconomic policy-makin,,-4contrui the money supply, taxation,Federal spendingas legitimate, but otherwisebelieving that Government intervention in eco-nomic affairs is to be tolerated mostly as a last`resort.

Entrepreneurship has been a driving forcefor American industrial development, withbusiness and Government coexisting ratheruneasily, but an industrial policy that woulddefer where posible to the private sector doesnot, them,iMply that Government plays no roleat all. Public policies aimed at promoting capi-tal formation would be consistent with suchan approach; so would, presumably, regulationof business practices widely considered.unfairor predatory, export promotion measures of at

"One of the more thoughtful of this viewpoint isRedefining Government's Role in the Market System: A State-ment by the Research and Policy Committee of the Committeefor Economic. Development (Washington. D.C. Committee forEconomic Development. July 1979).

least some types, and a trade policy that other-wise. supported American business interestsoverseas. What this altern live rules out firstand foremost would be atwmpts to develop sec-tor-specific policies targeting key industriesbe these housing or semiconductors or energy.

Such an approach to industrial policy wouldbe consistent with recent emphasis on reduc-ing Federal spending and trying to controlbudget deficits, scaling back regulations, andcutting corporate taxes. This policy directionwould, ideally, expand the financing availablefor dynamic and competitive firms while leav-ing them free to make their own business de-cisions. It would, at the same time, avoid sub-sidies for declining. industries or firms, just asit would eschew attempts by public officials toselect and support growth industries. To, those_favoring such a policy, it is the best hope ofthe United States for maintaining its interna-tional competitiveness into the future.

Central to.this policy option might be tax andother measures aimed at capital formation. Therationale for the Front), Recovery Tax Actof 19f31 (ERT A . I. !t..is. Governmentpolicies aimed at promoting savings and invest-ment were held to be the engines of growth forreviving the U.S. economy. As discussed inchapter 7, there is little evidence as yet thatERTA will serve this purpose, but its passagecould be taken as a sign of movement in thedirection of an industrial policy that wouldleave corporations, to their own devices.Nonetheless, such a policy direction .wouldhave to be judged one that, rather than mov-ing away from the fragmentation characteriz-ing past U.S. industrial policies, reinforces thisfragmentation. The reason is simple: withoutany evident justification, ERTA increases dif-ferentials in tax treatment across sectors of theeconomy (ch. 7). /

On the other hand, movement toward dereg-ulation of business activities represents a shifttoward greater policy coherence to the extentthat real progress is made in cutting back thetotal number of regulations, the conflicts thatmay exist among these, and the number of

498 International Competitiveness in Electron:cs

Photo credit: Universal Instruments Corp

Automatic assembly machine for placing leadless carriers on ceramic substrates

agencies that administer them. Of course, tobe effective in bringing a greater degree of con-sistency and harmony to regulatory policy, anymove toward deregulation must proceed with-out creating a set of 50 differing regulatorypolicies at the State level.

In years past, Government regulations grad-ually developed into a set of constraints onprivate enterprise thatdespite undeniablepositive benefits -have often been'judged in-efficient. While the adverse impacts of regula-tory policy on the competitiveness of Americanindustries have frequently been overstated,regulations have had at least a small effe.7,t in

50i

dampening overall rates of productivitygrowth.32 Even so, when Federal regulationsare examined sector by sector, there are fewcases of large and unambiguous adverse im-pacts on competitiveness. More often, the ef-fects of regulation have been"of the same gen-eral magnitude as for other pUblic policieshaving positive as well as negative effects ondifferent sectors. different comparies. At bot-tom, the most cogent criticism of Federal reg-ulatory policy is simply that individUal meas-ures have too often been implemented Without

32E. F. Denison, "Explanations of Declining ProductivityGrowth," Survey of Current Business, August 1979,P. 1.


any explicit attention to the consequences forproductivity, competitiveness, and economicefficiencyi.e,, lacking even a rudimentarybalancing of benefits, against costs, This ishardly surprising given that most though notall regulatory policies are directed at objectivesquite unrelated to competitiveness: clean airand water, safe products, minimization ofworkplace hazards. It may even be true thatregulatory policy as a whole has come to beconfused as well as sometimes inefficient, andthat regulatory rulings have hindered industrialdevelopment while not always being effectivein their avowed purposefor instance, protect-ing consumers. Certainly, advocates of de:illation cast U.S. policies in an unfavorablelight compared with approaches sometimestaken by other countrieswhere regulationsmay even be used, on occasion, to encourageeconomic development. Broadly speaking,.however, most advanced industrial economiessubject private business to regulatory re-quirements rather similar to those in the UnitedStates. Japan, for example, has instituted en-vironmental protection measures that are re-strictive by any standard. Nonetheless, manyin American busiriess community continueto 'Call for a rollback of safety and environ-mental regulations, as well as antitrust enforce-ment.

To some extent, this simply mirrors tradition-ally adversarial relations between business andGovernment in the United Statesattitudesthat in some cases must share responsibility fordeclining U.S. competitiveness. For instance,the differences in response among Japaneseand American automobile firms faced by reg-ulation of exhaust emissions and fuel economyin the U.S. market were striking. As might beexpected from examining the corporate strat-egies employed by Japanese firms in other in-dustries, automakers like Nissan, Honda, andToyota have looked at regulations as new op-portunities for finding a competitive edge.Within corporate headquarters, Japanese ex-ecutives may regard Government regulationjust as bleakly as their counterparts in De-troitbut they attempted to make the best ofthe situation, as a long string of new model in-troductions beginning in the 1970's attests.

As the discussion above implies, financial,tax, and regulatory issues would no doubt com-prise the core of a business-centered industrialpolicy. Tax reductions and deregulation havebeen at the head of corporate agendas for years.In the context of electronics, tax policy is muchthe mere important, regulations having seldomhad much impact. Chapter 7, on financial is-sues, shows the ability to fund expansion in theface of rising capital intensity to be one of thekey uncertainties for rapidly growing electron-ics companies. Fast-paced technic-A changemakin^. manufacturing equipinc t obsoles-cent: iogetiier high costs of design anddevelopment and rising levels of foreign com-petition create new financing pressures incomputers as well as semiconductors. Otherrapidly growing sectors of American industry,particularly where technology moves quidkly,can expect similar problemssteinming in part,from the common desire of U.S. managers tofinance growth with internally generatedfunds, as well as the declining role of stockissues as sources of financing for Americancorporations.

Would an industrial 'policy that cut taxes andreduced regulations, leaving other matters tothe business community, help high-technologysectors? The answer hinges on how they wouldfare compared with other portions of the U.S.economy. In terms of taxation especially, theissue comes back to differential affects. Taxpolicies, even when designed to be neutralacross industries, will never fully achieve this.The depreciation schedules enacted by ERTAare only one example. These will probably helpother sectors more than electronics, simply be-cause many electronics manufacturers wereable to depreciate production equipment quiterapidly under the old law; their capital costrecovery periods have sometimes been short-ened, but not nearly as much as in heavy man-ufacturing or primary metals. Firms earlier re-quired to depreciate newly purchased assetsover many years get much greater benefits interms of internally generated cash flows fromERTA. They may also find their ability to at-tract capital from external sources enhancedrelative to electronics firms.-Furthermore, ac-celerated depreciation tends to benefit corn-

502


panics with substantial profits rather than newand growing concerns that may still be spend-ing more than they take in. The larger comput-er firms, for example, may be helped more thansoftware vendors or semiconductor manufac-torers. Consumer electronics, where profitshave been low for years, is not likely to gainmuch.

Is the w4ample provided by ERTA typical ofwhat be expeted from an industry-cen-tered policy approach? If only because olderand larger firms and industries tend to havemore accumulated political power, the answeris probably yes; these sectors of the economymight be able to skew the policy process totheir advantage, with newer industries suffer-ingif only in a relative sense. This is a majorliability of an industrial policy that would deferwhere possible to business interests. Advocatesof this policy orientation must be prepared toaccept outcomes like the altered depreciationschedules in 7.,RTA. Policy directions wouldContinue to be determined largely by thepolitical process, and the greatest rewardswould probably go' to the sectors thattogetherwith their employeescould muster-the great-.est political strength. These are likely to beolder, well-established industriesparticularlythose whose employees are unionized. Wheresuch industries are suffering from internationalcompetition, they will seek tr .3hape policy inways that preserve their markets, profits, andjobs.

Of course, the electronics industry has beenactive and successful in its past lobbying ef-forts, and would be able to look out for itselfunder an industry-centered approach. ERTAlegislation included a number of measures that

electronics firms had actively sought, includingthe R&D tax credit and changes in tax treat-ment of income earned by Americans workingoverseas. These offer direct benefits to tha elec-tronics industry, particularly the R&D provi-sions. In addition to the tax credit, which per-mits a wrIteoff amounting to 25 percent ofspending for R&D above a base figure, equip-ment used in research can be depreciatedfaster. Deductions are also allowed for appa-ratus and equipment donated to universities.These measures are scheduled to expiry; in1985; until then, at least, they will assist firmsin portions of the industry with extensive R&Dactivities.

A further point that an industrial policy fol-lowing this approach would have to confrontis the extent to which firms pursuing economicself-interest may neglect objectives importantto the Nation as a whole. Basic researchthesort that does not promise immediate pay-offsL--provides one example. Nor is it likely thatthe health and safety of either the labor forceor the public at large would be served by anindustrial policy that deferred product andworkplace standards to industry. Regional im-pacts, along with questions of adjustment as-sistance for displaced employees are additionalcases where an industrial policy too heavilyoriented toward the desires of the businesscommunity might be perceived by other seg-ments of society as inadequate.

In the end, the question comes down to this:If other countries are developing ambitious andcomprehensive programs to support certain oftheir industries, can the United States assumethat absence of Government action is t.le best .

response?

Summary and ConclusionsThe competitive situations of the U.S. con-

sumer electronics, computer, and semiconduc-tor sectors differ greatly, but -they do have com-mon features. How then to summarize the pol-icy implications? The technological and marketleads of American electronics firms are nar-

rowing. Manufacturers in Japan especiallyhave successfully followed strategies based onselecting particular market niches, establishingthemselves in these markets, then expanding.This was their mode of entry into the U.S. con-sumer electronics market, it has allowed

Ch. 12Federal Policies Affecting Electronics: Options for the Uniied States 501

to deeply penetrate the American inn tket forsome types of ICs, and is the approach niey willfollow in computers. It has also helped Japa-nese firms to compete effectively in other partsof the world. Electronics companies in japanhave been aided by their goveinment, althoughthe form kind impact of the aid has varied agood deal across the industry; still, the pro-grams generated in recent years to stimulatethe expansion of high-technology sectors injapan, Western Europe, and several of the new-ly industrializing countries show a degree ofconcern for industrial development far out-stripping that in the United States.

At this juncture, U.S. electronics firms facenot only heightened competition, but also prob-lems in financing continued expansion and infinding well-trained people to fill their staffs.The industry exemplifies the structural trans-formations taking place in the U.S. economy:ever-growing requirements for skilled laborand creative management; dependence onR&D and the commercialization of new tech-nologies in order to establish new markets orretain old ones; rising capital intensity in theprocess technologies necessary to enhance pro-ductivity or simply to make state-of-the-artproducts; foreign competitors supported by theindustrial policies of host governments. Thereare a multitude of problems ahead for the U.S.electronics industry, and for others at the fore-front of economic developmentbiotechnol-ogy, robotics, communication and information'technologies. Congressional interest in the in-ternational competitiveness of the U.S. elec-tronics industry stems in part from the modelit provides for other kcy sectors.

At the same time, it would be misleading tooveremphasize the problems faced by indus-tries like electronics. U.S. capital markets con-tinue to function well. American semiconduc-tor firms have made rapid strides in improv-ing the quality of their products. The industryis still the world leader in technology, thoughnot so farnor so consistentlyahead. Policiesfollowed by the Federal Government haveaided American electronics firms by opening...-world markets for makers of computers andsemiconductors. Only infrequently have Fed-

eral policies been clear and direct obstacle -; toefforts by the industry to improve its competi-tiveness.

What is missing are the links between the bitsand pieties of Federal policy that affect thevarious portions of the electronics industry.Government policies cannot and will not trans-form this industry or others: the private sec-tor has provided the driving force for pastdevelopment, a pattern that will continue. Butpublic policies help create the environmentwithin which competition takes place, they setales, frame decisions. Industrial policy could

provide a setting conducive to capital forma-tion, R&D, education and training, free mar-ket competition. To the extent that Governmentpolicies support technological developmentand structural adaptation, they work in thelong-term interests of American industry andthe American labor force. A more coherent andconsistent industrial policy could make asignificant contribution to the competitive posi-tion of the U.S. electronics industry.

In the United States, industrial policy stillmeans different things to different people. Tosome, industrial policy is viewed much likesupply-side economics was several years agoas an untried theory. To others, it suggestsgovernment support for "sunrise" industriesor trade protection for threatened sectors likesteel or textiles. Some have argued that theAmerican political scene is so disorderly thatany attempt at a more consciously developed\industrial policy would be pointless if not',counterproductive. Despite the seemingly in- \cessant debates over the successes and failuresof industrial policies in Japan or Britain orTaiwan, all such views miss the essential point:industrial policymakin.5 is a routine activity ofall governments. In the United States, we cancontinue to leave industrial policy to the ran-dom play of events, or we can try to improvethe system.

Politics lies at the heart of finding a moreconsistent and coherent approach to industrialpolicy for the United States. The starting Pointis to recognize that industrial policy decisionsare being made all the time. The problems ofAmerican companies in consumer electronics,

504


automobiles, or clothing and apparel were notcreated by Government policy, but the absenceof a coherent approach to industrial policy hasvirtually guaranteed a devolution to special in-terest politics. Faced with seeming chaos in thepolitical arena, many have simply thrown uptheir hands. This implies accepting as inevita-ble long and torturous courses of events in in-dustries like color TVwhere final outcomesof trade complaints going back to 1968 haveyet to be determined, or steelwhere claimsby the AmeriCan industry of dumping by for-eign enterprises go hack at least to 1959. It alsoimplies relying on the blunt instruments ,.ifmacroeconomic policymaking Neither sup7ply-side economics nor public pump-primingof years past offer plausible remedies for thecurrent dilemmas of American industry. It iscertainly true that deregulation, lower rates ofinflation, and higher rates of overall economicgroWth will help a wide range of U.S. indus-trieS, but urgent needs such as technologydevelopment and diffusion, education andtraining for displaced workers, and seed capitalfor entrepreneurial businesses also call for at-tention by Government.

An industrial policy response following oneof the alternatives discussed in this chaptercould represent an attempt to find concretesolutions to particular problems. Such aresponse needs to be based on careful examina-tion of the situation of American industry ata given point in time. Advocates of a more co-herent industrial policy for the United Statesunderstand that Government decisions -affectthe activities of industry in many and often sub-tle ways' they would encourage policymakersto include competitiveness - and technologydevelopment more explicitly in the objectivesof policy; more consistently in its formulationand implementation. At the broadest level ofgenerality, this implies a "vision" of long-term

economic development interposed in the policyprocess; it means creating political constituen-cies for industrial policy rather than standingby while the myriad of interested parties attempt to promote their own typically narrowand short-term designs.

There is no doubt that improvement is possi-ble; policymaking can be a purposeful activi-ty characterized by learning from past experi-ence within a framework of empirically basedanalysis. Developing a more effective industrialpolicy must begin in this spirit, while recogniz-ing that the process is inherently political andalways will be.

Although a variety of policy instrumentscould be used in pursuit of industrial policy ob-jectives, in the U.S. context, it appears thatspecial stress should be laid on manpowertraining, R&D and technology diffusion, plusmeasures aimed at stimulating investment innew and innovative firms and an open environ-ment for international trade and investment,Such policy initiatives, emphasizing structuraladjustment, would help in building foundationsfor international competitiveness in electronicsand other industries.

The form that such an industrial policy mighttake would have to be determined by Congress,along with the executive branch and the manyinterest groups with a stake in the outcome. Tobe effective over the longer term, industrialpolicy must be based on practical understand-ing of the functioning of the economy on asector -by- sector basis, with forward-lookinganalysis of both problems and prospects. OTAhas outlined five alternative approaches to thistask; more than anything else, an effective in-dustrial policy for the United States requiresa clearer view of where industrial deVelopmentin this country is headed, and of the Federalrole in aiding this development.

Appendixes

APPENDIX A

Glossary*

acceptable quality level (AQL).The fraction ofdefective items permitted in a group of parts (e.g.,integrated circuits) that pass the statistical sam-pling tests agreed to by manufacturer and pur-chaser. An AQL of 0.01 percent, for instance,means that no more than 1 defect per 10,000parts, on the average, is allowed.

Pcz.zds time.The time required to retrieve the con-tents from a specified memory location in a com-puter system, commonly the average time to fetcha bit from an integrated circuit memory chip.

active device.An electronic component that cancontrol or regulate an electrical signal. Example3include most kinds of vacuum tubes, as well astransistors.

actuator.Causes mechanical force 'or motion inresponse to an electrical signal. Examples in-clude hydraulic cylinders (in conjunction withother control 7ystem components) and electricmotors.

A/D converter (analog/digital converter).A cir-cuit that transforms analog electrical signals intothe equivalent or proportional digital represen-tation, commonly so that they can serve as inputsto a computer system.

advance pricing (also termed forward pricing).Setting prices for manufactured goods below cur-rent costs in the expectation that,these costs willfall as production experience accumulates andscale increases. A common pricing strategy forintegrated circuits, where advance pricing hasbeen used to gain early market share, sacrificingimmediate profits for those in the somewhatlonger term.

analog.Refers to electrical signale that can varycontinuously over a range, as compared to digitalsignals, which are restricted to a pair of nomi-nally discrete values.

architecture.The overall logic structure of a com-puter or computer-based system.

binary.Number system in which all valueg arerepresented by combinations of a pair of sym-bolse.g., "0" and "1." In contrast, the familiardecimal or base 1.0 system represents all numbersby combinations of the 10 symbols "0" to "9."

bipolar.Transistors or integrated circuits in-which-electrical conduction takes place through

the motion of both negative and positive charges.

Many of the technical terms included in this glossary are explained inmore detail. often with examples. in chapter 3. See, in particular. appciiodixes 3A and 3B at the end of that chapter.

The negative charges are electrons, while thepositive consist of the absence of electrons wherethese negatively charged particles would ordi-narily be (the electronic vacancy, or "hole," leadsto a net positive charge). Bipolar ICs are fasterthan unipolar (MOS) chips, but not as dense.They also dissipate more electrical power.

bit.A binary digit, which can take on one of twovalues, typically written as "0" or "1."

bite (byte).A group of binary bits, usually 8.bubble memory.A solid-state microele-tronic

device in which binary data is stored in tiny mag-netic domains (bubbles) given one of two possi-ble polarities. Bubble memories, are nonvolatilebut not random access; to read or rewrite datain a given location, a string of bubbles must bemoved past a detector to reach the desiredmemory location. Thus the access time dependson where in the string with respect to the detec-tor that location happens to be.

bus.Circuit path by which electrical signals movebetween components (e.g., microprocessor andmemory chip) or between circuit boards.

capacitor.Passive uircnit eLament that stores elec-trical charge creating a voltage differential. Ca-pacitors can be fabricated within integrated cir-cuits, as we as in the form of discrete com-ponents.

captimA semiconductor manufacturer whoseoutput ryes primarily for intracorporate con-sumptinm

chip.An integrated circuit, either partially or fullycompleted.

circuit board.A card or board of insulating ma-terial on which components such as semicon-ductor devices, capacitors, and switches areinstalled.

clock.An electronic circuit, often an integratedcircuit, that produces high-frequency timing signals. A common application is synchroni2ationof the operations performed by a computer or mi-croprocessor-based system. Typical clock ratesin microprocessor circuits are in the megahertiranee, 1 megahertz equaling 108 cycles persecond.

core memory.Computer memory in the form ofmagnets -that can-have-one-of-two_statesi_thus_-__enabling storage of binary data, Now largelyobsolescent as a result of integrated circuitmemory chips.

CPU (central processing unit).The portion of a

507 505

506 International Co' ,;:etitiveness in Electronics

digital computer where logical operations areperformed under the direction of softwareinother words, the 131-.)rtion of a computer systemwhere the program is Executed.

CRT (cathode ray tube). ;Displays an image on ascreenas in a TV set or computer terminalin response to electrical signals.

D/A converter (digital/analog converter).Anelectrical circuit that changes a digital signal intoequivalent or proportional analog form. See A/DConverter.

dedicated.A piece of equipmentr;.g., a comput-er processorreserved for a single function, suchas aircraft flight control or the operation of a mi-crowave oven. Dedicated processors are oftenembedded within a more complex system so thatthey are invisible to casual users. This is the case,for example, with the computers used in elec-tronic banking terminals.

digital.Refers to equipment or systems whichoperate on electrical signals that can be repre-sented as strings of binary bits. In electronic cir-cuitry, a pair of nominally discrete voltage levelscommonly stand for the two possible values asso-ciated with each bit.

disk drive. Computer peripheral in which data isstored magnetically on a rigid or flexible (floppy)disk. In many cases, disks can be removed andreplaced. A drive unit rotates the disk beneathmagnetic heads for reading, erasing, and writingdata; a typical drive also includes circuitry andcontrol mechanisms for locating data and for in-terfacing with the processor.

distributed processing.Refers to computer sys-tems in which two or more CPUs are intercon-nected, with processing (program execution) car-ried outor distributedamong the linked CPUsunder the control of system software.

dop4ig.=Adding to a semiconducting materialsmall amounts of other elements (dopants) tochange its electronic properties.

electronic. Refers to devices, components, orsystems in whicii electrical signals are used pri-marily to convey and manipulate information.

etching.--In semiconductor fabrication, surfacesare etchede.g., with an acidto selectively re-move material.

European Community (EC); European EconomicCommunity (EEC).The EEC, established in1958 by the Treaty of Rome, joins Belgium, Den-

==mark-,-France,=West-Germany, Greece, Ireland,Italy, Luxembourg, the Netherlands, and theUnited Kingdom in a Cenimon Market.. TheEuropean Community links EEC, the European

Coal and Steel Community, and the EuropeanAtomic Energy Community. The Commission ofthe European Communities is EC's principalgoverning body.

European Free Trade Association (EFTA).Acommon market consisting of Austria, Finland,Iceland, Norway, Portugal, Sweden, and Switzer-land.

feedback control.Use of a signal generated at theoutput of a process to vary one or more processinputs so that the measured output is maintain-ed within a specified range.

fiber-optics.--Use of glass fibers to transmit lightfor communicating information as an alternativeto electrical signals: Typically, the light isgenerated by lasers.

firmware.Computer software stored in replace-able hardware components, generally integratedcircuit memory chips such as ROMs (read-onlymemory circuits).

floppy disk.A thin, flexible disk made of a plasticsuch as mylar and magnetically coated to be usedfor computer memory and data storage. See diskdrive.

forward pricing.See advance pricing.gallium arsenide.Semiconductor devices made

from this compound promise higher speeds thansilicon-based devices.

gate.A simple electronic circuit that can imple-ment a specified logical operation. In essence,gates act like switches. Computer processingunits depend on large assemblies of gates; as dointegrated circuit memory chips.

GATT (General Agreement on Tariffs and Trade).The nearly 90 nations which belong to GATThave agreed to work toward reduced barriers tointernational trade. The organization, the rulesof which also have the name GATT, serves as aforum for multilateral trade negotiations and fordispute resolution.

GDP (gross domestic product).The value ofgoods and services generated within a nationaleconomy, generally on a yearly basis.

GNP (gross national product).The value of GDPadjusted for revenues that enter and leave theeconomy as a resultof financial flows associatedwith foreign investments 'payments to foreign in-vestors are subtracted from GDP, revenues fromoverseas investments added).

hard-wired.Refers to an electrical circuit orsystem the operation of which is determined bythe hardware elementse.g., components andinterconnectionsand cannot be changedwithout changing the hardware configuration. In

App. AGlossary 507

contrast, the logical operations performed by acomputer or a microprocessor-based system canbe changed by loading a new software program.

input/output (I/O).Refers to provisions for enter-ing data into a computer system or for receivingoutput from the system. I/O devices are piecesof peripheral equipment with this purpose,notably terminals and printers.

instruction set.The group of logical operationsthat a microprocessor or computer can carry m .

integrated circuit (IC).An electronic circuit man-ufactured on a single substrate. Most ICs are pro-duced on small chips of silicon and are mono-lithici.e., fabricated on and within the chiprather than assembled to it.

integrated optics.Refers to devices in which in-formation is manipulated in the form of lightrather than electrical signals. Such devices mightbe interconnected by optical fibers (see fiber-optics).

interface.Circuitry, and often software, thatallows one piece of equipment to communicatewith another, as in the interface between a pairof computers or between a computer and a diskdrive.

Josephson junction, Josephson device.Madefrom superconducting materials (which must beheld at very low temperatures), these offer thepossibility of logic gates with very high switchingspeeds, hence very fast computers.

learning curve, experience curve.Graphicaldepiction of declines in manufacturing costs withtime Of with accumulated production volume.While often attributed to learning by factory per-sonnel and engineers, the curveswhich tend toshow rapid cost decreases for semiconductormanufacturing, leading to competitive strategiesbased on Fdvance pricingalso depend on manyother factors, including product design changes.

lithography.Processes related to photographyand printing by which patterns ale formed onsilicon wafers during the fabrication of in-tegrated circuits.

LSI (large-scale integration).Refers to integratedcircuits which contain of the order of 104 devices.

main memory, primary memory.The portion ofcomputer memory that the processor can addressdirectly (as opposed to mass or peripheral storageequipment such as disk drives). The main mem-ory normally holds the program being executedas well as the data being manipulated,

mainframe computer.A system, normally in-tended for general-purpose data processing, char-acterized by high performance and versatility.

Mainframes have grown steadily in capability assmaller and less expensive machines have pro-gressed.

mask.Stencil-like grid used in creating litho-graphic patterns on semiconductor chips.

mass storage.Refers to peripheral equipment forcomputer memory suitable for large amounts ofdata or for archival storage. Typical mass storagedevices are disk and tape drives.

mean time to failure, mean time between fail-ures.The common measure of reliabilityaver-age time between malfunctions that disable a sys-tem or substantially degrade its performance.Normally determined through statistical estimat-ing procedures.

microcomputer.Refers to integrated circuits thatcontain a microprocessing unit plus memory, aswell as to computers designed around microproc-essors or single-chip microcomputers.

microprocessor. An integlated circuit of whichthe major portion is a digitalprocessing unit.Microprocessor families consiit-of_groups ofsimilar chips each intended for a somewhat dif-ferent class of applications.

minicomputer.A small computer system, in-N.termediate in cost, size, and processing powerbetween a microcomputer and a mainframe.

MOS (metal oxide semiconductor).Oxide layersgrown on semiconducting substrates are used toform transistors and other circuit elements. MOSintegrated circuits are unipolar rather than bipo-lari.e., electrical currents are carried by either;positive or negative charges but not both.

packing density.VLSI circuits are packed moredensely than LSI circuits by making the individ-ual circuit elements and their interconnectionsphysically smaller. This has required steadyimprovements in manufacturing equipmente.g., for lithographyand careful control of theproduction process.

passive devices.Circuit elements, such as resis-tors, whose characteristics affect electrical sig-nals but which cannot be used to regulate or con-trol those signals.

peripheral.Equipment used in conjunction witha computer processor. Typical peripherals in-clude keyboards and terminals, mass storagedevices, and printers. .

peripheral chips.Integrated circuits designed tobe used in conjunction with particular micro-processors or single-chip microcomputers. Com-mon types include clock circuits, A/D converters,and keyboard interfaces.

PROM (programmable read-only memory).An

.rinQ


integrated circuit chip that stores data perma-nently. A PROM can be programed with this dataafter the chip has been manufactured; in-con-trast. data is stored in a ROM as part of the man-ufacturing process.

OEM (original equipment manufacturer).OEMsincorporate computers and other system compo-nents into their own end products. As an OEM,General Motors purchases microprocessors to beused in engine control systems and dashboardcomp uters.

quality.Measures the extent to which productsmeet specifications dealing with performanceand other functional parameters, and often ap-pearance as well. Quality is determined at a pointin timegenerally before the product enters serv-icein contrast to reliability, which is a measureof the ability to continue meeting specificationsover time. In most cases, quality is determinedby statistical sampling procedures based on datafrom testing and inspection that accompany orfollow manufacturing.

RAM (random access memory).Most common-ly, an integrated circuit that stores data in suchform that it can be read, erased, and rewrittenunder the control of a computer processor. Anymemory location in a RAM can be addressed di-rectly (random access) as opposed to sequential-ly or serially.

real time.Refers to computer operations thatparallel (in time) related external phenomena, asin real time control of industrial processes (seefeedback control). Real time processing oftenmakes heavy demands on the hardware elementsof a system, as well as the software.

register.Location in a computer processor wherebinary information is manipulated. Programs areexecuted by operating on strings of bits broughtto the registers.

reliability.Measure of the extent to which a prod-uct or system functions satisfactorily in service,commonly quantified as, the time betweenfailures that impair operation or degradeperformance.

resist, photoresist.Chemicals used in litho-graphic processing of integrated circuits which,much like photographic emulsions, can be ex-posed by light, X-rays, or other radiation to formpatterns.

ROM (read-only memory).Computer memory,typically consisting of integrated circuit chips,the contents of whichcan -be- retrieved at anytime but which cannot be changed by erasing orrewriting. Often used for program storage in mi-croprocessor-based systems.

second-sourcing, alternate sourcing.When onefirm designs- and develops a product that othersbegin to manufacture, the latter are referred toas second sources or alternate sources. Bothmilitary and commercial customers often insiston multiple sources for critical components suchas integrated circuits. An IC produced by a sec-ond-source supplier may be identical to the origi-nal design or it may differ in detail while remain-ing functionally interchangeable_Popular chipsare sometimes produced by a dozen or morefirms. Second-source manufacturers commonlynegotiate licenses or purchase technology fromthe originator.

semiconductors.Materials from which semicon-ductor devices are madee.g., silicon, germa-nium, gallium arsenideso called because theirelectrical conductivities are lowe than for goodconductors such as metals but higLer than for in-sulators such as glass. The devices themselves arealso called semiconductors. Discrete transistorsand integrated circuits are the most commontypes of semiconductor devices.

sensor.Converts a pressure, temperature, or otherphysical parameter into an electrical signal, oftenfor use in a control system. A digital speedometerfor an automobile transforms thebutput of a sen-sor into a miles-per-hour reading) as does anairplane's air speed indicator. In the case of theautomobile speedometer, rotary motion is con-verted into an electrical signal, while an air speedindicator depends on the pressure created by themotion of the airplane.

smart terminal, smart machine.In essence, asmart terminal is a small computer intended pri-marily for communicating with other, morepowerful computers. It can perform some proc-essing itself, in contrast to a dumb terminalwhich can communicate with a computer butcannot execute programs. A smart machinee.g., a numerically controlled lathecontains oneor more computer processors; these might bemicroprocessors or minicomputers.

software.Computer programs. More generally,instructions or procedural descriptions.

solid state.Refers to electronic components suchas transistors or integrated circuits in which thefunctions are carried out within a solid material(as opposed to a vacuum tube), or to systems (e.g.,'TV receivers) made with such components.

supercomputer.At a given time, machines at theupper limit of computing power--as measuredby computations per second or related measuresof performanceare called supercomputers.

system development.Software generation for

App. AGlossary 509

microprocessor-based systems is often referredto as system development. Commercially avail-able microprocessor development systems arefrequently used to help in the preparation of pro-grams that will implement the desired functions.Software accounts for a major portion of theengineering effort for systems that use embeddedor dedicated microprocessors or computers.

terminal.Generally includes a keyboard for dataentry, along with a display such as a video screenfor showing the input data, as well as output fromthe computer(s) that the terminal communicates

transistor.A solid-state electronic device whichcan control or regulate an electrical signal in re-sponse to a second signal, thus enablingamplification of the first signal.

TTL (transistor-transistor logic).Most commonof the families of bipolar logic circuits.

VLSI (very large-scale integration).Refers to in-tegrated circuits with of the order of 105 devices.64K RAMs are at the lower end of the VLSIrange.

VCR (video cassette recorder).Records and playsback TV images on magnetic tape. Common con-sumer models can record off the air or from videocameras. Also called VTRs (video tape recorders).

video disk.TV images mechanically encoded on

spinning disks can be played back using a stylusor laser beam. Current models cannot be used forrecording.

wafer.A disk of silicon (or other semiconductingmaterial) on which integrated circuit chips arefabricated. Today, wafers may be 4 inches ormore in diameter and accommodate hundreds ofchips.

Winchester disk.A hard disk for computer mem-ory v, ..ich rotates within a sealed enclosure andthus cannot be removed. Compared to removabledisks of the same diameter, Winchester drivescan store more data; however, they are not suitedfor archival storage.

word.The basic unit of informationhaving theform of a string-of binary bitsthat a computerprocessor works with. Typical word lengthsrange from 4 to 64 bits; more powerful machinesare generally designed to use longer words.

yield.In the production of microelectronicdevices, the fraction that survive all tests and in-spection, function correctly, and can be sold, orincorporated into the manufacturer's own endproducts. Production costs depend heavily onyields, which themselves depend on circuit de-sign, fabrication equipment, and control of themanufacturing process.

APPENDIX B

Offshore Manufacturing*

During the past two decades, many Americanelectronics firms have moved portions of their man-ufacturing operations overseas in search of lowerlabor costs. Offshore production has been a majorelement in cost reduction strategies, particularly inprice-sensitive portions of the industry such as con-sumer electronics -and semiconductors. Labor-intensive components and subassemblies for com-puters and many other products are also made inlow-wage developing countries. In electronics, asin automobiles, foreign investment has been amajor force in transforming national industries intointernational industries. Transfers of technology aswell as "capiial contribute to internationalization.

American electronics firms invest in overseasplants to serve foreign markets, as well as reimport-ing goods to the United States (ch. 4). The formerare often termed point-of-sale plants, the latter off-shore manufacturing or offshore assembly plants.It is offshore investment to serve the U.S. marketthat is the primary topic of this appendix. Otherarrangementsfor instance, subcontracting with /foreign firmswill not be covered-. Offshore man/ufacturing thus implies ownership and mane e-ment control by an American corporation. Virt allyall the major U.S.-owned consumer electronics andsemiconductor companies have offshore plants,mostly in Mexico and the Far East.

In both consumer electronics and microelectron-ics, the driving forCe for offshore investment hasbeen cost reduction. U.S. consumer electronicsfirmsprincipally television (TV) manufacturershave moved overseas to meet competitive pressuresand preserve existing markets (ch. 5). Foreign in-vestment has been largely a defensive tactic, a reac-tion to import penetration at home. In microelec-tronics, competition among U.S. firms has led totransfers offshore.

From the perspective of the United States, off-shore production has both positive and negative im-pacts. Compared to the plausible alternatives, thenet effects appear to be positive in most cases,much more so in the case of semiconductors.

This appendix is based largely on "Effects of Offshore and OnshoreForeign Direct Investment in Electronics: A Survey," prepared for OTAby R. W. Moxon under contract No:033-1400.

510

Economic Impacts of OffshoreManufacture

Offshore investments in electronics affect domes-tic employment, the balance of payments, nationalincome, and the future competitive abilities ofAmerican industry. The many studies of U.S. for-eign direct investment, while seldom focusing onoffshore manufacturing per se, yield insights intosuch investments. Even so, the evaluation of costsand benefits remains controversial, and the evi-dence gives no clear guide to public policy. Imme-diate impacts generally get the most attention,although longer term effects often prove quite dif-ferent than short-term consequences. Table B-1classifies the impacts.

Table B1.Possible Effects of OffshoreManufacturing Investments

Effects within the industry making the InvestmentsA. Domestic employment

1. Total U.S. employment in the industry (up or down).2. Changes in skill mix in the industry (increase or

decrease in blue-collar job opportunities, expansion inprofessional categories, etc.).

3. Regional employment shifts.B. Domestic value added by the industry

1. Changes in total wages and salaries paid to domesticemployees of the industry.

2. Profitability of companies in the industry.3. Tax payments by firms in the industry.

C. U.S. balance of payments1. Shifts in trade balance involving products of the

industry.2. Other current account flows.3. Capital account flows.

Effects in related industries (suppliers as well as customers)A. Domestic employment (with same subcategories as

above).B. Domestic value added (with same subcategories as

above).C. U.S. balance of payments (with same subcategories

as above).Longer term effectsA. Shifts in international competitiveness of U.S. indus-

tries.B. Changes in concentration and structure of U.S. indus-

tries.SOURCE: "Effects of Offshore and Onshore Foreign Direct Investment In Elec.

tronics: A Survey," prepared for OTA by R. W. Moxon under contractNo. 033-1400, p. 5,

App. BOffshore Manufacturing 511

Immediate Employment Impacts in theIndustry Making Foreign Investments

Offshore investment by U.S. firms creates jobs inforeign countries. To what extent do such jobs re-place employment opportunities in the UnitedStates? When a U.S. TV manufacturer moves itsassembly operations, offshore, some Americans losetheir jobs. But if the firm stays in business as aresult of the cost savings from offshore assemblyand if it might have failed without this movethenthe net effect can be to preserve some U.S. jobs. Ingeneral thee, if foreign investment improves thecompetitive position oUthe American firm, the ef-fects on domestic employment can be positive; theinvestment may create foreign jobs while savingdomestic jobs.

Demonstrating unambiguously that this has orhas not happened is, unfortunately, seldom possi-ble. The matter turns on a counterfactual question:What the outcome have been if the foreigninvestment had not been made? Largely because ofthis, past studies of the employment impacts of thesame investment have resulted in estimates rang-ing from losses in employment opportunities ofmore than a million to gains of half a million.,

Foreign investment may also affect the mix ofjobs available domestically. Even if net employmentincreases, certain job categories may suffer. Mostof the foreign workers in offshore manufacturingplants perform unskilled production tasks. Theseare the kinds of jobs that tend to be lost in theUnited States. Thus, the domestic skill mix general-ly shifts in the direction of the more highly skilledand professional jobstechnicians, engineers, man-agers. Unfortunately, unemployment in the UnitedStates is concentrated in the ranks of unskilled andsemiskilled workers. Moreover, since the electron-ics industry is geographically rather concentrated,offshore investment can have significant local andregional impacts.

Immediate Effects on Domestic Value Added

Closely related to employment is the impact onU.S. national income, or value added. Value-addedeffects can, in turn, be divided into several catego-ries: wages and salaries, profits, tax payments (tableB-1). Offshore investments generally substitute for-eign for domestic value added. The magnitude ofthese effects depends, however, on changes in the

q. Segall, "Introduction to the Conference," The Impact of Interneetimid Trade and investment on Employment: A Conference on Depart-ment of Labor Research Results (Washington, D.C.: Department of Labor,1978), p. 5.

competitive position of the firm making the invest-ment. In some cases value added may increase bothin the United States and abroad.

Foreign investments can also affect the distribu-tion of income among the categories of wages andsalaries, profits, and taxes. If offshore manufactur-ing substitutes foreign jobs for U.S. employment,value added will tend to move from wages and sal-aries toward profits and tax payments. But a sharpenough swing toward highly paid skilled and pro-fessional workers in the United States could reversethis effect. Offshore manufacturing may also createopportunities for firms to reduce their U.S. tax bills.On the other hand, if the company's competitiveposition improves sufficiently as a result of offshoremanufacturing, net tax revenues could go up.2

Effects on the Balance of Payments

Offshore investments are reflected fn the U.S. bal-ance of payments through both the current and cap-ital accounts. Foreign manufacturing generates im-ports, which show up on the current account, butthese will be partially offset by exports of materialsor components to the offshore plant. In the semiconductor industry, wafer fabrication has generally,remained in the United States, with wire bondingand other labor-intensive assembly operations mov-ing overseas. The wafers shipped to offshore plantsby American firms later return as finished inte-grated circuits (ICs); the latter are counted as im-ports, the former as exports. A substantial fractionof U.S. trade in semiconductor devicesroughlythree-quarters in the case of imports (ch. 4, table28)represents intrafirm transfers of this type.

The U.S. capital account shows outflows whenAmerican firms invest abroad, but moneys maygradually return in the form of profits or other pay-ments flowing back to the United States. Onceagain, the primary question is: What would havehappened in the absence of the investment? Has it'enhanced the competitive position of an Americanfirm? Or has U.S. competitiveness declined? Thesequestions are central to any evaluation of costs andbenefits.

Some of these questions are seemingly imponder-abl,sor at least subject to widely differing answers

rWhlIe overseas investment by American firms often displaces U.S. in-vestment, resulting in losses of domestic output and decreases in U.S.tax payments, foreign earnings remitted to the United States can offsetthese losses. The net result may be only a small net decrease due to the(:,reign investment. Perhaps more important, the distribution of nationalincome tends to be shifted toward capital, and away from labor. See P.B. Musgrave, Direct Investment Abroad and the Multinationals: Effects,on the United States Economy, Subcominittea on Multinational Corpora-tions, Committee on Foreign Relations, U.S. Senate, August 1975.

512 International Competitiieness in Electronics

thus estimates of the net impacts of foreign in-estritent on the balance of payments cover a range

just as broad as for employment effects 3 Again, thecrucial points involve the extent to which invest-ment overseas displaces investment ai home, andthe extent to which offshore production may dis-place or. alternatively, stimulate U.S. exports. Suchmatters can seldom be addressed on other than acase-by-case basis.

Indirect Impacts on Supplier IndustriesOffshore manufacturing in consumer electronics

or semiconductors generally cuts into the sales ofU.S. firms that supply these industries. Overseasplants normally buy expendable supplies and ma-terials locally; they may also purchase parts, com-ponents. and su!:3ssemblies from foreign ratherthan American firms. U.S. firms supplying suchcomponents as switches, circuit boards, resistors,and capacitors to the TV industry suffered heavylosses in sales as American consumer electronicsmanufacturers moved overseas.* U.S. suppliershave seldom been able to meet price competitionin overseas markets; when they lose sales to foreigncompanies, domestic employment and value addedsuffer. As their customers have moved offshore,some U.S. component manufacturers have, not sur-prisingly, followed.

Technology Transfer and OtherLonger Term Impacts

Most of the effects outlined above have long-term,as well as more immediate, aspects. Beyond directemployment or financial consequences, what possi-ble shifts in the competitive position of U.S. in-dustry could result from transfers of technologythrough offshore plants? If U.S. investments accel-erate processes of technology acquisition by othercountries, the competitive advantages of Americanfirms in electronics and related industries coulderode. Such a result is more likely in rapidly indus-trializing countries like South Korea and Taiwan,which have already emerged as significant competi-tors in consumer electronics, helped to consider-able extent by transferred technology.

When multinational corporations invest in devel-oping countries, they must generally train workers,typically drawing on the local population not only

Dufbauer and F. M. Adler, Overseas Manufacturing Investmentand the Balance of Payments (Washington, D.C.: Department of the Treas-ury. 19681.

1.. Marion, "TV Parts Makers Face Offshore Threat," Electronics, May24. 1979. p. 102.

for blue- and grey-collar employees, but for foremenand. often, middle managers. In electronics, the ex-perience that these people get has proved to be asubstantial benefit to indigenous -firms; not onlydoes a pool of workers, both skilled and unskilled,become available for locally owned companies tohire, but the managers of these companies are oftenpeople who got their start in a foreign-owned plant.

While it is easy to point to examples of this sort,where foreign investment has accelerated industrialdevelopment, technology diffusion is in any caseinevitable. Offshore investments may speed theprocess, but consumer electronics technology wa..,accessible to firms in Taiwan regardless of U.S. in-vestments there. Technology moves international-ly by multiple paths, some of which are quite inde-pendent of investment patterns. Furthermore,American electronics firms are not the only onesto invest in developing countries. Japanese com-panies have been quite active in moving electronicsoperationsparticularly those that are lower tech-nology and/or more labor intensiveto other Asiannations. In consumer electronics, developing coun-tries can probably learn more from companies likeMatsushita or Toshiba than from American manu-facturers. Technology transferred abroad via U.S.investments often helps to build foreign competi-tiveness, but the recipients could generally get thesame technology from other sources.

Evaluating Impacts

As pointed out at several places above, the under-lying difficulty in trying to evaluate the conse-quences of offshore investment comes in the com-parison of what did happL JA hat wi, havehappened if the-investment i..id not been made. Theanswer to such a question depends on judgmentsabout how markets would have been served with-out the investment, which in turn calls for analysisof comparative costs and other factors in the com-petitive environment. Reaching conclusions onwhat has taken place can be difficult enoughwit-ness the length of this report. But it is easier thandetermining what would have happened if a giveninvestment had not been made. Still, logic and theavailable information can yield some insights.

Critics of offshore manufacturing by U.S. firmsoften assume, perhaps implicitly, that the productsmade abroad could have been produced here in-stead, contributing not only to domestic sales butto U.S. exports. If true, U.S. employment, nationalincome, and balance of payments would all havebenefited from continued domestic production.Critics also tend to assume that American compa-

nies choose foreign investment over domestic man-ufacturing in order to increase their own profits,and that the company's competitive position wouldnot be seriously threatened if it chose not to investabroad.

Most defenders of offshore investment acknowl-edge that jobs are transferred .abroad in the shortrun, buf argue that the situation would in the longerrun be even worse without these investments. Theyemphasize that most such investments are defen-sive reactions to competitive threats, domestic orforeign. When the primary competitors are foreign,and American firms do not respond by moving off-shore, supporters of offshore manufacturing arguethat the United States would end up importing thesame goods from foreign-owned rather than U.S.-owned plants. Offshore investment thus preservesat least some benefits for the United States,'becauseexports will go to the offshore facilities, profes-sional and skilled jobs remain here, and the balanceof payments will look better than it otherwise wouldhave.

The counterresponse of the critics is generally asfollows. If the primary intent of the offshore invest-ment is to help U.S. firms meet import competition,then the proper response is simply to restrict im-ports. Interest groups that accept this argumentmay then combine; as they did in the Burke-Hartkebill, a call for protection against imports with a callfor restrictions on offshore investment. As in somany questions involving shifting comparative ad-vantage and the consequences for industrialpolicies, when the economics of the situation arecloudyas they are herepolitical considerationstend to become dominant.

Motivations for Offshore Investments

American electronics firms establish offshoremanufactui ing facilities to take advantage of low-cost foreign labor, Investing companies see costreductions as critical for meeting competitivethreats from foreign enterprises, or to expand out-put and sales in competition with other domestic;firms, or both.

Cost Savings for ProductsManufactured Offshore

American TV firms make monochrome sets off-shore, as well as subassemblies and complete chas-sis for color receivers. Production is labor-inten-sive, with low skill requirements, involving suchtasks as inserting components in printed circuitboards, assembling tuners, winding coils, and mak-

App. BCffshore Manufacturing 513

ing subassemblies for picture tubes. Offshore semi-conductor manufacturing has generally been lim-ited to assembly, primarily wire-bonding and en-capsulation. In recent years, some testing has beenperformed overseas as well, usually as an aid toquality. control. Many U.S. semiconductor firmsalso subcontract to local companies in developingcountries.

As table 18 (ch. 4) indicated, wages are muchlower in developing countries than in the UnitedStates or even Japan. Although labor productivityin such countries may also be low compared to do-mestic plants, large savings also can still result. In1980, the average hourly compensation for Ameri-can workers in the electrical and electronic equip-ment industry was $9.59; in the more popular loca-tions for offshore American subsidiaries, it rangedfrom $1.13 in Singapore to $2.40 in Mexico.5 Al-though wages have been increasing more rapidlyin offshore locations than here, offshore productionhas continued to be attractive in making both TVsand semiconductor devices. To some extent, firmshave responded to wage increases by moving onto other countries. For instance, two Americancompanies have announced plans to invest in SriLanka, where wage levels remain very low.6

Because costs for wafer fabrication and testingmake up a much larger percentage of the total forcomplex devices, offshore manufacture yieldsgreater savings for discrete semiconductors andsimple ICs. Table B-2 illustrates this, based onroui cost structures for simple and complex de-vices, and applying two arbitrary ratios of U.S. tooffshore wage rates. Substantial savings are possi-ble at either a 10-to-1 or a 5-to-1 wage ratio, but themargins are much larger for the simple device.

In TV manufacture, the net savings are smalleras a percentage of total production costs. Never-theless, for some kinds of subassemblies they canbe substantial, and in a highly price-competitivemarketas TVs have beenany saving can be im-portant. Zenith estimated in its annual report for1977 that the transfer to Mexico and Taiwan of cir-cuit module and chassis assembly for color setswould lower its unit costs by $10 to $15.

Strategic ImplicationsIn consumer electronics, offshore manufacturing

was a reaction to severe import competition, pri-

'Information from Bureau of Labor Statistics, Office of Productivityand Technology. Hourly compensation in Japan averaged $5.15.

°L. Antelman, "Harris to Construct $19 (sic) IC Facility in Sri Lanka,"Electronic News.'Feb. 8. 1982, p. 39. Motorola is the second U.S. firmplanning a factory there,


Table B-2.Cost Comparison for Offshore Assembly of Semiconductorsa

Discrete devices or simple integrated circuits Large-soale integrated circuits

Offshore assembly

Domestic assembly Domestic asst: 'blyWage ratiob Wage ratiob

10:1 5:1 Offshore assembly 10:1 5:1

Cost of chip $0.015 $0.015 50.015 $1.00 $1.00 S1.00

Assembly cost 0.050 0.500 0.250 0.15 1.50 0.75

Packaging cost 0.050 0.050 0.050 0.50 0.50 0.50

Testing cost 0.020 0.020 0.020 0.75 0.75 0.75

Reject cost L.015 0.015 0.015 1.00 1.00 1.00

Total $0 .150 $0.600 50.350 $3.40 $4.75 54.00

aThe basic costs Used in this table are from A Pefrort on the U.S. Semiconductor Industry (Washington, D C.: Department of Commece, Sep:ember 1979), p.73. Thesecosts ao not apply to specific devices. nor are they necessarily current. The purpose is simply to illustrate the magnitude of the cost savingsavailable through offshore

assembly°Assumed ratio of U S wages to wages in offshore plant.SOURCE. "Erects of Offshore and Onshore Foreign Direct investment in Electronics: A Survey,- prepared for OTA by R. W. MOxon under contract No.033-1400, p. 29.

manly from Japan. Sales had been lost, and profitscut to low levels or to losses; a number of smallerAmerican TV manufacturers succumbed du:ingthe period that RCA, Zen'th, and GE were movingoffshore. The story ii.: roelectronics is quite dif-ferent. Importso.xcl:.inixio of those from subsidi-aries of American firmswere not a major factorwhile U.S. firms were transfe..ring production over-seas; in the 1960's, imports from foreign-ownedcompanies accounted for only 1 or 2 percent of U.S.sales.

For semiconductors, the primary motives behindoffshore assembly were:

1. Cost Reduction as a Stimulus to Sales. Pricedeclines have lcd to a continuous stream ofnew applications of semiconductr.irsin otherwords, demand is highly price-elastic. As salesmount, costs drop through learning curve ef-fects. Offshore assembly accelerated price de-clines still more, opening further markets.

2. Capital Investment Constraints. Semiconduc-tor firms have had to continually increasecapital spending to keep up with exploding de-mand and advancing technology, but have notalways generated the profits needed to fundcapital investment internally (ch. 7). Given theneed for investment in costly wafer fabricationand testing equipment, offshore assembly of-fered an attractive way to expand capacitywhile conserving capital.

3. Risks of Large Capital Investments. Especiallyduring the 1960's, when many offshore plantswere established, semiconductor firms werewary of capital investments in automated pro-duction equipment. The fear was that techno-

logical change might quickly make them obso-lete. For example, semiconductor packaginghas changed a good deal, first as discrete de-vices gave way to ICs, later as ICs grew morecomplex. Several companies suffered as a re-sult of automating at the wrong time. Offshoreassembly offered flexibility without the risk oftechnological obsolescence. When technologyand/or demand stabilizes for a given product,automation becomes more attractive, and as-sembly is occasionally brought back to theUnited States.

Once some American firms succeeded in cuttingcosts by moving offshore, others were forced to fol-low; later, Japanese semiconductor manufacturersdid the same.

Alternatives to Offshore ManufactureAmerican firms invest overseas becaus--- to them

this seems the best course of action given their com-petitive situation. If this possibility were foreclosede.g., by Government policywhat other avenuesare open? The following appear to be the primarychoices:

1. Maintain production in .the United States,using labor-intensive processes similar to thosethat have been followed in offshore plants.

2. Maintain production in the United States, in-vesting in automated equipment.

3. Subcontract production to an independent for-eign manufacturer.

4. Discontinue production and sales of the prod-uct or products in question.

These four possibilities are briefly examinrd below.

Maintain U.S. Production on aLabor-intensive Basis

For some consumer electronics products. wherethe savings from offshore sourcing have been rela-tively small. this 1.,:ould probably be the alternativechosen. Nevertheless. the loss of the cost savingsfrom offshore assembly would hurt the competitiveposition of U.S. firms, some of which would prob-ably move to lower (Jim areas within the UnitedStates.

This alternative has little to offer for semiconduc-tor companies faced with increasing competitionfrom foreign manufacturers. Substantial cost penal-ties would hurt sales. especially for mature prod-ucts.

Automate Domestic Production

For many products that are now assembled oshore, automation is technicallsiblii.-A ericanTV inanufacturers_aready-t[se automatic compo-nent insertion to a considerable extent (ch. 6); in-vestments in this and other automated manufactur-ing methods could ! e accelerated. Automation hasbeen spreading rapidly in the semiconductor indus-try:Although automation is not at present feasiblefor all types of semiconductor productssome-times for technical reasons, other times becauseproduction runs are shortfinding the capital re-quired is a more central issue for many firms in theindustry. Smaller firms especially would have trou-ble financing extensive automation. As chapter 7pointed out. funds are scarce and capital-intensityincreasing in semiconductor manufacturing; man-agers' priorities place automation fairly low as longas there are feasible alternatives. Investments inautomation would divert funds from advancedwafer fabrication equipment, as well as from re-search and new product development--withoutwhich, in this fast-moving industry, automated pro-duction equipment would be useless.

Subcontract Manufacturing toForeign Enterprises

Subcontracting labor-inten, Ive production opera-tions to foreign firms has short run consequencesfor the United States not unlike those of direct for-eign investment, and the U.S. semiconductor indus-try has in fact made considerable use of foreign sub-contracting. Some American consumer electronicsfirms do the same. Subcontracting contributes toflexibility in responding to competitive pressures.Disadvantages come with respect coordination


in matters such as production schedules and costor quality objectives. And while subcontractingsaves capital compared to direct investment, directproduction costs will be higher because of the prof-its sought by subcontractors.

Especially in the semiconductor industry. butalso in consumer electronics, this option might wellbe the first choice of American companies unableto establish their own foreign subsidiaries. The at-tractions are especially great for low-volume prod-ucts where a foreign subcontractor with severalcustomers might be able to achieve scale econo-mies.

Discontinue the Product

Unless a firm had already decided on such a step,this would not be the first choicebut it might not

be the last. Whether American companies wouldstop making_ssirriducts if prevented from mov-ing offshore depends on the extent to which theirother options are practicable and cost effective.

Offshore Manufacturing Comparedto the Alternatives

Of the four options, U.S. consumer electronicsfirms would probably adopt a mix of the first three,depending on their product lines and competitivecircumstances. In particular, the smaller consumerelectronics manufacturers are much more limitedin investment possibiliti ,si.e., in automationthan companies like GE or RCA. In the semicon-ductor industry. tt f-t cost savings from offshore pro-duction are so la: getable B-2that most Ameri-can merchant firms would no doubt subcontract toforeign enterprises if they could not invest overseasthemselves. Some production would be transferredback to the United. States, probably high-volumeproducts made by larger companies.

What would be the impacts on the U.S. economyof the four alternatives compared to offshore invest-ment? To address this question, the effects ondomestic employment, balance of_pay_ments, andthe other categories listed in table B-1 could be com-pared. At least in principle, scenarios could be con-structed for the alternatives, singly or in combina-tion, most likely to be chosen by a given companyor industry. Ideally, estimates would cover a periodof years, because an offshore investment might, forinstance, initially cause an outflow of capiial whichin later years could shift to an inflow. In any suchprocedure, assumptions would have to be madeconcerning the future competitive environment forAmerican firms.

51.6 Internaric)-4,41 C,:tyl,-titivoness in Etectrontcs

A Case Example

Rather than pursuing an abstract analysis likethat outlined above. the methodology can be ap-plied to a simple example, a real company which.for purposes of the case study, has been renamedSystek.7

Systek, in Nog. decided to build it plant inTaiwan for assembling both complete automobileradios and subassemblies. All prouuction was to besold to Systek's U.S. operations, where final assem-bly and testing would take place. The company'smanagement chose to make this investment be-cause of a deteriorating competitive position; bymoving (ashore, the company felt that it could cutits production costs and prepare for upcoming bat-tles with Japanese producers. Automobile radiossold largely on the basis of price, and Systek's ma-jor customers, U.S. auto manufacturers, continuallysolicited and compared price quotations from var-ious suppliers. By the late 1960's, the automakershad begun to receive bids from Japanese electronicscompanies; Systek's management felt that the corn-pany would soon begin losing sales to the Japaneseunless it could significantly lower its own costs andprices.

Systek evaluatod several alternatives before build-ing its Taiwanese plant. The company had alreadyautomated its U.S. factories as much as it judgedpractical; the only option it saw for cutting costs-

"I tie nor fir,1 a Pi69 and in revised formi, .0, liy the liiircard Business Sarno!.

while remaining in the United States was to moveproduction from its urban site in the north to oneof the Southern States, where costs would be lower.Management judged this to be no more than a tem-porzry solution. Systek also considered subcon-tracting the assembly of its line of auto radios toa Japanese firm, but could see little advantage inthis choice because Systek had the resources andexpertise to estaVish its own foreign subsidiary,which would have lower costs than a subcontrac-tor could offer.

After a detailed feasibility study, the offshore al-ternative was chosen; Systek-Taiwan began produc-tion in late 1969. Operations went smcothly for thefirst few months, but then sales began to suffer be-cause of a decline in the U.S. economy. Productionhad to be cut back in Taiwan. As sales continuedto fall, the manager of Systek's U.S. plant placedfewer orders with Systek-Taiwan; fi;.131131 theseorders stopped entirely, and most of the workersin Taiwan had to be laid off. At this print Systek-Taiwan's management was authorized to seek otherbusiness, and by mid-1971 had begun doing elec-tronic assembly work for a number of Canadianand European companies.

Tables B-3 and B-4 examine the balance of pay-ments and employment effects of Systek's invest-ment in Taiwan. The tables are based on the com-pany's pro-forma projections for the first 5 yearsof operations to illustrate the expectations ofSystek's management at the time the decision wasmade. The actual results in terms of both employ-ment levels and flows of funds turned out to be

Table B-3.-U.S. Balance of Payments Flows With and Without Systek Investment

Capital LoanFiscal yea, outflow repayment

Capital flow (thousand of dollars)e

U.s. exportsof capitalequipment

U.S. exportsof components U.S. imports

Royalties Dividendsand fees and interest

Other payments tothe United States Net flow

With investment (Systek projection)1969 (4 months) $5,900 .$1,440 . $1,140 . $528 $2,930 4 $41 + $319 -$5,360

1970 - . 850 - 1,580 - 14,000 .238 +147 + 1,140 -10,000

1971 ..... ... - . 1,010 - . 1,310 -17,100 .237 .147 . 1,360 -13,000

1972 - . 700 - * 773 -19,900 +238 .174 4 1,580 - 16,400

1973 - - * 858 -22,000 .242 + 220 1,760 - 18,900

1974 - - - 946 - 24,200 .272 241 . 1,930 - 20,800

Total $5,900 . $4,000 $1,140 4 $6,000 - $100,000 +$1,230 $970 . 4$8,090 $84,500

Without investment (estimated)1969 (4 months) - - - -1970 - - - -$5,700 $570

1971...... - - - - 13,400 .1,340 - 12,060

1972 - - - - -23,000 +2,300 -20,700

1973 -33,700 +3,370 - 30,300

1974 - - - - -45,900 +4,590 _41,300

Total - - $121,700 +$12,170 - $109,000

aPlus indicates inflow to the United States, minus indicates outflow.

SOURCE 'Effects of Offshore and Onshore Foreign Direct Investment in Electronics1. 44, based on company records and author's estimates.

A Survey," prepared for OTA by R. W. Moon under contract No. 0331400. p.


Table B-4.U.S. Employment Levels Withand Without Systek Investment

Fiscal year

Systek employment in theUnited States (number of workers)Production

workersOtner

employees TotalWith investment (Systek projection)1969 (4 months) 1.480 452 1.9321970 1,283 393 1,6761971 1.204 368 1.5721972 1,124 341 1,4651973 1.120 341 1,4611974 1,115 339 1,454Without investment (estimated)1969 (4 months) 2,021 641 2,6621970 1.767 554 2,3211971 1,439 472 1,9111972 1,025 340 1,3651973 542 181 7231974SOURCE -Effects of Offshore Onshore Foreign Direct Investment in Elec-

tronics A Survey." t.:.c:...ared for OTA by R. W. Moxon under contractNo 0331400. p 44 Based on company records and authors es:imates.

heavily influenced by the business downturn in theUnited States. in tables B-3 and B-4, Systek's pro-jecti--. ns are compared with estimates by OTA'scontractor of the probable consequences if the in-vestment in Taiwan had not been made. Table B-3gives the estimated flows of funds, table B-4 theemployment comparison. The assumptions form-ing the basis for the estimates are discussed in de-tail below. The net effect of the investment in Tai-wan, obtained by subtracting the "without invest-ment" case from the "with investment" case, ap-pears in table B-5.

Based on the assumptions made, the initial. im-pacts of Systek's investment are negativeboth cap-ital and jobs are transferred to Taiwanbecomingpositive as time passes. This is typical of foreigninvestments for purposes of offshore assembly; theshort-term impacts tend to be negative, but over thelonger term the trend reverses, provided the invest-ment is assumed necessary for maintaining compet-itiveness.

In table B-3 the major flow of funds category isthat associated with imports; other financial flowsare much smaller. Imports have been assumed toincrease much more rapidly in the absence ofSystek's investment in Taiwan; in fact, as can beseen in table B-4, by 1974 it has been assumed thatSystek would no longer be making automobileradios in the United States under the "no invest-mon" scenario.. and its _domestic employmentwould fall to zero. How realistic is,this scenario?

Table B-5.Net Effect of the Systek Investment inTaiwan on U.S. Balance of Payments and Employment

YearBalance of payments flows'

(thousands of dollars)

Err ployment'number ofemployees)

1969 (4 months) .... - $5,360 7301970 -4,870 -6451971 -940 -3391972 +4,300 +1001973 , 11.400 '7381974 +20,500 +1,454aplus indicates int:ow to United States, minus indicates outflowSOURCE: Derived from tables and 8-4.

Photo credit: RCA

Consumer electronics assembly

The assumptions, based on events elsewhere in theconsumer electronics industry, are as follows:

If Systek had not invested in Taiwan, it wouldhave moved the same manufacturing opera-tions to a lower cost region of the UnitedStates.If Systek had done so, foreign mtinufacturerswould have had a cost advantage.Because of Systek's market knowledge. reputa-tion, and established working relationshipswith U.S. automakers, it would have been able


at first to hold on to part of the market evenwith a cost disadvantage.Although foreign-owned firms suffer initial dis-advantages in terms of proven ability to deliv-ery high-quality radios on schedule. they wouldmanage, over lime, to penetrate Systek's mar-ketOnce the foreign producers gained a substan-tial foothold in the U.S. market, their takeoverwould be swift.

Past rates of penetration in products like mono-chrome TVs or radios for home use indicate thatit might have taken Japanese and other foreign-owned companies about 5 years to penetrate Sys-tek's mz...rket more-or-less completely. This is theassumption behind the estimates in tables B-3 andB-4. Of course. the actual rate of penetration byforeign auto radio manufacturers might have beensomewhat faster or slower, but in the end thiswould not make mu':h difference.

As a result of these assumptionsthat the Taiwanplant helped Systek retain markets that it otherwisewould h4ve lost completelythe long run impacton U.S. employment and balance of payments turnspositive. The lystek investment still results in jobsbeing transferred overseas; it even accelerates theprocess somewhat. But job losses, and increased

imports, would most likely have occurred in anyevent, and could have been much greater.

The Systek case is also an example in which U.S.management acted to preserve American jobs bykeeping some production in the United States.Faced with falling sales as a result of recession, andthe need to cut output and lay off workers, Systekchose to stop production in Taiwanwhere the av-erage wage was less than one-tenth that in theUnited Statesrather than reduce its domestic op-erations still further. The plant in Taiwan was shutdown, with only the supervisors retained on thepayroll, until the company's management foundoutlets in Canada and Europe for products thatcould be made in Taiwan.

Typical Impacts of OffshoreManufacturing

The Systek case, by itself cannot be generalized,but it is suggestive; together with the earlier discus-sion of offshore sourcing compared to four alterna-tives, it points to some tentative conclus;ons. TableB-6 summarizes in a qualitative way the alternativesto offshore assembly outlined earlier. The table in-dicates the probable affects if alternatives other

Table B6.LiFely Effects of Alternatives to Offshore Manufacturing

Labor;ntensiveproduction in the

UNted StatesAvtomate domestic

productionSubcontract to

foreign enterprisesDiscontinuethe product

U.S. employment In electronics and related !ndustriesTotal domestic employment Positive in sarly years.

probably negativelater

Small positive in early years,probably negative later

No major change Negative

Proportion of skilled jobs Small posr,ibledecreasr,

Geographic distribution of jobs Move to low-wageareas

Small increase likely

No major change

No major change

No major change

Not relevant

Not relevant

U.S. value added in electronics and rotated industriesWages and salaries Positive in early years.

probably negativelater

Positive in early years,possibly negative liter


Profits Negative Possibly negative Probably negative Negative

Tax payments Negative Possibly negative Probably negative Negative

U.S. balance of payments for electronics and related IndustriesTrr...de balance (exports-imports) Positive in early years, Positive in early years,

probably negative in. possibly negative laterlater years


Other current account items Negative(principally investmentincome)

Negative Negative Negative

Capital account flows Positive in early years Positive in early years Positive in early years Positive In early years

Long-term effect on competitiveness of U.S. Industry due to technology transferSlightly positive Slightly positive Slightly negative Domestic industry

eliminated

Changes :n structure of domestic electronics IndustryWeaker firmsthreatened

Smaller firms weakened No major effect Domestic industryeliminated

SOURCE Office of Technology Assessment, based on 'Effects of Offshore and Onshore Foreign Direct Investment In Electronics: A Survey,- prepared for OTA byR. W. Moxon under contract No 033-1400, p. 47.

App. 8Offshore Manufacturing 519

than offshore assembly were chosen by a companyunder -pressure to reduce manufacturing costs. The

impacts follow the classification presented in table33-1.

As table 13-6 indicates, the labor-intensive domes-tic manufacturing alternative would keep produc-tion and employment in the United States, andtherefore have initially positive effects, but thesewould become negative in later years, the result ofa gradual decline in the ability to compete with low-wage foreign countries. This was the situation Sys-tek anticipated. Employment would drop, profitsdeteriorate, and tax payments fall. Positive effectson the trade balance (the result of lower importsin early years because of the absence of offshore,production) would soon be offset by shipments

. from foreign competitors.On the other hand, the domestic. manufacturing

alternative(s) would probably slow the migration ofU.S. technology overseas. Developing countries getboth tangible and intangible benefits from offshoreplants, including learning and experience thatstrengthens local industries. Over the longer_term,the result could be a relati"e weakeniiii-of the posi-tion of U.S. firms. flow serious is this possibility?

For offshore plants that ship most of their pro-duction back to the United States, labor-intensiveoperations, moEtly assembly, are performed over-seas. Although the general skills learned by produc-tion workers and supervisors are relevant, assemblytechnology itself is of little significance competitive-ly. The situation is rather different for point-of-salesemiconductor plants, but most of these are inEurope, where local firms already possess much ofthe technology associated with wafer fabricationand related processing Steps.

Automating domestic manufacture might havesomewhat similar results, but evaluation of this al-ternative is more problematic because the technol-ogy of automated production has been advancingrapidly. Electronics firms have guessed wrong atvarious times in their own evaluations, and the sec-ond column in table B-6 should be viewed tenta-tively. Employment would probably decline, but thecompetitive positions of U.S. firms that chose toautomate might or might not improve, dependingon circumstances. Purchases of automated manu-facturing equipment would stimulate the U.S.capital goods industry to the extent that this equip-ment was purchased domestically.

This is a difficult alternative for smaller com-panies with limited capital for investment. In con-sumer electronics, RCA has been perhaps the most

active U.S. firm in automating; it is no accident thatthis company as one of the largest and most diver-sified in the industry.

Most of the effects of the foreign subcontractingoption would be similar to those of offshore manu-facturing, as table B-6 outlines. This choice wouldharm domestic suppliers, who would have difficul-ty selling to overseas subcontractors.

Discontinuing production, the last alternative,has negative consequences for the U.S. economy,although t'-.n capital released could be invested inother industries.

Summary and Conclusions

American manufacturers in many industries aremoving some of their production overseas. At thesame time, foreign firmsfor various reasonshave begun to invest more heavily in the UnitedStates. In general, the advantages and disadvan-tages of either type of investmentfrom the stand-point of impacts on U.S. employment, and the U.S.economy in generalcan only be evaluated on acase-by-case basis. In most instances, the net impacts of offshore manufacturing by U.S. electronicsfirms seem to be relatively small.. But even if thenet effects are small, the consequences for the indi-viduals and firms affected can be seriousfor work-ers who lose their jobs, for suppliers who lose sales,for communities and regions where industrial ac-tivity has diminished. To call these short run ad-justment problems does nothing to mitigate them.

Moreover, the shift of unskilled and semiskilled .

jobs overseas seems in the end detrimental to U.S.interests. This country already has a large numberof unemployed job-seekers, many of whom are real-istic candidates for unskilled or semiskilled manu-facturing jobs but not for work demanding highskill levels. That overseas investment may some-times help maintain the competitiveness of Ameri-can firms and industries seems small recompensefor those who lose jobs or,job opportunities. On-shore investments by Japanese and other foreignelectronics companies may provide something ofa counterweight, but thus far many more jobs havebeen lost than gained. On the other hand, policiesthat would restrict overseas investments by U.S.firms seem generally counterproductive. As dis-cussed at some length in chapter 8, the alternativeof choice would appear to be a strong commitmentto upgrading the U.S. labor force so that transfersof unskilled work overseas will be less damaging.

521

APPENDIX C

Case Stuthes in the Development andMarketing of Electronics Products*

Consumer Electronics: The 700-WattPower Amplifier

The Product

In early 1970, Robert Carver, an engineer turnedentrepreneur with a passion for music and high fi-delity sound reproduction founded a small com-panyPhase Linearin Seattle to manufacturehigh-power, state-of-the-art stereo amplifiers. Thefirm began as a limited partnership but was incor-porat:0 later that year. Carver became the majori-ty stockholder, while his partner and an SBIC (SmallBusiness Investment Corp. see ch. 7) were minori-ty shareholders. During the early years of the com-pany Carver made all the major decisions. PhaseLinear Corp.'s first product was a 700-watt poweramplifier for use as a component in home audio sys-tems. Carver tried to bring out one new producteach year, and by 1974 Phase i'inear had threeamplifiers on the market.I

Stereo amplifiers range in power output from afew watts per channel on up to 350 watts per chan-nelthe Phase Linear 700's capabilityor more.The main feature differentia_ing the Phase Linear700 from others on the market was its great power;one of the first advertisements touted it as "the mostpowerful, most advanced high-fidelity solid stateamplifier in the world." In a 'February 1972 articlein the magazine Audio, Carver described severalof the design problems overcome in achieving thispower level. The main obstac!e had been transistorvoltage breakdown. While 350 watts at 8 ohms foreach of two channels requires a power supply capa-bility of more than 200 volts, the best existing audiotransistors had sustaining voltages of only 120 volts.Carver solved the problem by working with a ma-jor semiconductor manufacturer to modify a 600 -volt television horizontal sweep transistor so thatit would he suitable for use in audio amplifiers.2

Crossover distortion created another barrier. Insmall, low-power amplifiers, this form of distortioncan be avoided by allowing an "idling current" to

.i.rileSe case studies are based on reports prepared for OTA by J. I.

Wheatl6y, D. M. McKee, S. R. Barnes. L. E. Hartmann. and D. f. Keithunder contract No. 033-1190../Interview with Robert Carver.

R. Carver, "A 700 Watt Amplifier Design." Audio. February 1972.

520

flow continuously from the output transistors. Atlower powers, the idling current does not generatemuch heat, but in a 700-watt amplifier with 24 out-put transistors this approach is impractical. A novelbiasing circuit, eliminating crossover distortionwhile operating without idling current, solved theproblem.

The IndustryMarket Growth.During its early years, the

high-fidelity industry catered to a small market con-sisting mostly of the wealthy. The cost of early high-fidelity equipment made it a sign of status., In the1930's, few could afford the $3,000 to $10,000 priceof a Capehart record changer. As the price of audiocomponents fell during the 1940's and 1950's, anew market for high-fidelity equipment grew, cen-tered on hobbyistsaudiophiles and music enthu-siasts willing to spend several thousand dollars toassemble systems built around separate tuners, am-plifiers, turntables, arid speakers. Sales levels re-mained modest, but continuing technological im-provements led eventually to the present mass mar-ket. Factors contributing to the expansion of high-fidelity equipment sales in the United States since1960 include:

rising levels of disposable income;the introduction of stereophonic sound record-ings in 1959;approval by the Federal Communications Com-mission of FM stereo-multiplex broadcastingin 1962;solid-state equipment designs beginning in themid-1960's, which sharply reduced manufac-turing costs as well as improving reliability;andprogressively lower tariffs on imports, leadingto more intense price competition (duties onspeakers and amplifiers were cut from 15 to7.5 percent, and on tuners and receivers from12.5 to 10.4 percent, between 1968 and 1972).3

Demographic trends helped catalyze demandduring the 1960's and 1970's. Fifteen to thirty-fiveyear olds buy most high-fidelity equipment; at thetime this was the-most rapidly growing segment of

'E. Ashkenazi. -The Executives' Corner,- Wall Street Transcript, June

11. 1973.

522

App. CCase Studies in the Development and Marketing of Electronics Products 6 521

the U.S. population. The flowering of the music-oriented youth culture in the 1960's also boostedsales. In the latter part of that decade, audio prod-ucts became the fastest growing portion of the con-sumer electronics industry, with demand especiallystrong in the 15 to 24 age bracket. Few Americanmanufacturers were able to capitalize on thisgrowth, as imports made major inroads into theU.S. market. Japanese equipment often surpassedthe products of U.S. companies in performance,while sel!...ng for less: Continued technological im-provements, creative new product developments,and lower foreign labor costs helped imports eataway at the market shares of American firms.

Amplifier Technology.The evolution of thepower amplifier is marked by a long series of in-cremental design improvements aimed at reducingdistortion in the reproduction of music. Amplifierscreate two kinds of distortion. Clipping is the mostserious; it occurs when the music being reproduceddemands a higher instantmeous power level thanthe system can deliver. Normally, these extraor-dinary power demands are fleeting: the high C inan aria, the climax of a thundering crescendo ina baroque score. On an oscilloscope display, clip-ping appears as a flattening of the peaks and valleysof the waveforms.

The second type, crossover distortion, occurs atlow instead of high volume levels. Crossover distor-tion gets its name from the small notches in wave-forms seen on an oscilloscope as the polaritycrosses from plus to minus or vice versa. Resultingfrom nonlinearities in transistor characteristics atlow current values, this form of distortion producesharmonics that are approximately constant in levelregardless of output power, hence only audible dur-ing quiet passages.

Consumer Behavior.Buyer psychology was oneof the keys to the market for the Phase Linear 700,as illustrated by the opening paragraph of an arti-cle in a 1976 issue of Saturday Review:4

When I first got into hi-fi nearly 20 years ago,everyone knew that you needed a minimum of 10watts of amplifier power for good high fidelity. Andso I swapped my table radio, with the servicemaninstalled phono input (one watt of power, if I wasvery lucky), for a fashionable 10 watt amplifier. Afterthat came a 25 watter, then my first stereo amplifier(35 watts in each of its two channels), then a 60 wattper channel amplifier. Today I have one with two200-watt channels. At each step of the way, I've beenperfectly in fashion. But what else, if anything, haveI gained from my power hunger?

1. Berger. "Power Plays,- Saturday Review, Jan. 6. 1976. p. 40.

The author goes on to point out that sound qualitydid in fact improve, but in small increments andat high cost.

The Phase Linear 700.When introduced, thePhase Linear amplifier was not only more power-ful than others on the market, but offered morepower for the money. In 1971, the Crown DC 300(150 watts per 'channel) listed at $685, compared tothe Phase Linear's $749.5 The Phase Linear stoodout as a bargain, offering more than twice as manywatts-per dollar, and helping establish a new mar-ket category. In the early 1970's, a "sPper-power"amplifier was considered to be anything deliveringmore than 50 watts per channel. The entrants inthis class included, in addition to the Crown, DC300: the Pioneer SA-1000 (60 watts per Channel,$230); the Harman-Kardon Citation 12 (60 watts perchannel, $298); the SAE Mark III (120 watts perchannel, $700); Sony's. TA-3200F (130 watts perchannel, $359); and the OM 911 (120 watts perchannel, $540.8 The Phase Linear-700 surpassed allthese by a large margin. So successful was PhaseLinear in opening up a new market niche that itfaced no direct competitionfrom either Americanor foreign firmsduring its first 3 years.

The Competition.To the extent that it providedmore power at a lower price, the Phase Linear 700was able to capture buyers from other companies.These competitors were mostly large or medium-large, and well-established. Crownmaker of theDC 300was a division of International Radio andElectronics. The privately held firm sold most ofits products to profesSional musicians and institu-tional purchasers such as churches. Marantz, pro-ducer of another powerful amplifier, was a whollyowned subsidiary of the Superscope Corp. Super-scope had been incorporated in California in 1954,and in 1966 purchased the Marantz Co., Inc., add-ing 50-percent interest in Marantz japan, Inc. in1971. Marantz products were manufactured in thecompany's Tokyo plant. Superscope had alsoserved al; exclusive U.S. distributor of Sony prod-ucts since 1957. A third maker of high-power am-plifiers was the McIntosh Corp., also privately held.A small company compared to Crown or Super-scope, McIntosh produced high-end stereo equip-ment almost exclusively.

From an international perspective, that the Japa-nese presence in the super-power category wassmall' may seem remarkable; Japanese producershad by 1970 captured an overwhelming share of theU.S. audio market. The explanation appears to be

°"1072 HiFi Preview Directory." Audio. September 1971.°Ibid.

52r4

522 - International Competitiveness'in Electronics

simple: the market for separate amplifiers of allsizes was not very large, and within it, the marketfor the very largest amplifiers, with their high pricetags, was even smallerless than 10 percent.' TheJapanese approach had been to concentrate theirproduct development and marketing efforts on thelargest selling st,:ireo components, such as in-tegrated tuner-receivers. Japanese firms like Akai,

Kellwood, Nikko, Pioneer, Panasonic, andL'ansui also offered separate amplifiers in the morepopUlar power ranges. Many of these companieswere large-and diversified. Panasonic, for example,is a brand name of the Matsushita conglomerate,whose export sales to the United States in 1971totaled $358 million. Pioneer had 1970 U.S. salesof $120 million, and was already one of the largesthigh-fidelity equipment producers in the world.

These firms relied on their ability to sell equip-ment perceived to be of good quality at low prices.Their strategy had been to concentrate almost ex-clusively on mass-market products, leaving the ex-pensive, high -end components to smaller Americanfirms. Phase Linear thus faced little competitionfrom the major Japanese electronics firms in itsearly yea rspri maiiiy,becau6e of the relative small-ness of the market for super = power amplifiers. Buteven though Phase Linear achieved its initial suc-cess by appealing to audiophilesand while doingSo acquired a reputation for high qualitythe com-pany soon began selling to a wider range of buyers,largely because of its modest prices.

Distribution.Most audio equipment manufac-turers sell through networks of franchised dealersserve:L.4 regional sales representatives. Dealersgenerally take delivery from the factory, althoughPioneer and some of the other large firnis maintainr-elton;91 warehouses. The greatest portion of retailsales tii:-&-rna,d__,F through audio. specialty storeshigh-volume, low - margin outlets emphasizing theheavily advertised, low-priced Japanese brands. Asecond type of retailer, the "audio salon," tends tobe individually owned, and to specialize in moreexpensive products. In addition to prestige linesof the mass-market firms, these stores sell high-endequipment made by smaller and less well-knowncompanies. Other major outlets include: mail orderhouses; discount and department stores; appliance,radio, and TV dealers; and catalog showrooms.Generally, these limit themselves to the more mod-erately priced and popular components.

.. 'frier:few with Don Prewett, Phase Linear Corp.

Product DevelopmentThe super-power amplifier for home stereo sys-

tems was largely the brainchild of Phase Linear'sfounder, Robert Carver. Very high power.as a routeto better sound quality at all listening levels was anovel idea when Carver began experimenting in hishome workshop. He built a series of amplifierswhose power capability surpassed anything on themarket. The tests he ran backed up his insights;music sounded better to lihn played through theprototypes.. Measurements of audio distortion sup-ported his subjective judgments.

Convinced of the virtues of a stereo amplifierwith a wattage rating more than double anythingthen available, Carver set out to create a designsuited for commercial production. Most of thiswork he did himself. While Carver enlisted the aidof several Motorola engineers to solve the transistorvoltage breakdown problem, the ideas were basical-ly his.°

Phase Linear placed a premium on technology inthose early days. Carver, highly regarded in theaudio industry as a gifted designer, wanted to buildan amplifier of unprecedented power, but he alsowanted to build one that was affordable. This sec-ond objective, more than anything else, called forthe creative use of technology in order to reduceproduction coststhe simplest possible design thatwould deliver very high power levels.

Carver did not have the financial resources to domuch marketing research, but his experience toldhim that a low-priced, high -power amplifier wouldsell. After showing prototypes built in his homeworkshop to dealers, and being assured that theywould carry the product, he decided to go into lim-ited production. Manufacturing began in an oldSafeway store leased .for the purpose.

MarketingAt. first, Phase Linear. took a rather ad hoc ap-

proach to distribution and marketing. As %-'ord ofthe Phase Linear 700 spread, the company accepteddirect orders from anyone individuals, as well asdealers large and small. In 1971, Phase Linear hireda marketing manager who set up a system of com-pany sales representatives, but the firm still founditself with a growing backlog of orders from an un-wieldy assortment of some 603 buyers.° Two yearslater, a new marketing manager took overDonPrewett, a recent MBA graduate. Prewett began set-ting up a new distribution system.

"-A 700 Watt Amplifer Design.- op. cit."Prewett. op. cit.

App. CCase Studies in the Development and Marketing of Electronics Products 523

It took 2 years to reorganize Phase Linear's sys-tem of dealers and sales representatives. Prewettfound representatives with overlapping territoriesand Phase Linear products stocked by competingstores. The firm's managers decided that PhaseLinear products should be sold primarily throughlarge chains and retailers of stereo equipment,which at that time were growing at a phenomenalrate, whiloavoiding small, specialized outlets. Twoof the highest volume stereo chains in the countrybecame major outlets for.Phase Linear. With the de-mand for stereo equipment exploding in the early1970's, these retailers were opening many newstores. By 1980, the company was selling its prod-ucts through 16 sales representatives to 275 dealersoperating about twice that number of retail outlets.Dealers and.designated repair shops handled serv-ice in the field.

Phase Linear's strategy was to dominate itschosen market niche and expand from there. The700 faced no real competition for more than 3years. As it became apparent that the first part ofthis strategy would be successful, the companyquickly began to extend its product line. In January1972, it came out with the Phase Linear 400, whichoffered 200 watts per channel. This amplifierproved even more popular than the 700. By 1974,two to-three times as many 400s were being soldas 700s, and the company's annual sales hadreached $3.5' million.

Thanks to imaginatively simple design, produc-tion costs were low; becauseof the lack of competi-tion, the Phase Linear 700 could be priced to yielda healthy profit. As a result, the firm was able togenerate virtually all the funds needed for expan;sion from internal sources.

The Industry Reaction

Phase Linear's entry into the high-fidelity in-dustry was inconspicuous. At the time of its incor-poration in early 1971, the company counted onlya few employees. Most of the industry regarded its

- product as an oddity with limited appeal.Japanese firms, which constituted the dominant

force in the audio industry worldwide, had over-looked the potential of extremely high-power am-plifiers, which did not seem to fit their export-oriented approach. Efficient production technologyand effective advertising, sales, and distributionenabled them to drive their American competitorsout of the market for mainstream products likestereo tuner-receivers. But the Japanese manufac-turers did not regard the stereo separates marketas big enough to deserve much attention. Firms

such as Kenwood, Sansui, and Pioneer maintainedseparate stereo component lines, but mostly for thesake of product mix and the lustre that high-endcomponents added to their image.

As a result, competitive response to PhaseLinear's products was slow in taking shape. Onlyin 1975, after the company had already secured thelargest market share of any entrant in the marketfor separate amplifiers-15 to 17 percentdid sev-eral firms, both American and foreign, introducesuper-power amplifiers of.their own. For the mostpart, Phase Linear's U.S. competition came fromsmall and relatively new companies. One of these,the Great American Sound Co., came out with amodel called the Ampzilla aimed at the heart ofPhase Linear's marketthe 20- to 35-year-old malehi-fi hobbyist. Bose Co., which had been primarilya manufacturer of speakers, also introduced a /super-power amplifier. One of the largest firms to Ienter at this time was Marantz, mentioned earlier./These companies constituted the first wave of com/Petition. A second wave came as the huge Japanesemanufacturers, including. Mitsubishi and Yamaha,finally began making super-power amplifiers.

The response of Pioneer Corp., the sales leaderin the industry, was the most belatedbut most sig-nificant by far for Phase Linear. In 1978; U.S.Pioneer, a wholly owned subsidiary of Pioneer-Japan, bought Phase Linear. Just. prior to. the ac-quisition, Carver had sold his stock interest backto the corporation. At this point, U.S. Pioneer pur-chase the company from the remaining stockhold-ersG. giver's ex-wife, the SBIC, and his formerpartnerior a price reported to be in the middleseven-figure range.

How did this change of ownership affect PhaseLinear? While leaving the company's managementteam intact, .Pioneer placed at Phase Linear's dis-posal a wide range of new resources. The companynow had ample financing for product 'developmentefforts, and new sources of technology. PhaseLinear's marketing capability was' strengthenedbecause ft could use the parent company's exten-sive U.S. retail network. Pioneer also became a sup-plier of component parts for Phase Linear's prod-ucts.

What was Pioneer's motive in purchasing PhaseLin, 'T.? The major reason was probably a desire tost hen its position in a. rapidly growing marketsegment. Partly because of their successful strategyof dominating the mass market, many of, the Japa-nese brands lacked the quality image necessary forsuccess at the upper end. The best evidence for thethesis that Pioneer acquired Phase Linear primari-ly for its prestige value is the succession of new

523


stereo components introduced under the PhaseLinear name. Within 2 years of the acquisition,"Phase Linear" turntables and tuners were beingexported from Japan to the United States. Theseproducts were developed by Pitneer design engi-neers and manufactured at Pioneer facilities. Ex-cept for their high price tags, they did not differgreatly from comparable Pioneer products:1° Ratherthan any siviificant transfers of technology to orfrom the United States, the effects of the takeoverseem to have been restricted to marketing and fi-nancial matters.

In recent years, Phase Linear has fared well in-ternationally, with 30 percent of its 1979 sales com-ing from exports, two-thirds of these to Europe. But

-Japan is one major market that Phase Linear, alongwith almost all other U.S.-based audio manufactur-ers, has not been able to crack. Only two AmericancompaniesMcIntosh and JBL, the latter a manu-facturer of speakershave established distributionchannels in Japan. Their entries took place shortlyafter the end of the war. Since then, no major Amer-ican manufacturers of consumer electronics havebeen able to sell their products within Japan in anyvolume. This inability stems at least in part fromdistribution problems. The task is not impossible,but costs for deciphering and meeting the manyproduct regulations, as well as establishing market-ing channels, are great. Even so, the distributionof electronic products and household appliances inJapan is less complex than for goc such as foodor kitchenware. A major reason is e e emergenceof a few large manufactur onsumer productsand household appliar . -e.g., Matsushita .which have taken the initiative in organizing sim-pler marketing channels. Still, over 80,000 retailers,more than three-quarters quite small, handle con-sumer electronic products.

Phase Linear executives had their eye on the Jap-anese market for some time prior to the 1978 ac-quisition by Pioneer, but report that Pioneer'spolicy has been to refrain from encouraging effortsby Phase Linear to export back to Japan. In par-ticular, Pioneer apparently has no intention or mak-ing its domestic marketing channels available to itsAmerican subsidiary. This might reflect: 1) simpleexclusion; 2) a decision that i.1 would be too costlyto undertake a mar' ' program in Japan; or3) a market-dividin, iratc ry whereby Pioneer de-cided to promote Phase Linear only in the UnitedStates and Europe.

"Intelviews with stereo dealers.

ConclusionWithin the high-fidelity industry, qualitative dif-

ferences between products of similar price tend tobe small. Industry executives generally believe thatthe successful firms are those that market most ef-fectively. Phase Linear was typical; its rapid riseto a position of leadership in one sector of the in-dustry was largely due to effective marketingde-signing and building a product that others had over-looked but that consumers were ready to purchase.Robert Carver began to pursue his ideas based onintuition about the market. At the time, the notionthat real demand could exist for a super-power am-plifier would probably not have gotten much of ahearing in a large, established company.

Technology played a crucial role in the secondstage, the actual development of the product, whereCarver's sense of design led to a simple, low-costamplifier. Phase Linear's critics sometimes re-marked that they were "designed to the bone,"meaning that they gave maximum power while of-fering little in the way of backup or protective cir-cuitry. But it was apparently just this quality ofbrute power that younger buyers of stereo equip-ment wanted. Nonetheless, -the company also rec-ognized that demand for a 700-watt amplifier wouldbe limited, and quickly moved to broaden their of-ferings.

Robert Carver later started another company; in1982, Carver Corp. began advertising a power am-plifier featuring "750 Watts/chan. DynamicHeadroom for just $..,-99."

Semiconductors: The 4K DynamicMOS RAM

The ProductElectronic data processing, at one time solely a

matter of computers, has spread to a wide rangeof products: industrial controllers, automatedmachine, tools, "smart" terminals, calculators, evenhousehold appliances. These systems need mem-orythe ability to store and retrieve information(see ch. 3). Random access memories (RAMs) canretrieve or rewrite digital data stored in an arbitrarylocation on command. Most integrated circuitmemory is of the random access type. Both majortransistor technologies are used in semiconductormemoriesbipolar and MOS (metal oxide semicon-ductor, ch. 3), with MOS now the largest seller byfar.

L 94., 0


MOS RAMs can either be static or dynamic. Stat-ic RAMs hold their contents indefinitely, providedthey are supplied with power. Dynamic RAMs relyon capacitance for storage; they must be "re-freshed" every few milliseconds. Static RAMs re-quire inore complex memory cells than dynamicRAMs, and are thus not as dense, taking up morearea on the chip and costing more.

In 1974, when the 4K dynamic RAMwhich canstore 4,096 bits of informationwas introduced, anew generation of memory circuits was appearingabout every 30 months. Each design generation hadbeen four times larger than the previous one, thesequence being 256 bits, 1K, 4K, then 16K, andin the early 1980's-64K. By 1983, 256K RAMSwere in pilot production. One explanation for thefourfold density increment is that, while technologi-cal capability in terms of circuit density roughlydoubled each year, design costs were high enoughso that, if new designs came out every 12 to 15months, they would not generate enough cumula-tive sales to be profitable. By the end of the 1970's,the intervals between RAM generations had length-ened to several years.

The newly introduced 4K dynamic RAMs werehailed in mid-1974 as far outdoing 1K types as thecheapest way to satisfy user needs. Despite spottyavailability during that year, they were quickly de-signed into microcomputers, minicomputers, andperipherals; manufacturers of mainframes waitedfor price decreases and assurances of productreliability before switching from 1K to 4K chips.

The Industry Setting

While some captive semiconductor manufactur-ersnotably IBMhave designed and built theirown RAMs for internal use, this case study treatsthe competition for sales in the merchant market.Development of 4K chips for merchant sales beganin the early 1970's, with samples available by late1973. As the 4K RAM moved into volume produc7tion, the semiconductor industry entered the mostsevere downturn in its history, the result of ageneral recession in the U.S. economy beginningin 1974. Semiconductor firms furloughed 50,000employees, and idled $750 million in productioncapacity."

As the 4K chip emerged and economic recoverybegan, 1K RAM sales declined. The 4K RAMs ac-counted for only $14 million in sales during 1974,but $45 milliem the next year. By 1976 1K sales had

""New Leaders in Semiconductors." Business Week. Mar. 1. 1976. p,40.

fallen to $42 million, while 4K sales soared to $161million.12

Intel Corp.'s 4K RAM design was first onto themarketvia a licenseebut the competition quicklybecame intense, complicated by production prob-lems at several firms, including the industry'slargest manufacturer, Texas Instruments (TI). Onlyat the end of 1975 had firms such as Intel, TI, andMostek ironed out most of their processing dif-ficulties; while earlier projections had been forshipments of 10 million chips during the year, ac-tual output was perhaps half this. The 4K RAMposed the greatest difficulties the industry hadfaced up to that time in moving a product into vol-ume production; indeed, before the 4K Rikivfreached high volumes, 16K RAMs had been an-nounced.

It took several years for an industry standard 4KRAM configuration to emerge. Three chip designswere vying for dominance, with the situation inconsiderable flux:"

Intel/TI's 22-pin package, announced by Intel andthen modified by TI, uses TTL voltage levels for alladdress, data-in, and data-out lines; it requires onlyone high-voltage clock level but needs three powersupplies.

Motorola/AMI's [American Microsystems, Inc.] 22-' pin package differs in having an extra reset pin,which must be energized when power is first ap-plied.

Mostek's 16 -pin package takes up less board spacethan ,the other two, at the cost of some added systemcomplexity in clocking and interface logic, since thedevice must be multiplexed; it is also TTL-compatibleat all inputs, including the clock input.By the end of 1976, sales of 16-pin designs were

increasing at the expense of 22-pin devices. The22-pin part was larger; the extra pins also led togreater assembly cost. A second focus of techno-logical competition was access timethe time, onaverage, to retrieve a bit of information from thememory. Access time for memory chips is normallymeasured in nanoseconds, 1 nanosecond (ns) be-ing 10-9 seconds. For RAMs, an access time of 100ns is considered fast; 500 ns is slow.

The Competitors

Capital requirements for manufacturing 4KRAMs were not, in the mid-1970's, a significant bar-rier to entry. Many of the competing firms had

"Dataquest. Oct. 7, 1977. pp. 18.6-9."L. Altman, "Semiconductor Random-Access Memories." Electronics,

lune 13. 1974, p. 109.

52'7


begun operations only a few years earlier, and were. still relatively small.

Microsystems International Ltd.The first com-pany to bring a 4,096-bit RAM to marketin late1972was Microsystems International of Ottawa,Canada, a licensee of Intel. Th 227pin, 3-transistormemory cell chip was based on proprietary processtechnology. with the company benefiting fromearlier experience as a licensee for Intel's 1KRAW. Although first with a working part, Micro-systems International never became a major factorin 4K RAM sales.

Intel.Intel's 4K chip followed an immenselysuccessful 1K productwith the possible exceptionof IBM's proprietary 1K design, the most widelyused semiconductor memory circuit up to thattime. Judging that the product lifetime for a 2KRAM in volume production wouldprobably be nomore than 6 or 8 months, Intel jumped to a 4K chip,introducing--in the summer of 1973a slow (600ns access time) 4K device designed for small sys-tems. The .company planned to introduce a high-speed version later in the year; both were to havea 22-pin, single-transistor memory cell design. Thehigher speed chip, with maximum access time of150 ns, would be better suited for large computersand was projected to take over most of Intel's 4Kproduction during the first half of 1975. Intel hopedto capture as much as half the potential market.

Meanwhile, customer desire for greater circuitboard density was prompting movement away fromthe 22-pin package. At the end of 1974, Intel an-nounced plans to introduce its own 16-pin device.The company thereafter continued to build both 22-and 16-pin RAMs.'

Mostek.In September 1973, Mostek was sam-pling an innovative 16-pin RAM, one in whichsome of the pins served two functions (called mul-tiplexing). The chip enjoyed a two-to-one densityadvantage over the competition. Eventually, aftera redesign reduced the size even further, it becamethe de facto industry standard. Howe-ter, Mostek,along with other, chipmakers, suffered throughyield and quality problems which cut into its abili-ty to capture early market share.

Despite the pioneering features of its 16-pin de-sign, the firmin common with the rest of the in-dustrydid not rely on patents to protect its tech-nology. Mostek's 1977 Common Stock Prospectusstated ". . . the Company believes that success inthe semiconductor industry is not dependent uponpatent protection but is dependent upon engineer-ing and production skills and marketing ability. Itdoes not anticipate that the grant of any patent ap-

plication will significantly improve its competitiveposition."

National Semiconductor.National developedboth one- and three-transistor cell designs of itsown, By mid-1975 it was marketing 22- and 18-pinchipsthe 22-pin part faster than, but compatiblewith, that of TI. National's strategy of seeking fasteraccess times is part of the explanation for its deci-sion not to build a 16-pin device; National's engi-neers felt, incorrectly as it turned out, that theMostek approach did not lend itself to speed im-provements that would prove great enough. The18-pin choice allowed good board density and highspeed without requiring the multiplexingi circuitryof 16-pin packages. Two other firms quickly linedup as alternate sources for National's 18-pin part.

Texas Instruments.TI was the first to drop its4K RAM price below the cost to purchase four 1Kchips. By September 1974, TI was producing more4Ks than anyone else, having solved its earlier yieldproblems. At the close of the year, TI added an18-pin package to its existing 22-pin 4K catalog;both the 18-pin and the new 22-pin part offered ac-cess times of 200 ns. TIs' second source for its 4KRAMs was Advanced Micro Devices (see below).

Other U.S. Entrants.Fairchild became Mos-tek's second so) rce, offering a pin-compatible ver-sion of Mostek's unit while also producing anotherdesign, with faster access times, based on the pro-prietary Fairchild Isoplanar processing technology.Meanwhile, American Microsystems, Inc. (AMI)and Motorola developed their 4K RAMs jointly,sharing masks and processing technology. AMIwas particularly confident of its product "even forthe chronically confident semiconductor indus-try"and expected its entry to become the industrystandard; its speed and power characteristics,single clock design, pin configuration, and second-source at Motorola all seemed to the company tojustify this belief." AMI's porter,. M)terola. wasrelying on this new 4K RAM , volesales to its semiconductor division, after a coupleof false starts tivith.early memory products."" Still,Motorola also sought other alternate sourcingarrangements.

Japanese Firms.Semiconductor manufacturersin Japan were developing their own 4K RAMs overthe same time period. Nippon Electric Co. designeda 4K RAM described as an improved and enlargedversion of the company's three-transistor cell, 1Kpart." Hitachi hoped to have a 300 ns chip on the

"H. Wolff. '4.096-Bit RAMS Are on the Doorstep." Electronics. Apr.12. 1973. p. 76.

'511. Wolff.Customers Sweat Out 4,096 -Hit RAMS." Electronics, Mar.21. 1974, p. 70.

'5"4.096Rit RAMs Are on the Doorstep." op. cit.


market by the end of 1973. Fujitsu's 4K RAMs usedthree-transistor cells, but one-transistor productionversions were anticipated. Toshiba was also devel-oping a 4K design.

These development efforts attracted little atten-tion in the United States. Shipments of 4K RAMsfrom .Japan did not begin to enter the U.S. marketuntil 1977, and then only in small quantities. IfJapanese competition appeared to be no more thana minor threat, European firms posed even less ofonein part because most had neglected MOS tech-nology, continuing to concentrate on bipolar. In976, U.S. firms had 90 percent of the world marketfor MOS devices of all types, with the Japaneseholding must of the restlargely as a result of salesat home..

Initial Japanese entry into they U.S. market wasbased on a combination of low prices and high qual-ity, with special emphasis'given the latter (chs. 5and 6). Although the Japanese were a minor factorin the case of the 4K RAM, they persisted in thisstrategy with the 16K RAM and other semiconduc-tor products.

The Market

Demand.As table C-1 shows, fewer than amillion 4K RAMsat $15 to $20 eachwere soldin 1974. Volume increased as prices broke the $10barrierdropping to $6 late in 1975and main-frame computer manufacturers began to buy inlarge quantities. Sales peaked in 1978, before 16KRAMs took over.

The companies involved grew rapidly as 4K RAMvolumes jumped. Intel's sales in 1970 totaled only$4.2 million; by1974 they were $134.5 million, andby 1979 had reached $663 million. This was not alldue to the Li K RAM, but that device played a major.role.

Distribution.Within the United States, mostsemiconductor firms sell directly to large customers

Table C1.Worldwide Sales of4K Dynamic MOS RAMs

YearSales

(millions of units)1974 0.71975 5.01976 28.01977 57.11978 76.51979 69.21980 31.21981 13.0SOURCE: Dataquest.

as well as through independent distributors. Dur-ing the mid-1970'S, many of the firms producing4K RAMs were rather small, with little marketingexperience. However, a well-developed network ofindustrial distributors such as Arrow Electronicsand Hamilton/Avnet served the many smaller cus-tomers for memory products.

Product Development

Top managers in semiconductor firms devote agreat deal of attention to product and process de-velopmentthe two go togetherbecause of therapidly evolving technology. Many industry ex-ecutives have technical backgrounds.

Planning.At Intel, product planning commit-tees are organized for each of the firm's "strategicbusiness segments." The committeese.g., that forRAMsoperate with a 5-year time frame. Planningresponsibilities may take a third of a committee

.member's working hours. Intel's approach has beento look for high-growth products where the com-pany's technology can provide an advantage. Pro-posals emerging from the planning process are pre-sented to an executive group that includes thechairman and visx-chairman of the board, the presi-dent, and the vice presidents.

Texas Instrumentsanother technology leaderemphasizes projecr-orientod teams for planningfuture activities, while. a more conventionaloperating hierarchy looks after current operations.

Not all firms in the industry try to be innovators.Instead, managements `may opt to become alternatesources for products introduced by others. Thisstrategy is sometimes dictated by costssince theextensive research and development (R&D) neces-sary to come up with a proprietary design mayseem too risky, particularly for a company withouta position of technical leadership. It does requirethe ability to duplicate (and perhaps improve on)the device in question, and get into productionquickly.

In the case of the 4K RAM, American entrants'followed one of two approaches to R&D. In the firstgroup were firms such as Intel and Mostek, whichattempted to take the lead, hoping that their designswould become de facto industry standards. In thesecond were companies like Advanced Micro De-vices, that aimed at becoming alternate supplierswith a competitive advantage in attributes such asquality or performance. Technical leadership in thesemiconductor industry requires two kinds ofscarce resources: money and skilled engineers. Thechoice of strategies depended on these, but evena second-source supplier needs clever designers

528 International Competitiveness in Electronis

and still more, competent process engiciet\s. Dur-ing the 1960's and 1970's, successful U.S. merchantfirms were sometimes built around the abilities ofthree or four inventive circuit designers, Even so,the R&D emphasis in many firms during develop-ment of the 4i: RAM was heavily on process tech-nology. The case of Advanced Micro De Vices(AMD) illustrates the point.

The Example of Advanced Micro Devices.AMD was founded in 1969 with an initial capitalinvestment of $1.5 million. Research, design, anddevelopment activities began immediately; by theend of the first year half a million dollars had beenspent, although sales had not begun. R&D playedonly a limited role in the strategy adopted by thecompany. The president and chairman of the boardW. J. Sanders, Illwhile an engineer, had re-signed a position in marketing at Fairchild to startAMD. Sanders chose to emphasize second-sourcingof chips developed by larger firms. Not only didAMD have limited funds for developing new prod-ucts, but initially the company had no proprietarytechnology.

Product design and development throughout theindustry was almost exclusively a technical activi-ty during these years. Marketing research was in-significant by comparison. Neither Intel nor AMDhad internal marketing research staffs. One reasonwas a pervasive 'feeling that production capacitywould limit total sales over the foreseeable future.

During its first few years of operation, AMD fol-lowed a strategy of introducing devices that couldbe put into production quickly to serve existingmarkets; R&D spending remained low until 1974,when it reached about $1.5 million. The companytried to concentrate on high-volume chipsfor ex-ample, targeting customers who might be able togrow rapidly in their own industries, which in turnwould permit AMD to expand more rapidly thanits competitors. At first, the firm concentrated itssales efforts on 25 to 30 customers worldwide. Inthe late 1970's, AMD began to modify its approach,Pursuing new products of its own.

The choice of integrated circuits to second-sourcewas Critical for AMD. To fit the company's productdeVelopment strategy, a proposed new integratedcircuit would: 1) be marketable in high volume at'a price attractive to AMD's customers, implying;2) that it would be complex enough to be a cost-effective substitute for existing devices, but not;3) so complex that it became, on the one hand, dif-ficult to make, or, on the ether, so specialized asto limit its market. The essential links are betweendesign engineersthose at the semiconductor firm

and those at the customerrather than betweensales staff and purchasing department.

The three fundamental steps in producing inte-grated circuitswafer fabrication, assembly, andtestingare now all essentially mass productionprocesses. During the peak period of 4K RAM pro-duction, however, assembly and testing were botl'quite labor-intensive. Because of this, AMDlikeits counterparts in the U.S. industryhad estab-lished offshore plants in low-wage countries.AMD's offshore facilities were in Manila and inPenang, Malaysia.

AMD's approach to the 4K RAM market typifiesits strategy during the mid-1970's. The firm pro-duced two 4K chipsone an 18-pin design with twopower supply voltages, the other a 22-pin part re-quiring three voltages. Both were interchangeablewith 4K RAMs manufactured by TI, but AMDmade a number of design changes aimed at reduc-ing power consumption, improving noise immuni-ty, and meeting military standards. The last hasbeen a centerpiece of AMD's marketing approach;by advertising that all its chips met militaryspecifications, the firm sought to establish an im-age of high quality and high reliability. AMD's um--phasis on making modest improvements in theproducts they chose to manufacture, adhering thigh quality standards, and concentrating f7n s+ l-

ard devices foreshadowed the Tap nese ratew.if fM,V airs later.

Demand for AMD's 4K RAMs came mostly fromcomputer companiesabout 10 in numberalongwith another 150 firms manufacturing,systems andequipment ranging from typesetters to computerperipherals, and including a half-dcii-en telecom-munications accounts as well as 10 or 12 militarycontractors. Each customer was, potentially, a high-volume purchaser. A mainframe computer with 8megabytes of memory, for instance, needed 18,000to 20,000 4K RAMs.

Pricing and ProfitsAs part of its overall strategy, AMD attempted to

hold its prices somewhat above those of the com-petition by stressing quality. Prices for semiconciuc-tor products tend to be high at first, declining rapid-ly as production volumes and the number of en-trants grow. Manufacturers sometimes set pricesin anticipation of future cost reductions. Eventual-ly, product obsolescence puts still more downwardpressure on prices. At the time the 4K RAM wascoming onto the market, the semiconductor indus-try was in a deep recession, leading to even moreprice cutting than normal.


Intelone of the acknowledged technical leadersalso charged premium prices, but on the basis ofoffering the most up-to-date products; the companycontinues to pursue a strategy of building uniquecircuits when possible, thus maintaining healthyprofit margins. Intel has, in maw: years, beentunong the most profitable companii indus-trytable C-2. The tables lists profit els for anumber of U.S. merchant ':irms in 19Th, the peaksales year for 4K RAMs ai,,l a generally good onefor the industry; profitability ran somewhat aheadof that for U.S. industry as a whole, represented bythe Fortune 500 average.

The 4K RAM Lifecycle

Before Intel's pioneering 1K RAM entered themarket, semiconductor firms often simply copiedeach other's products. Customer demands for alter-nate sourcingmore than one suprl:f, lot a givenpartprovided an easy avenue into titi: ittarket fora new company. Mostek's Executive Vice Presi-dent, Berry Cash, remarked: )ne usedcopy everyone else_ About the o;,.'y do, you coulddo when you got sei,liething good vv.tls run like helland work on new pi av'ticts to obsolete it.-17

This pattern changed, partly as a result of ex-perience with the I K RAM. A number of firms triedbut failed to duplicate Intel's chip; after 3 years onlytwo or three other companies had learned to buildit. As a consequence, companies began to negotiateformal alternate sourcing agreements for the nextgeneration 4K RAM. Thiough these agreements,firms could acquire design rightsand sometimeslithographic masks. Thus, as pointed out earlier,

l'"Boorn Times Again for Semiconductors,- Business Week, Apr. 20,1974. p. fib.

Table C2.Profit Levels for U.S.Semiconductor Firms, 1978

After-tax earning.;As percent

of salesAs percentof equity

Advanced Micro Devices 7.1% 17.6%Fairchild Camera and Instrument 4.6 12.0Intel 11.0 21.6Mostek 7.1 15.8Motorola 5.6 16.6National Semiconductor 4.6 17.1Texas Instruments 5.5 17.6Unweighted average ....... 6.5% 16.5%Average for Fortune 500

indwitrial firms 4.8% . 1A.3%SOURCE: Annual reports.

Fairchild negotiated a second-source agreementwith Mostek, while Motorola and American Micro-systems worked jointly oi, 4K RAM development.

During the early years of the 4K RAM productcycle-1973-76Iniel enjoyed a major share ofsales, hi:' the end inn, observers rated MostekCie 'win' tut of Lie 4X RAM competition.

(any other entrants henefitc I in terms of profits__and demonstrated viability in the rapidlgrowingmemory market. Mostek's success was dle not onlyto customer acceptance of its 16-pin design, but alsoto the head start it got when Tls' 22-pin device en-countered production problems.

The situation in 1977, the year before outputpeaked, is illustrated in table C-3. While Mosteksold the most of any one design, TIs' total 4K RAMsalescpread over three designswere slightlygreat,:, There wi;, n ,i,1;!.. no losers, especiallysince ru,ii.,,facturing capacity constrained sales.;Ionetheless, Mostek's 16 -pin RAM found the great -cst eventual acceptance in the marketplace; 1977'},, veer co' .,. ')v the tablemarked the sales

ur 22- -in told while the 16-pin alternativedid not peak until 1979. The world market shareof 4K RAMs for Japanese firms was 18 percent in1977, with NEC the clear leader. The Japanese weresplitting their efforts between 16- and 22-pindesigns.

Table C-4 gives market shares from 1977 to 1981.For the first years, AMD's second-source strat-egy led to an increasing proportion of a decliningmarket, while Intel's share declined in part becauseit began moving into new products. The marketshare of Japanese firms actually fell over this peri-od. By 1980, several manufacturers had begun toabandon the 4K RAM market.

Conclusion

The 4K RAM reached its unit sales peak in 1978(table C-1). Dollar volume had been greater the yearbeforea common phenomenon in the industry.While volumes have since tapered off, 4K RAMswill continue to be widely marketed at least throughthe mid-1980's. Where a dozen companies made thedevices in 1980, the number has since been cutperhaps in halfthose who can still make a reason-able margin on sales remaining. The lifecycle of the4K RAM proved somewhat longer than that for IKchips, illustrating a trend toward lengthening prod-uct cycles for RAMs that is expected to continue.One factor in the longer lifecycle was strong pricecompetition; as 4K prices fell, mass acceptance ofthe next-generation 16K RAM was delayed. Onlywhen 16K prices came down to the point where one


Table C-3.-Estimated Worldwide Sales of 4K Dynamic MOS RAMs, 1977

Shipments (mil'ioris of units)16-pin 18.pir, 22.pin Total

United States:Advanced Micro Devices - 0.55 1.84 2.39Fairchild 2.08 - - 2.08Intel 2.0 - 8.4 10.4Intersil 0.5 - 0.4 0.9Mostek 11.8 - - 11.8Motorola 1.55 - 1.19 2.74National 0.38 - 3.3 3.68Signetics 0.54 - 0.33 0.87Texas Instruments 0.9 5.9 5.6 12.4

U.S. total .-:-.---..".16.8 (82%) 6.45 (96%) 21.1 (80%) 47.2 (83%)

Japan: -----_.''Fujitsu 1.8 - 1.1 2.9Hitachi 0.45 - 0.46 0.91Nippon Electric Co. (NEC) 2.15 0.25 3.7 6.1

Japan total 4.4 (18%) 0.25 (4%) 5.26 (20%) 9.91 (17%)

World total 24.2 6.7 26.4 57.1

SOURCE: Dataquest.

Table C-4.-Worid Market Shares of 4K Dynamic MOS RAMs

Share of unit sales1977 1978 1979 1980 1981

United States:Advanced Micro Devices 4.1% \ 8.6% 14.6% 9.4% 12.7%Fairchild 3.6 \ 1.2 - -Intel 18.2 114.4 8.7 3.2Intersil 1.6 . \ 0.5 1.4 3.9 1.1

Mostek 20.7 22.2 20.1 22.8 17.3Motorola 4.8 . \7.4 9.2 16.2 24.9National 6.4 7.3 11.3 14.6 . 16.1Signetics 1.5 1.5 0.7 - -Texas Instruments 21.7 21\8 15.3 2.6 7__.

U.S. total 82.6 % 84.9 % 81.3%, 72.7%.....

72.1%

Japan: \'Fujitsu

'its chi5.1%1.6

2.5°2.3

1.1%1.2

2.1%0.6

1.5%-'ippon Electric Co. 10.7 8.0 7.9 4.7 2.5

Japan total.... 17.4% 12.8% 10.2% 7.4% 4.0%

Europe:In 2.0% 7.4% 16.0% 18.9%SGS-Ates 0.4 \ 1.1 3.9 5.1

2.4% 8.5% 19.9% 24.0%SOURCES: 1977-table C3.

1978.1981-Dataquest.

App. CCase Studies if. the Development and Marketing of Electronics Products 531

chip cost about the same as four 4K devices did thenew generation parts begin to take over. Similarforces were at work during the early 1980's as 64KRAMs entered the marketplace.

Computers: A Machine forSmaller Businesses

The Product.

Before 1970, the computer industry was domi-nated by a few relatively large manufacturerswithIBM holding by far the greatest market share. Asthe decade progressed, advances in hardware cre-ated numerous opportunities for newer firms to sellsmall computers in markets as yet untapped by es-tablished mainframe-oriented companies. The newentrants at first aimed their minicomputers atoriginal equipment manufacturers (OEMs) and atsophisticated customers who could put small proc-essors to work in science and engineering. Betweenthe minicomputer market and that for large, generalpurpose mainframes lay a vast pool of potential cus-tomersmany of them small businesses largelyunfamiliar with the esoterica of computer hardwareand software, and without the capability to plantheir own data processing im!allations. Companiesfrom both the mainframe and minicomputer por-tions of the industry began to design small businesscomputers (SBCs) to attract such customers.

Small business systems range in price from about$5,000 to perhaps $100',000. Typical installations in-clude a central processing unit, one or more ter-minals for input, disk storage, and a serial or lineprinter for hard copy output. By the late 1970's, 80to 90 suppliers were marketing nearly 300 differentSBC systems." Among these, the IBM System /32the focus of this casefell near the middle in costand features. When first introduced, the System/32could be leased for $770 to $1,085 per month, orpurchased for $33,100 to $40,800. It had been de-signed for businesses with sales in the range of $1million to $10 million, and as many as 200 or 300employees. The complete systemconsisting of thecentral processing unit, up to 32 kilobytes of mainmemory, a keyboard, display, printer, a single flop-py disk drive, and a nonremovable hard disk formass storagewas housed in a desk-sized enclo-sure. Software was unbundled, with everything butthe operating system sold separately. In 1978, thesoftware available included three programing Ian-

"Datapro Feature Report: A!! About Small Business Computers (Del ran.N.J.: Datapro Research Corp., September 1978), p. 70C-010-30a.

guages and a series of industry-specific applicationspackages.

Hardware.Thirty-two models of the System/32were available, differing mostly in the capacity ofthe hard disk-3.2 to 13.75 megabytesand printerconfiguration." Printer options included a serialprinter with speeds ranging from 40 to 80 charac-ters per second, and line printers of 50 to 155 linesper minute. The basic machine came with 16 kilo-bytes of memory; model changes could be made inthe field.

The System/32 could be operated in batch or in-teractive modes and also function as a smart ter-minal or a satellite processor linked to other com-puters. For example, a System/32 can easily be setup to communicate with: another System/32; anink-jet document printer; an IBM Office System6/430, 6/440, or 6/450; an IBM Mag Card II type-writer; IBM Systems /3, /7, /360, or i370; some ofthe equipment in a 3740 Data Entry System; or a5230 Model 2 Data Collection System.

Software.IBM supported the System/32 by reg-ularly offering new software. The three program-ming languages available were: Report ProgramGenerator II (RPG II), a commercially oriented lan-guage; COBOL; and FORTRAN IV. A utility pro-gram aided in file preparation and management;other software supported the communications fea-tures mentioned abovee.g., use of the System/32as a remote work station for a 370 series main-frame. Other miscellaneous software included:word processing; form letters; a library of math-ematics subroutines; statistics; critical path analy-sis; and a manufacturing management package forscheduling purchases, fabrication, and shipments.

Much more software was made available throughthe 14 Industry Application Programs (IAPs) sup-plied on IBM-owned floppy disks and written inRPG II; these could be customized still further ifnecessary. The 14 IAPs handled tasks associatedwith: accounting firms; medical groups; bulk mail-ing; construction; hospitals; manufacturing; distri-bution; law firms; lumber dealers; food distributors;student administration; motor freight; financial in-stitutions; and retailing. Typical IAP functions areaccounting, analysis or control of cost/time/inven-tory/sales, management of files and records, andplanning and scheduling.

Upward compatability was one of IBM's designgoals. System/32 purchasers had two possiblegrowth paths: into a System/3 Model 8, 12, or 15;

"Datapro Reports on Minicomputers: IBAf System/32 (Delran, N.J.:Datapro Research Corp., January 1978). p. M11-491-601.


or into a Systeml34. With minor modifications RPGII programs from a System/32 could be run on aSystem/3 or vice versa; this also meant that Sys,tem/3 users could move into 32s or add 32s to theirnetworks. For those wanting to move into a Sys-tem/34, the System/32 RPG programs were source-compatible with the Systeml34, allowing IAPs torun without change.

Support by the manufacturernot only new soft-ware packages, but also hardware updates and serv-icingare important to most SBC customers. Thosewho rented the System/32 could get service 24hours a day, 7 days a week. Purchasers had serviceavailable 9 hours per day, 5 days per week underthe Minimum Monthly Maintenance Charge, orcould buy 24-hour service for an additional fee.IBM also emphasized ease- of usethe minimaltraining needed to run IAPs, indeed to operate theentire system.

The MarketThat the SBC market offered enormous growth

potential was a truism of the early 1970's. The poolof prospective customers for an SBC costing lessthan $1,000 per month included as many as half amillion small organizationsvirtiuiny none ofwhich had prior electronic data processing ex-perience. Besides the sheer numbers involved, thismarket was important for another reason. Com-puter customers exhibit high loyalty to the firmfrom whom they buy their initial system. By cap-turing first-time purchasers, a supplier gains a bigfuture advantage. Computer manufacturers whochose not to compete vigorously in the SBC marketran a real risk of seeing their future market shareand competitive position eroded. Adding to thepotential market were expanding applications indistributed data processing, where many SBCmodels could be used as remote job entry stations.

These markets brought a pair of reqUirements fora competitive SBC. First, it would have to besimpleuser-friendlyso that a customer with lit-tle or no data processing background could learnto operate it quickly. The second requirement wascompatibility with other machines in a networkingor distributed processing environment. Upwardand for larger SBCs, downwardcompatibility wasan important selling point. so that customers couldexpand or upgrade their installations.

The established mainframe computer manufac-turers had, at least initially, advantages in all theseareas. Their nationwide sales and service staffswere accustomed to dealing with customers hav-ing business applications, rather than the OEMs

and technically trained users who bought minicom-puters. They also had considerable software experi-ence; mainframes were ordinarily marketed withextensive software support compared to the mini-computers of that era. Moreover, a mainframe man-ufacturer could design an SBC compatible withother parts of its product line; existing customersthen comprised a readyinade market base. As afinaland very importantweapon, the large, es-tablished firms had brand recognition. Not onlyIBM, but companies like Burroughs, NCR, and Uni-vac were familiar names. Many new purchasers,bewildered by competing claims and fearful of thepitfalls involved in purchasing a computer, auto-matically turned to one of these companies. A dec-ade later, IBM reaped similar benefits when itentered the personal computer market. Despitethese putative advantages, most of the establishedmainframe manufacturers had a good deal of trou-ble adjusting to the competition for SBC sales.

Minicomputer firmsfor many of whom SBCswere upward rather than downward extensions oftheir product linesfaced serious handicaps incomparison. Unlike the mainframe companies,minicomputer suppliers such as Digital EquipmentCorp. (DEC) had little experience selling to end-users. OEMs or engineering organizations did notneed extensive support; minicomputer firms hadneither large service networks nor large marketingstaffs. They competed most heavily on hardwarefeatures and price. Software was less critical; mostusers could write their own. Minicomputer custom-ers who needed software or other support frequent-ly bought from "systems houses" or other middle-men. Systems houses purchased hardware in bulk,supplying customized software and assembling asystem to meet customer requirements. Not re-stricted to any one manufacturer, they could puttogether processors, terminals, storage units, andother peripherals from a variety of sources to cus-tomize a given installation. By taking advantage ofthe lower hardware prices it could command, a sys-tems house miet be able to supply an entire instal-lation, including ,of ware, for less than the hard-ware cost to .a si:,-'1E-unit purchaser. Qantel andBasic Four both had considerable success as sys-tems houses before entering the SBC market withequipment of their own designs.

Minicomputer makers also experimented withother marketing channels aimed at small end-users.An example is the "software representative" cre-ated by Datapoint to locate potential-customers.The sale was between Datapoint and the end-user,with the representative getting a commission andafterwards supplying software and other servicesindependently.

App. CCase Studies in the Development and Marketing of Electronics. Products 533

Development of the System /32

IBM had been slow compared to Burroughs andNCR in exploiting the SBC market. Prior to the in-troduction of the System/32 in January 1975, IBM'sonl. offering was the Systeml3. Originally pricedat t`.., ery top of the SBC rangealthough laterme rte cost much lessthe Systeml3, whichre. 11,:qi .111e market in 1970, had gone on to sellmel th,in any other computer in IBM's his-tor ; .11 1075, ever 30,000 had been purchasedworldwide. The success of the System/3 was amajor retnon for IBM's decision to expand its'' leof SI3C into the lower price ranges.

In many respects, the System/32 was a direct de-scendent of the System/3 Model 6. The central proc-essing units of the two were quite similar, andmany of the software products offered for the Sys.tem/32 were adaptations of those developed for theSystem/3 Model 6. Changes included faster print-ers, simplified operation, and improved applica-tions programs. With the System/3 Model 6, onlylimited applications software had been available,and those wishing to write their own programs hadto master a complicated operating system.

Development of the System/32hardware, soft-ware, and the market studies leading up to thesewas the job of IBM's General Systems Division(GSD). The GSD emerged from a major corporatereorganization in 1972 that split the former DataPressing Group into three divisions. With this re-orga,,ization, the GSD was given development andmanofacturing responsibility for the System/3 andrelated peripherals. Responsibility for all small.business applications within IBM followed in 1974,at which time the GSD was given its own marketingarm.

I3y the rvxt year, the GSD marketing and salesforce had grown to some 4,500 sales representativesworking out of 67 sales offices, plus nearly 3,000field engineers. The System/32 entered the marketaccompanied by an extensive advertising cam-paign, along with exhibits at trade shows, directmail, and iM:person sales, calls. These promotionalefforts were tailored to potential customers with lit-tle or no computing experience. Initial sales of the

competition, but IBM's accurate perception of theneeds and concerns of the SBC market. Indeeci,from a hardware perspective the System/32 was arather limited machine. Instead of being designedfor multiprogramming, it was restricted to ex-ecuting one program at a time. It had less disk stor-age than a number of other SBC systems. Further-more, because the disks were hard and nonremov-able, only on-line storage was available. The tech-nology utilized in the System/32 did reflect the stateof the art in using MOS integrated circuits in bothmemory and processor. -

Probably the most innovative aspect of the Sys-tem/32 was its software. While IBM had been ac-customed to writing customized software for main-frame purchasers, such an approach was not prac-tical given the large number of SBC customers.Hence the Industry Application Programs, aimedat meeting perhaps three-quarters of user needs.The remainder could be supplied by IBM or an in-dependent vendor at extra cost. The IAP conceptwas not unique, but the design, distribution, andsupport for these programs was a major undertak-ing. Unbundling the software was another new de-parture; IBM had traditionally supplied hardwareand software together at a single package price.IAPs, in contrast, were sold for an initial one-timepayment plus a monthly support fee. The WholesaleFood Distribution IAP, example, carried an in-itial charge of $3,120, plus $147 per month for sup-port. By emphasizing reliable hardware, imaimaimaintenance, and off-the-shelf software, IBM WESable to continue its "116nd-holding" approach tomarketing while supplying large numbers of ma-chines.

The Competitive ResponGe

Burroughs and NCR were4e two companiesmost affected by IBM's entrarice into the SBCmarket. While both offered -broad product lines,SBCs had come to represent a significant share oftheir total revenues. Both had seen the importanceof SBCs early, and sought to utilize their sales andmarketing organizationswhich were much moreextensive than those of the minicomputer suppli-

System/32;whichwasmadeinRochester,- Min- --ersto establishthemselvesinthispartofthe--nesota and Virmercate, Italywith components market.and subassemblies coming from other IBM facili- Burroughs, then the dominant force in SBCs, hadtiesexceeded the company's projections. formed two special marketing groups the "general

The System/32 offered a price-performance com- accounts force," and the "selected accountsbination not available in other IBM systems. Heavy force"to handle smaller machines. The general


be served by small computers. Thus Burroughs ex-plicitly recognized the dual nature of the SBC:stand-alone for the small enterprise, distributedprocessing for larger customers.

Burroughs had originally entered the lower endof the SBC market in 1973 with the first of its B700'series, selling more than 2,000 in the first 31/2 years.When the System /32 was introduced, Burroughs re-sponded immediatelydoubling the main memoryof the B700 and bringing out six new models. InApril 1976, Burroughs announced the B80, whichwasunlike IBM's System/32capable of multipro-graming and multiple terminal support. This ma-chine was well received initially, but suffered fromsevere software probl -ms; i! was soon replaced bythe 1190. Burroughs' share of the SBC market beganto slip, a. considerable concern to a company that,as early as 1973, got 30 percent of its revenues fromthe low end of its productline.20

NCR also made an early entry into the SBC market.ln 1972 it had introduced the NCR 322, a mini-computer priced in the $15,000 range, followed bythe Century 8200, the firSt of a series of SBCs. Thesetwo models represented the first results of a thor-oughgoing and painful reorganization at NCR, acompany that few observers at the time believedcould survive. NCR was seen as tradition-bound,still producing electromechanical products thatcould not compete with electronic equipment.Then, under a new president in the early 1970's,NCR invested nearly $300 million in product de-velopment, one thrust being interactive systemsdesigned specifically for business applications.21 Aturnaround followed, as the company went froma loss of $60 million in 1972 to earnings of $72million in 1975.

The new commitment to electronic products alsobrought changes to NCR's marketing organization.The old system of branch offices was dismantled,to be replaced by a "vocational sales" organization.Under the earlier system, each salesperson hadbeen responsible for a group of products: some-times two NCR representatives found themselvescompeting for the same sale. Under the new ar-rangement, salesmen were responsible for sellingto one of four vocational groups: retail stores; finan-CIO organizations; commerciaThindustrial enter-prises; and a residual group consisting of medical,educational, and government organizations. In ad-dition, the entire field engineering forcesome

7,000 peoplewas retrained to service the newelectronically based product line.22 ,

Sperry Univac was the last of the major main-frame manufacturers to move into the SBC market,introducing the BC/7 in 1977a machine featuringmultiple terminal concurrent data entry capabili-ty, a great deal of available storage, and removabledisks none of which were available on the System/32. Sperry Univac had created fully staffed mar-keting organizations in 18 cities, with plans for fur-ther expansion, just for the BC /7.23 A further indic-tion of their commitment to the SBC market wasthe acquisition of Varian Data Machines, a majormanufacturer of minicomputer products. Nonethe-less, 'the BC/7 family suffered from applicationssoftware that did not compare favorably with thecompetition, and could, capture but 2 percent of theSBC market.

Among the minicomputer firms, DEC was andstill is the largest and most successful. The com-pany. which had developed the first commerciallysuccessful minithe PDP-8probably had the mostto lose in competing for SBC sales. DEC had estab-lished itself by mass-producing "black boxes" soldprimarily to OEMs. In the 1970's, this market waycoming under increasing pressure from other com-panies, including those making microcomputers,and DEC realized its greatest growth prospects layin small business and other end-user markets.

When the System/32, was introduced, DEC wasthe first to respond, counteringonly 10 days afterIBM's announcementwith the Datasystem 310,which played to DEC's own strengths. It wasslower, with less memory than the System/32, butcost a third less. DEC retained its establishedmarketing practices, selling networked systemsdirectly while relying on independent distributorsfor simple turnkey sales. These distributors boughthardware at a discount, added software, and thensold the systems at approximately the same priceDEC would have charged for the hardware alone.After purchase, DEC provided hardware mainte-nance, with the distributors responsible for soft-ware.

Another minicomputer manufacturer, Warg Lab-oratories, also responded quickly, releasing a seriesof computers=the WCS series,--that proved quitesuccessful.24 Like other minicomputer manufac-turers, Wang stressed its low prices and proven

u"NCR's New Strategy Puts. It in Compulers to Stay," Business Week,


hardware. However, the company realized thatprice competitiveness alone would not assure suc-cess, and moved to establish a dealership and serv-ice network to provide the services that SBC cus-tomers expected. One function of Wang's new dis-tribution system was to provide applications soft-ware, including customized programs.

Conclusion

When IBM moved into the SBC market, otherfirms rapidly- cut prices on existing models andsought to upgrade and expand their product lines.More companies entered the fray, perhaps feelingthat IBM's entry had :9gitimatized the SBC. Buyerscould choose from more sophisticated systems atlower prices. Business Week estimated that, by1975, IBM had captured about 28 percent of SBCsales, with Burroughs around 12 percent and NCRjust under 5 percent. IBM's share of this marketcontinued to climb; by 1978 it was put at 37 per-cent, with Burroughs and NCR together still ac-counting for less than 20 percent.25-

z''''Fhe Syndrome.- op. cit.

By 1980, the Systemi32 in its basic configurationsold for $23,490a reduction of $10,000 comparedto the original price 5 years earlier. While the'System /32 was more successful at the outset thananticipated, sales declined rapidly once its suc-cessor, the System/34 was introduced. The System/34 could handle multiple work-stations; it offeredmore processing power, multiple programing capa-bilities, and more storage. Selling in.the same pricerange, the System/34 continued the trend towardgreater performance/cost ratios.

Beyond its brand recognition and "safe" image,IBM's immediate success with the System/32 camefrom its decision to stress applicationsan obviousstrategy, but one that IBM executed better than thecompetition. The technology in the System/32 wasnot much different from its predecessor. In the SBCmarket, technical wizardry counted for little com-pared to cost and convenience.

Index

Advanced Micro Devices (AMD), 190, 191, 192, 269,-108. 433, 527, 528

AVI.C10, 332Alabama, 128Amdahl. 205, 211. 269. 271, 433American Electrenics Association, 309American Microsystems. Inc., 526American Society for Training and Development, 312Ampex. 70Apple, 103. 148, 205: 212Atari. 268AT&T, 14. 27, 168, 139, 179, 189, 313, 465Audio. 526Australia, 135

BASE, 211.13(41 Laboratories, 27, 108. 139, 179, 189, 313. 46513raniff International. 258Brazil, 57. 135, 151. 166, 206, 388Burroughs. 82, 134, 147, 201, 212Business Week. 535

Canada, 136, 149, 175Carver, Robert, 520-524Cash, Berry. 529China, Peoples Republic of, 185, 387

China Technical Services Corp.. 387Electric Power Ministry. 387industrial policies. 387Information Processing and Training Center. 387National .Electric Technology Import Corp., 128National Plan for Development of Science and

Technology, 387Stith: Administration of Computer Industry, 387

Cincinnati Milacron, 24(1, 242coin pet it iVelleSS

in computers, 15-19, 28, 149, 200-212Japan

government assistance, 208marketing strategies and multinational

operations. 210United States

impacts of microelectronics and reliability,201.204

international aspects, 205product strategies, 204strategic patterns, 201

in consumer electronics, 5, 11-13, 112, 115, 160-186American response, 182Japan, 180

__dumping, 168,199economics of. 164-168

comparative advantage, 164evaluation of, 177-179

international business strategies, 179market distortions. 168market share, 173in semiconductors, 13. 24, 25, 27, 28, 132, 133. 137,

187-200current trends, 194international dimensions, 192Japan, 196-200

linear to digital circuits, 196strategy and structure, 198

mergers, 195products and prices, 190United States, 188-196

technology, role of. 169computers

Amdahl 470-V, 81applications of, 90, 91central processing unit (CPU), 82, 83, 85, 99characteristics of different categories, 83characteristics of generations, 90in China, 387competitiveness, 28, 149, 200-212Cray-1. 93distributed processing. 85Eniac, 88in Europe. 152Fairchild F-8, 88growth of computing power, 87IBM System/32, 531-535IBM 370/168, 99IBM 650, 87, 88in Japan (see Japan)mainframe, 83, 84memory, 99-101microcomputer, 83, 84minicomputer. 83, 84

price trends. 89networking, 85PDP-8, 88, 146process control, 102small business computers (SBCs), 531-535software, 85, 531

costs, 86development, 206product strategies. 204reliability, 220

systems aspects, 29types of, 82Univac I, 81, 87, 101U.S. industry, 145-150

competitiveness, 2001 -212evolutionT147-exports, 150imports, 149, 150


production. 29sales. 146. 149, 202structure, 146

world markets. 151congressional interest. 7. 9, 36. 46, 270consumer electronics

amplifiers. high-fidelity. 520-524in Asia. 126, 127. 128automation (see manufacturing)cathode-ray tube (CRT). 68, 71,-102color TV sales, 114competitiveness. 5. 11-13. 112. 115. 180-186employment, 114. 353exports. 118in Japan (see Japan)offshore U.S. TV manufacturing plants, 118, 510produil differentiation, 23research and development, 68, 69, 168sit ucture, 58, 114TV receiver componentry, diagram of, 69U.S. sales and imports, 112, 114, 116U.S. TV 'manufacturing facilities, 114video cassette recorders (VCRs), 68, 70, 112, 122,

156, 165, 180, 186/ video disk, 70

wage rates. 127Consumer Reports, 231. 232Control Data Corp., 201, 206, 212, 268Costa Rica. 185Cray, 206, 271

Data General, 200, 20.5Datapoint, 204, 205Deming, W. E., 224Digital Equipment Corp. (DEC), 31, 88, 111, 146, 147,

179, 201, 204, 205, 240Domestic International Sales Corp. (DISC), 445dumping, 42,168, 169, 'n7.. ,,39

education and training, 35, 36, 302-318international differences, 317 -318secondary school education

in Japan, 304in UnitedStates, 3G3

university and continuing education, Japan, 314company-run training programs, 316continuing education and training, 315engineering, 314enrollments in colleges, 315

university and continuing education, United. States, 305, 312community colleges and local initiatives, 311engineering.. 306 -308engineering technology, 310industry initiatives, 309sunnlv and demand. 308

Electronics Education Foundation, 195employment

in computer manufacturing, 40in consumer electronics, 39, 40effects of import penetration and offshore

assembly, 359-363consumer electronics, 360foreign investment in U.S., 363semiconductors, 362

future patterns, 366-372impacts of technical change in electronics, 344-349

automation debate, 344factors affecting employment levels, 347

in semiconductors. 356trends in U.S. electronics industry, 349-358

computers, 353, 357consumer electronics, 351, 353semiconductors, 352, 356

European Community (EC), 43, 143, 153, 185, 411,412, 432, 433, 434

European Economic Community (EEC), 107, 123, 124European Free Trade Association (EFTA), 433exports, U.S.

color TV receivers, 119computers, 150semiconductors, 135, 136

Fairchild Camera and Instrument, 72, 140, 143. 191,195, 288, 433

Federal policies affecting electronics: U.S. options,463-502

alternative perspectives on industrial policies,467-500

competitiveness of American industries,promoting the, 491open trade and industrial policy, 492product cycles and structural adjustment, 491

critical industries, support for, 475capital for investment, 482national defense, 477U.S. economy, 480

infrastructural and adjustment policies, 485adjustment, 488design and implementation, 489

private sector, 496-500protecting domestic markets, 468-475

bilateralism in trade, 469market-protection strategy, 474pros and cons of protected market, 470

current environment, 464-467Ferranti, 142financing---accelerate-d-depreciation 32, 279;-295;- 500--bankruptcy risks, 263Chase Financial Policy study, 256, 257, 26 'S, 261,

Index 541

effects of financial leverage on tax payments, 262financial structure-international comparison,

268-294France. 289-292japan

continuing change. 286-289external funding, reliance on, 281-284government involvement, 284-286postwar trends in financial system, 280

United Statescapital supplies and financing costs for

electronics industry, 277-280entry, costs of, 270financing startups and growth, 274-276internal and external sources of funds, 272-274venture capital, 31, 268-270

West Germany, 292-294inflation and banking practices, 260-261, 278internal and external financing, 257international differences, 32, 33, 34R&D tax credit, 32, 168risk, 258

aversion behaviors, 259compensation for, 259subsidization, 260

risk absorption in Japan. 262savings rates, 264size and diversification, 263

Ford Motor Co., 110, 332France, 46, 123, 124, 151, 175, 268, 281

CII-Honeywell Bull, 55, 107, 126, 152, 291, 382,394, 396

Commissariat du Plan, 401Cornpagnie Generale d'Electricite, 394Compagnie Internationale pour L informatique

(CII), 291..395Electronics Industry Task Force, 399French Atomic Energy Commission, 398Government Program for Development of

Electronics, 399industrial policies, 394-400Institut de Developpement Industriel (IDI), 395Le Plan Calcul, 291, 395, 396, 397Le. Plan Circuits Integres, 481Le Plan des Composants, 190, 396, 397Le Plan Periinformatique, 398Ministry of Economy and Finance, 395Ministry of Industry, 394Ministry of Research and Technology, 394President Mitterrand, 394, 395, 398, 400Prime Minister Barre, 399Saint-Gobain-Pont-a-Mousson, 397

Fujitsu, 151, 153, 154, 158, 199, 201, 209, 211, 240,269

GCA Coro.. 110

Geareral Electric (GE), 12, 14, 107, 111, 122, 123, 133,147, 181, 183, 195, 245, 277, 332

General Motors, 331, 332Gold Star Co., 128, 384gross domestic product (GDP), 171, 172, 175gross national product ((GNP), 346, 348. 364, 383GTE-Sylvania, 115, 123

Hewlett-Packard, 111, 126, 148, 151, 200, 204, 205,210, 230, 247, 248, 249, 268, 313, 322

Hitachi, 151, 153, 154, 158, 209, 211, 240, 262, 277,317

Honda, 285Hong Kong, 126, 127, 193Hughes Aircraft, 313human resources

labor, 318-325comparison, U.S.-Japan, 325market trends, 319mobility, 321

in Japan, 323in United States, 322

organization and management, 326-334decisionmaking in Japan, 328management styles, U.S.-Japan, 333, 336-339Matsushita's purchase of Quasar, 338worker participation, 330

quality control (QC) circles, 331, 332preparation for work, 37quantity and quality, 34

engineering education, 35, 36

IBM, 14, 18, 27, 28, 42, 55, 59,.82, 86, 87, 95, 108,111, 125, 126, 132, 141, 144, 145, 147, 155, 158,179, 139, 194, 201, 205, 211, 227, 240, 271, 277,313, 322

imports into United Statescolor TVs, 24, 116, 183computers, 150consumer electronics, 182, 435

industrial policiescomparisons, 383-422

developing countriesChina, 387others, 388South Korea, 383Taiwan, 385

European Community, 412France, 394-400

finance, 395future prospects, 400Le Plan Calcul, 291, 395, 396, 400Le Plan des Composants, 397, 398, 400planning, 395recent developments, 398-400

Moan. 413-422


policymaking, 416institutional setting. 415recent trends, 421role of universities. 419supports and subsidies. 419

United Kingdom, 400-405early experiments. 401other policies toward electronics. 403recent context. 401research and development. 403

United States, 389-394antitrust, 390policymaking, 39:3procurement and R&D. 391taxat;on, 392trade and foreign economic policies, 390

West Germanyeffectiveness of policies, 412Fraunhofer Gesellschaft, 410future policies, 411institutional setting, 405policies toward electronics. 407policymaking processes, 406R&D support. 408

congressional interest, 7, 9, 46, 47, 48, 49critical industries, 46, 48domestic market base, 46. 47economic adjustment, 10education and training, 9Federal policies affecting, 467-500infrastructural development, 49lack of U.S. Ito; cy. 61OTA definition, 7tools of iqdustrial policy. 381

foreign investment controls, 382government procurement. 381investment grants, 381tariffs, 382

integrated circuits (ICs.), 23, 24, 25, 26, 27, 29, 32, 58,71, 101 (see also manufacturing;semiconductors)

design, 73-74, 77development, 72learning curves, 76-77manufacturing, 78-81, 96-99microprocessors, 95technology, 93-99types of, 93

Intel Corp., 31, 74, 190, 323, 324, 525, 526, 527International Brotherhood of Electrical Workers, 361international electronics industry, 107-159

competitiveness, 163-214J?panese industry, /.19-123

development, 11'i, 120, 184, 185government su2ports,.121

consumer electronics. 111. 112-129investment, 175offshore assembly in color TV, 116 (see also

offshore manufacturing)sales trends, 109semiconductors, 111, 112-129

wage rates and investments, 127-128Western Europe, 123-126

Phased Alternating Line (PAL), 123, 124sales, European, 126

International Labour Office, 344Intersil, 14, 33ITT, 184Iowa State University, 308Ireland, 205

JapanAgency for Industrial Science and Technology, 245Bank of Japan, 266, 280, 285, 286Computer Development Laboratory, 418computerscompetitiveness, 15, 151, 205, 206-212

government assistance, 208production and exports of VCRs, 25production, imports and exports, 18, 153, 154sales, 156, 157

consumer electronics, 23, 119-123. 180-182automation, 23color TV exports to the United States, 116competitiveness, 115, 117, 180-182development, 119dumping, 42, 182foreign markets, 184government supports, 121industrial structure, 121reliability, 23, 70, 229television, 69, 70, 115, 117, 120, 121, 122,VCR development, 24, 70, 122, 165wage. rates and investment, 127

Council for Science and Technology, 416Development Bank, 155, 208, 245, 286, 415, 420Economic Planning Agency, 415education and training, 304, 314, 315, 316Fair Trade Commission (FTC), 415finance, 33, 34 (see also financing)foreign investment, 175Foreign. Investment Council, 193industrial policies, 413Institute of Standards, 225international electronics industry, 107-159Japan Information Processing Development Center

(JIPDEC), 416. 417, 418Kansai Electronic.Industry_Develornent Center,...

181labor (see human resources)

Index 543

(M1TI). 119, 121, 140, 145, 154, 180, 193, 194,196, 197, 198. 206, 208, 209, 245, 415, 416. 417,415, 419, 420, 448, 449, 454

Ministry of Posts and T,lecommunications, 416,418

Pattern information Processing System (PIPS)project, 207i 417

Plant Export Policy Committee, 121Reliability Center for Electronic Components, 224Research Association for R&D on New Function

Elements, 48robotics (see manufacturing)Scieno and Technology Agency (STA), 415semi(- 1,(ctors

capc,1 investments. 138competitiveness, 13, 132, 133, 187, 190, 193,

tf":..].)0development, 139government role, 140, 193, 197international trade, 141major producers, 140, 158manufacturing equipment, 144production, 139protectionism, 193rates of spending by firms, 33sales, 44, 138. 198structure, 137

trade, 43, 431computers, 458consum:-:' electronics, 451dumping in the United States, 439Orderly Marketing Agreements, 446; 447semiconductors, 452tariffs, 432, 433, 434

VLSI project, 80, 145, 197, 198. 207, 208, 238, 416,417, 418, 454

Japan Electronic Computer Corp. (JECC), 155, 208,209. 245, 414. 420

Japan Robot Leasing Company, Ltd. (JAROL), 245J. C., Penney, 116, 181, 4,40John Fluke Manufacturing Co., 110

J, Nt. 224

Karatsu, Hajime, 224K Mart, 181Korea Development Bank, 384

legislationBurke-Haitke bill, 513Clayton Act, 390, 465

--Economic Recovery Tax--Act-of 1981 (ERTA), 32,51, 173, 276, 278, 279, 381, 392, 393, 497, 500

tz.molovee Retirement Income Security Act of 1974

Smoot-Hawley Trade Bill, 429Tariff Act of 1930, 443, 445Trade Act of 1974, 444, 446, 447. 448, 449, 450Trade Agreements Act of 1979, 45, 430, 431, 438,

439, 441, 443Trade Expansion Act of 1962, 184

Library of Congress, 101Linear Technology, 189Lucky Group, 128

Magnavox, 12, 70, 123, 181Magnuson, 271Malaysia, 135, 193manufacturing

automation, 233-246in electronics

consumer electronics, 236reasons for, 235semiconductors, 236

fixed and flexible, 234robotics. 239-246

international trenc s, 244in manufacturing, 240market growth, 242robots and jobs, 241

'technology, 239use of, 243

color TVs, quality and reliability of, 231-233integrated circuits, 78-81, 96-99, 226

acceptable quality level (AQL), 228, 229customer requirements, 227failures in, 227quality and reliability comparisons. 229testing, 226, 228testing and screening Japan, 228

Japanese approach, 224organizing for quality. 225quality consciousness, 224standards, 225

offshore (see offshoi manufacturing)quality, 219

organizing and managing for, 221reliability. 219, 220

importance of design, 222in United States and Japan, 237

Matra-Harris, 291, 397Matsushita, 24, 115, 120, 122, 126, 144, 184, 186, 210,

224, 225, 231, 238, 256, 262, 285, 316, 338,.451Mexico, 13, 57, 117, 135, 149, 151, 166, 193, 206, 388,

447, 510, 513Microelectronics & Computer Technology Corp., 195,

205microprocessors and memory, 74-75, 95Microsystems International Ltd., 526


National Advanced Systems, 149, 231National Aeronautics and Space Administration, 314Nationai Committee to Improve the Quality of Work

Life. 331Ndt !midi Si:ic,tite Center for Communications and

Electronics, 314National Science Foundation. 303, 313, 314, 320, 393National Semiconductor, 107, 108, 158, 189. 191, 211.

322, 526National Venture Capital Association. 269NCR {formerly National Cash Register), 82, 111, 132,

134, 135. 147. 155, 201, 270New York Times, 308New York University, 312Nippon Electric Co. (NEC), 137, 138. 151, 154, 155,

190, 209. 211. 262, 277, 387, 433, 526Nippon Telegraph and TelepL,one (NTT), 107, 138,

197, 261. 381, 418, 437, 481North Americ:irt Philips, 115, 156, 451North Carolina Schoot.61 Science and Mathematics,

312

offshore manufacturing. 137, 193, 226, 510-519alternatives to, 514, 518balance of payments, effects on, 511case example. 516-518economic impacts, 510employment impacts, 511motivations for investments, 513technology transfer. 512

Oki Electric, 108. 197, 481Olivetti, 205. 311. 397Orderly Marketing Agreements (OMAs), 13, 42, 62,

63. 112. 113, 117, 120, 121, 128, 169, 184. 356,360, 437, 446, 447, 448

Organization for Economic Cooperation andDevelopment (OECD), 344, 348, 430

PhaSe Linear, 180. 520-524Philco, 183, 236Philipines, 135. 193Philips, 70, 122; 142, 153, 186point-of-sale manufacturing, 135

Quasar, 12, 115, 231, 332

Radio Shack. 74, 148. 186RCA, 12, 13, 55, 63, 70, 107, 108, 111, 113, 114, 128,

132, 147, 172, 181, 183, 186, 199, 277, 282, 451RCA Institute, 312Reagan administration, 465, 466Remington Rand, 147, 201research and development

in computers, 207, 208, 209in consumer electronics. 68. 69. 168

Saturday Review, 521Schlumberger, 195Scholastic Aptitude Test (SAT), 303, 304Sears and Roebuck Co., 115. 116. 181, 440SELJINSA, 211semiconductors

acceptable quality level (AQL), 228, 229American dominance, 129analog circuits, 67, 73, 94, 230in Asia, 143, 193automation (see manufacturing)bipolar, 93. 94captive producers, 132, 133competitiveness, 13, 24, 28, 132, 133. 137, 187-200digital circuits, 67, 73, 94, 196employment, 137, 356, 357European industry, 141

sales. 142exports, U.S., 135, 136future developments, 81gallium arsenide, 81imports, U.S.. 135, 136integrated circuits (ICs), 71. 72, 73, 74. 81

design, 77, 78, 92, 130lithography, 79, 80manufacturing of, 78, 96-99production levels for captive manufacturers, 134quality and reliability of, 226, 229, 247technology, 93-99terminology and classification. 94types of, 93

international positions, 26Josephson junctions, 81large-scale integration (LSI), 73, 74, 93learning curves and yields, 76major producers, U.S. and Japan, 140major U.S. companies, 130, 131manufacturing equipment, 144market trends, 131merchant manufacturers, 132, 133metal oxide semiconductors (MOS), 27, 73, 78, 93,

94, 99, 188, 189, 397microprocessors, 73.74, 75, 77, 95, 132, 190, 199military sales, 132offshore facilities, U.S., 137, 193, 220, 514production in the United States and Japan, 139profits, 137programmable read-only memory (PROM), 96quality dumping, 230random-access memory (RAM), 96, 99, 130. 196,

247, 248, 2499K dynamic MOS RAM, 524-531

read-only memory (ROM). 96, 176, 190, 323read-write memory (RWM), 96rpcinl rr1.1 nnri rlivirnlevrArts...1 07 00 nn Ines

Index 545

U.S. firms. international operations, 134. 13:3U.S. sales, 26. 130, 131, 198

in RAMS, 25. 57, 176, 187, 529-530very large-scale integration (VLSI), 73. 76. 79. 81.

82. 92. 194. 196Semiconductor Industry Association. 309. 313Semiconductor Research Cooperative, 28. 195Siemens. 142, 143. 205. 211. 269Silicon Valley. Calif.. 31. 55. 311. 313, 322, 323Singapore. 126. 127, 135, 193. 205Small Business Investment Corp. (SBIC), 269, 520Sony. 12. 115, 122, 186. 231. 285South Korea

color TV exports to the United States. 24. 112-118competitiveness, 58, 61. 166consumer electronics, 12. 23, 126, 128, 129Development Bank. 384dumping. 42Economic Planning Board, 384electronics production. 129, 384Hyundai Group. 129industrial policies, 383Institute of Electronics Technology, 384Ministry of Commerce and Industry. 384Ministry of Communications, 385semiconductor manufacture. 143

Soviet Union. 108Spain. 122. 211Sri Lanka. 127Standard Industrial Classification (SILJ categories,

350, 365Stanford UniversitN, 308, 310, 313Sylvania. 181, 183

Taiwan. 12, 23, 24. 42. 57. 58, 61. 63. 117, 126, 135,151, 166, 513

Electronics Research and Service Organization(ERSO), 386

industrial policies, 385-386Industry Technology Research Lstitute (ITRI), 386

Tandy Corp., 148, 179, 186, 205,.;tariffs (see trade)Tata Elxsi, 388Taylor, Frederick W.. 221, 326Texas Instruments (TI), 72, 111, 131, 136, 138, 140,

144, 179, 186, 189, 191, 193, 194, 197, 199. 205,240, 277, 322, 525, 526, 527

Thomson-Brandt, 126Three Mile Island, 103Tokyo Round Multilateral Trade Negotiations, 45,

169, 430, 432, 436, 437, 438, 443, 451Toshiba, 158, 185, 210Toyota, 317, 436trade

policies and their effects. 429-460

computers, 457consumer electronics, 451semiconductors, 452

quantitative restrictions and the escape cause.446-450

escape clause proceedings in color TV, 447OMAs for color TV imports, -146quotas and other nontariff restrictions, 445the escape clause, 449-450

subsidies and countervailing duties, 442-446indirect taxes, 444unfair practices, 445

tariffs. 43, 45, 430-438agreements from Tokyo Round. 436-437changes in Tokyo Round N1TN, 432-433effects. 431, 433, 434offshore manufacturing, 434-435

Trilogy Systems. 205, 271TRW, 211

Underwood, 147Unidata, 205Unimation, 240, 242Union of Japanese Scientists and Engineers (JUSE),

224United Auto Workers, 331United Kingdom

Advisory Council for Applied Research andDevelopment, 402

Central Computer Agency, 402computer development, 81consumer electronics, 124Department of Trade and Industry, 401government subsidies, 268industrial policies, 400-405Industrial Reorganization Corp., 401International Computers Ltd. (ICI.), 55, 200, 205,

211, 400, 403Microelectronic Industry Support Programme, 403Microprocessor Applications Project (MAP), 403Ministry of Technology, 4u1National Economic Development Council (NEDC),

401National Enterprise Board (NEB), 381, 402, 404,

409Science Research Council, 402Thatcher administration, 403

United Nations Council on Trade and Development,388

United Tech- 'logies, 133Univac, 82University of California, Berkeley, 313Unix, 29U.S. Bureau of Census, 81

1,f111.5. FliirPail Lahnr StntiOiru cl l',41 371

546 Internatio jai Competitiveness in Electronics

Ada-computer langUage, 107l'itrastnall Electronics Research. 391Very iligh-Speed Integrated Circuit (VHSIC)

program. 15. 79, 80, 209, 381, 391, 393, 4651..5. Department of Education, 303, 3131.7.S. Department of Energy, 314U.S. Department of Justice, 55, 390, 465

Departm.mt of Labor, 320, 356U.S. Department of State. 431U.S. Department of the Treasury- 38, 439, 443. 444,

445Customs Service, 440. 441

U.S. Federal Communications Commission, 464, 520U.S. Eedera, Trade Commission. 187, 390, 465U.S. General Services Administration, 436U.S. International Track, Commission (ITC). 133. 361,

439, 440, 443, 445, 448, 449, 45111.S. Small Business Administration, 38117.5. Steel, 110U.S. Supreme Court, 441, 444. 448U.S. Tariff Schedules, 117, 118, 136, 434U.S. Trade I.Znpresentative, 431

value-added tax (VAT), 444, 445Visicorp, 110

Wang, 204, 205, 387Wang Institute, 309Warwick Electronics, 115. 183, 223Western Electric, 14, 33, 59, 130, 132, 144, 196, 227,

277, :366

West GermanyAct for the: Promotion of Stability and Economic

Growth, 405AEG-Telefunken, 123, 407, 408, 411Association of German Engineers, 412Commerzbank, 293Council of Economic Adviser-, 406Deutsche Bank, 293Deutsche Bundesbank, 406Dresdner Bank, 293Fraunhofer Gesellschaft (Association of Institutes of

Applied Research, FhG), 28, 381, 410gross national product, 408Grundig AG, 411industrial policies, 405Institute for Solid State Technology, 410Japanese TV manufacturing in, 122Labor Market Office, 349Ministry of Economics, 406, 410Ministry of Finance, 406Ministry of Labour and Social Affairs, 331, 407Ministry of Research and Technology, 331, 406Nixdorf, 148, 152, 201, 408, 411secondary schooling, 304Technical University of Munich, 410

Xerox Corp., 125, 196

Zenith. 12, 13, 55, 68. 107, 111, 113, 114, 117, 128,172, 181, 183, 186. 444, 448, 451

Office of Technology Assessment

The Office of Technology Assessti ent (OTA) was created in 1972 as ananalytical arm of Congress. OTA's basic function is to help legislative policy-makers anticipate and plan for t:ie consequences of technological changes andto examine the many ways, exp,:cted and unexpected, in which technology,affects people's lives. The assessment of technology calls for exploration ofthe physical, biological, economic, social, and political impacts that can resultfrom applications of scientific knowledge. OTA provides Congress with in-dependent and timely information about the potential effects--both benefi-cial and harmfulof technological applications.

Requests for studies are made by chairmen of standing committees of theHouse of RepresentatiVes or Senate; by the Technology Assessment Board,the governing body of OTA; or by the Director of OTA in consultation withthe Board.

The Technology Assessment Board is composed of six members of theHouse, six members of the Senate, and the OTA Director, who is a non-voting member.

OTA has studies under way in nine program areas: energy and materials;industry, technology, and employment; international security and commerce;biological applications; food and renewable resources; health; cOmtnunica-non and information technologies; oceans and- erivii-Onnietit-r and:science;transportation, and innovation.

Competition; *Electronics; Employment Aigher E - ERIC

Documents

Transcript of Competition; *Electronics; Employment Aigher E - ERIC