Remedies to Exclusionary Innovation in the Hight-Tech Sector

Remedies to Exclusionary Innovation in theHigh-Tech Sector: Is there a Lesson from the

Microsoft Saga?

Maria LillaÁ MONTAGNANI*

In the high-tech sector innovative behaviour may have exclusionary effects: In both the Microsoft III andIV cases exclusionary practices centred on innovation were deemed generating both anti- and pro-competitive effectsÐwith the latter overcoming the formerÐto the extent of requiring an antitrust lawintervention. Such intervention presents two problems though. Firstly, the question arises as to wheninnovation becomes detrimental to competition thereby requiring antitrust remedies to be applied. Secondly,it is to assess how antitrust remedies should be designed in order not to stifle the pro-competitive effects thatinnovative behaviour, although exclusionary, still generate.

This article addresses both issues above mentioned. The potentially exclusive nature of innovation is thestarting point to explore the interface between remedies on the one hand, and (incentives to) innovation onthe other, in order to offer considerations on a sound remedies policy.

I. INTRODUCTION AND TERMINOLOGY

Innovation has historically been deemed a competition driver and, therefore,

rewarded through the granting of intellectual property rights (IPRs).1 However, in

high-tech markets innovation may be twofold. On the one hand, it promotes

competition and deserves reward, such as IPRs; on the other hand, whenever the actor

is monopolist or enjoys a dominant position, innovation may generate anti-competitive

effects to the extent of turning these into an exclusionary practice. The reason why this

occurs can be found in the specific features of the high-tech market,2 such as lock-in,

network, winner-takes-all, and the like, which, coupled with low marginal costs,

amplify anti-competitive effects.3 In addition, in such markets actors' positions are

strengthened by IPRs not only rewarding rightholders for their previous ``innovative

efforts'', but also protecting their freedom to further innovate. Although changing

* LL.M., Ph.D., Assistant Professor of Commercial Law at Bocconi University, Institute of ComparativeLaw, Milan. This article is based on a presentation addressed to the 1st Annual Conference of the EPIPAssociation, Policy, Law and Economics of Intellectual Property, 7/8 September 2006, in Munich, on ExclusionaryInnovation: How to Avoid Innovation to be Stifled. Remedies Identified for the Software Sector, (available at <http://www.antitrustisti.net>). I am grateful to the conference participants for the useful suggestions and hints. I am alsograteful to Laurent Manderieux and Federico Ghezzi with whom I shared the work preliminary to the Conference.Although this work represents a further analysis and development, without the discussions and their comments itwould not have been the same. All the views herein expressed are however strictly personal and can be furtherinvestigated by emailing the author at: [email protected].

1 Frederic M. Scherer, Industrial Market Structure and Economic Performance, Rand McNally College, Chicago, 1970,p. 350. See also John M. Clark, Competition as a Dynamic Process, Brooking Institution, Washington, 1961, pp. 178±270.

2 Ilka Rhanasto, Intellectual Property Rights, External Effects and Antitrust Law, Oxford University Press, Oxford,2003, p. 183.

3 John Temple Lang, European Community Antitrust Law: Innovation Markets and High Technology Industries, inBarry E. Hawk (Ed.), International Antitrust Law and Policy, Bender, New York, 1996, p. 519.

World Competition 30(4): 623±643, 2007.# 2007 Kluwer Law International. Printed in The Netherlands.

design or updating (upgrading) patented or copyrighted products are rightholders'

choicesÐas well as bundling two different productsÐnevertheless, exploitation of

IPRs may heighten barriers to entry.

Innovation becoming a means of exclusion is not a subject new to the law and

economics literature where a line between ``beneficial'' and ``detrimental'' innovation has

been tentatively struck by those scholars theorizing a ``predatory innovation doctrine''.4

However, herein the aim is not to further develop that doctrine, rather, to use as a starting

point the assumptions that innovative behaviours may generate both pro- and anti-

competitive effects, and that in high-tech markets the latter can be amplified by specific

market features and the IPR presence. Even though the predatory innovation doctrine is

used here as the starting point of the analysis, in this article a slightly different terminology

will be adopted, and a further distinction will be implemented. Amongst all the

behaviours falling into the category of predatory innovation, those merely aiming at

driving and keeping competitors out of the market are defined ``predatory innovation

practices'': since they falsely innovate, their unlawfulness is beyond doubt and they will

not be analysed. The focus will be on those highly controversial practices centred on

innovation which, regardless of the actor's purpose,5 generate, at the same time,

detrimental and beneficial effects on competition. Given the potentially exclusive nature

of such innovative practices, they are defined ``exclusive innovation conduct'' and may

need an antitrust intervention aiming at eliminating the anti-competitive effects, such as

occurred in the cases of Microsoft's integration of Internet Explorer (IE) and Windows

Media Player (WMP) in Microsoft III and IV respectively.

However, competition law intervention in high-tech markets faces two relevant

issues. Firstly, the standard to distinguish detrimental from beneficial innovation needs

to be formulated. Secondly, the question arises as to how to tailor general antitrust

remedies to that sector's specific features without restraining incentives to innovate.

As to the first question, the principles emerging from the Microsoft III and IV

decisions are likely to be that as long as IPRs work as incentives to innovate, a system of

rivalry is to be preserved; however, when their exploitation by rightholders becomes a

means to limit competition in the market, then IPRs lose their incentive to innovate

function and an antitrust intervention is needed in order to maintain markets open and

the innovation process ongoing.

As to the concern of fine tuning general antitrust remedies, it does not appear to be

thoroughly addressed yet. Indeed, once innovation becomes detrimental to competi-

tion, the various categories of remedies ought to be carefully considered in order to

compare their advantages and drawbacks.

4 Janusz A. Ordover and Robert D. Willig, An Economic Definition of Predation: Pricing and Product Innovation,(1981) 91 Yale L. J. 8. Contra J. Gregory Sidak, Debunking Predatory Innovation, (1983) 83 Colum. L. Rev. 1121.

5 Although in some cases courts have considered the innovator's intent to adopt exclusionary practicesrelevant to the assessment of predatory practices centred on innovationÐas in Bard v. M3System, 157 F.3d 1340(Fed. Circ. 1998, analysed by Herbert Hovenkamp, Federal Antitrust Policy, The Law of Competition and its Practice,West End ed., St. Paul Minn., 1999, pp. 327±332; and Federico Ghezzi and Maria LillaÁ Montagnani, Software edinnovazione predatoria, (2004) 23 AIDA 436-441)Ðin this context the actor's purpose is not taken into account.

624 WORLD COMPETITION

In this article I will first illustrate the twofold nature of innovation using Microsoft's

practices, and the problems that they generate (sections II and III). Then, in section IV, I

will discuss the legal standard to distinguish between beneficial and exclusionary

innovation. Finally, in section V the central issue of tailoring efficient remedies in order to

limit the anti-competitive effects of exclusionary innovation in the software market (and

more generally in the high-tech sector) will be addressed. The remedies applied in Microsoft

III and IV will be considered as a starting point to formulate considerations on a sound

remedies policy in the high-tech sector as mentioned in the conclusion (section VI).

II. THE TWOFOLD NATURE OF INNOVATION IN THE SOFTWARE MARKET

The exclusionary potentiality (or the potential exclusiveness) of innovative

behaviours in high-tech markets emerges in the recent cases dealing with Microsoft's

middleware integration practices and its refusal to disclose information.6 Both the well-

known American and European Microsoft decisions address the issues of software

integrationÐrespectively IE and WMPÐinto Microsoft operating system (OS), and of

protocols and interoperability information foreclosure.

Before entering into the merits of these decisions it is worth pointing out that the

software market is a classic example of a high-tech market presenting specific features.

In this sector one product or standard tends towards dominance because of high direct

and indirect network effects.7 The utility that a user derives from the consumption of a

good increases with the number of other users consuming that good. Thus, once one

product or standard achieves wide acceptance, it becomes entrenched.8 On the other

hand, software programs are protected by copyright and this protection has been

progressively extended from source code to other elements,9 not least communication

interfaces, information protocols, and the like. Besides, patents are granted for software

programs showing technical effects, which further expands the protection granted to

these products.10 Middleware (or more broadly software) integration can then

6 United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001) (Microsoft III); and Commission Decision of24.03.2004, relating to a proceeding under Article 82 of the EC Treaty (Case COMP/C-3/37.792, Microsoft),C(2004)9 final (Brussels 21.4.200) (unofficial publication) (Microsoft IV). The Commission's decision has beenrecently upheld by the Court of First Instance (Case T-201/04) but for the monitoring trustee to be appointed bythe Commission from a list of persons drawn up by Microsoft. For this reasonÐand since this article has beenfinalised before the adoption of the CFI's decisionÐreference will herein be made to the Commission's decision.

7 This mechanism has been defined competition ``for the field'', instead of ``in the field'', by Harold Demsetz,Why regulate utilities?, (1968) 11 J.L. & Econ. 55, 57 n. 7.

8 Joseph. A. Schumpeter, Capitalism, Socialism and Democracy, Allen & Unwin, London, 1943, pp. 81±90. TheAuthor argues that this entrenchment is temporary because innovation may alter the field altogether. At least thelatter is deemed the mechanism to keep the software market (and, generally speaking, the high-tech markets)competitive. Recent cases, though, such as the European Microsoft case, demonstrate that network effects and IPRstend to strengthen the innovator position. On the assessment of market power in high-tech markets see HerbertHovenkamp, Post-Chicago Antitrust: A Review and Critique, (2001) Colum. Bus. L. Rev. 332.

9 For a complete overview of software protection under US copyright law see Mark A. Lemley et al., Softwareand Internet Law, Aspen Law and Business, 2nd Edn, New York, 2003. For a comparative approach, see AdrianSterling, World Copyright Law, Sweet & Maxwell 2nd Edn, London, 2003.

10 On software patentability see, amongst many, Micheal Guntersdorfer, Software Patent Law: United States andEurope Compared, (2003) Duke L. & Tech Rev 6.

REMEDIES TO EXCLUSIONARY INNOVATION 625

exemplify the issue analysed: innovative behavioursÐadopted by IPR holdersÐ

generating both pro- and anti-competitive effects. Moreover, when adopted by a firm

leading an upstream market (such as the operating system market), software integration

may be capable of hampering competition in an aftermarket (such as the server or

browser markets). Similarly, a firm vertically integrated that refuses to supply

information necessary to develop complementary products (such as media players or

work group servers) does raise barriers to entry in the secondary markets. However,

since software market operators are rewarded for their innovation through the granting

of IPRs that, in turn, are supposed to provide an incentive to further innovate, they can

allege that those behaviours fall within the boundaries of their exclusive rights.

Such a complex scenario makes it difficult to strike a balance between pro- and anti-

competitive effects generated by exclusive behaviours centred on innovation, and to

ascertain when innovation turns into ``exclusion'', or to decide how and when intervention

should take place in order to eliminate the anti-competitive effects without stifling the pro-

competitive. The US Court of Appeal in the Microsoft III and both the European

Commission and the CFI in the Microsoft IV decisions appear, however, to provide some

hints. For these reasons the following paragraphs will deal with each decision in turn.

A. MICROSOFT III AND THE BUNDLING OF INTERNET EXPLORER

In Microsoft III, amongst the several challenged behaviours such as withholding

crucial technical information, predatory prices, contractual restrictions on original

equipment manufacturers (OEMs) in order to affect distribution channels and the like,

the software integration at issue was IE integration in the Windows 98 OS.11 To end-

users this did not directly constitute a burden since it did not affect the Windows 98

purchase price, rather, as it emerges in the decision, it indirectly affected the variety of

browsers available on the market. The integration, thus, did matter to Microsoft

competitors. In particular, Netscape and Sun Microsystems argued the integration to be

an exclusionary practice since (i) it prevented their productsÐnamely, Netscape web

browser and Java class librariesÐfrom competing with the Microsoft products in the

OS and browser markets; (ii) it prevented them from entering the OS market; and (iii)

it drove them out of the browser market.

While all the complaints were accepted by Jackson J.Ðwho even required

Microsoft to submit a proposed plan of divestiture in order to split the company into an

OS business and an applicative software business12Ðthe Court of Appeal upheld the

sole monopolization offence. Microsoft was deemed to possess monopoly power in the

OS relevant market and to have adopted an anti-competitive behaviour in order to

maintain its position. The existing monopoly power was derived not only from

11 Specifically, challenges were of monopolization; attempt of monopolization; and tying.12 United States v. Microsoft Corp., 97 F.Supp.2d 59 (D.D.C 2000).


Microsoft's market share, but also from the OS market structure, such as the presence

of network effects as barriers to entry.

Although the Court of Appeal was in principle sceptical about an integration

having anti-competitive effectsÐsince integrations were still considered mainly pro-

competitiveÐin this case even the second instance court showed concern for the way

the integration could affect the emergence of products alternative to Microsoft OS (i.e.

middleware). Since IE was bundled into the OS code, it was deemed to prevent OEMs

from pre-installing other browser programs, thereby reducing rivals' usage share and

developers' interest in the rivals' application programming interfaces (APIs) as an

alternative to those exposed by the Microsoft OS. In other words, both Courts appear

to be aware of the integration refraining middleware potential competition to OSs to

the extent of constituting an exclusive practice.13

As to the other claims, they were remanded to the lower court which never

pronounced on them because the parties signed a consent decree putting an end to the

proceeding,14 and imposing many obligations on Microsoft.15 Amongst them it is

worth mentioning: (i) the disclosure of the new APIs in use between OS and browser,

and (ii) the compulsory licences of Microsoft's IPRs to enable licensees (i.e. OEMs) to

promote non-Microsoft middleware. As to the former, despite the fact that the APIs

may be considered protected under copyright and trade secret laws, their disclosure was

deemed necessary so as to enable software developers to create new browser versions

competitive with the new IE version. As to the latter, past behaviour of Microsoft had

showed how IPRs could be a means to prevent OEMs from promoting non-Microsoft

products. Therefore, compulsory licensing was considered a viable solution to this.

B. MICROSOFT IV: THE TIE OF WINDOWS MEDIA PLAYER AND THE REFUSAL TO

SUPPLY INFORMATION

The starting point for Microsoft IV is the above mentioned consent decreeÐas

implemented in Kollar-Kotelly J.'s final judgmentÐand its alleged violation. Although

the Microsoft III and IV decisions are often compared, and the Microsoft IV decision is

13 In order to understand this point some technological background is needed. Specifically, OSs and softwareprograms interact through Application Program Interfaces (APIs) which also enable software developers to writeprograms compatible with an OS. However, OSs are not the only software programs exposing APIs. Other non-OS programs do the same, among them: Netscape and Sun Microsystem products. Programs like these are called``middleware'' since they rely on an OS and its APIs, but, at the same time, they expose their APIs to softwaredevelopers. While the Microsoft III controversy was taking place, no middleware exposed enough APIs to offer a fullrange of applications. Users still needed to rely on an OS. However, a middleware such as Netscape would havebeen capableÐonce program developers had written enough middleware applicationsÐto satisfy all user needs.Had this happened, users might have chosen less expensive middleware compatible applications rather than OSdirect compatible applications. For this reason, Jackson J. stated that: ``[t]he growth of middleware-basedapplications could lower the costs to users of choosing a non-Intel-compatible PC operating system like the MACOS'' [United States v. Microsoft Corp., 83 F.Supp.2d 9, 18 (D.D.C. 1999)].

14 United States v. Microsoft Corp., 231 F.Supp.2d 144 (D.D.C. 2002).15 A thorough analysis of Microsoft III remedies has been carried by David S. Evans, Albert L. Nichols and

Richard Schmalensee, U.S. v. Microsoft: Did Consumers Win?, (October 2005) NBER Working Paper SeriesWorking Paper 11727, available at <http://www.nber.org/papers/w11727>.


deemed stricter than Microsoft III,16 they do not appear to involve identical practices,

rather, to address different aspects of a similar behaviour.

In the European case, Microsoft was challenged for abuse of dominant position

under Article 82 of the EC Treaty with regard to both the OS and work group server

markets and the WMP integration in its OS, which was deemed to constitute a much

stronger competitive advantage than the previous integrations due to the WMP new

functionalities.

In both cases the challenged behaviours were defined softwareÐor better

middlewareÐintegration. However, these practices need to be divided (or at least

seem to have been divided in the Microsoft III decision)17 into technological integration

(or technological tying) and contractual tying (the latter appearing a means to

strengthen the former in both the American and European cases). Technological tying

was the focus of the assessment carried out in Microsoft III, and, in this respect, the

behaviour was evaluated under the monopolization offence (the contractual tying not

having been assessed but remanded). In Microsoft IV, instead, since the technological

tying complied with the consent decree as to the disclosure obligations (i.e. the new

APIs in use between WOS and the new WMP were disclosed), the assessment dealt

more broadly with other non-technological components of the tying practiceÐi.e.

``contractual tying'' (those same practices remanded to the lower court in Microsoft III

and never assessed because of the settlement).

What strikes in the Commission reasoning is the consideration given to potential

competition,18 whose stifling may make consumer choices impossible due to the

absence of alternative products. For this reason, Article 82 EC of the Treaty was

deemed applicable even though Microsoft complied with the consent decree, and

affected consumer choices only indirectlyÐby affecting access to the market and

16 Rudolph Peritz, Re-Thinking U.S. v. Microsoft in Light of the E.C. Case, (22 March 2004), NYLS LegalStudies Research Paper No. 04/05-4 <http://ssrn.com/abstract=571803>.

17 The distinction between technological and contractual tying is stressed by Herbert Hovenkamp, as note 5supra, at 303, arguing that ``the most obvious difference between the §2 tying (. . .) and the traditional §1 or ClaytonAct §3 offences consists in the lack of any agreement requirement in the former. Monopolization is a unilateralpractice. So when a dominant firm unilaterally imposes tying (. . .) under circumstances where a qualifyingagreement cannot be proven, the practice may still constitute an antitrust violation''. Specifically, in Microsoft III theCourt of Appeal identified four behaviours: two of themÐnamely preventing OEMs to uninstall or remove IEfrom WOS desktop; and designing WOS so as to withhold from consumers the ability to remove IE by use of theAdd/Remove Programs utility in WindowsÐwere deemed technological tying and assessed through themonopolization offence test; the remaining twoÐnamely: Microsoft's requiring WOS licensees to license IE as abundle at a single price; and Microsoft's designing WOS so as to override the users' choice of default web browserin certain circumstancesÐwere deemed tying, and needed for a rule of reason analysis by the remanded court. Onthe difference between technological and contractual tie-ins, see David S. Evans, A. Jorge Padilla and MichelePolo, Tying in Platform Software: Reasons for a Rule-of-Reason Standard in European Competition Law, 25 W.Comp. 4,509, (2002).

18 Ian S. Forrester, Article 82: Remedies in Search of Theories?, in Barry E. Hawk (Ed.), International Antitrust Law& Policy, Bender, New York, 2005, p. 167 (stressing that in the last decade the European Commission's applicationof Article 82 EC Treaty has looked forward ``at the desired conduct rather then looking backward and ensuring thediscontinuation of the abuse'').


behaviours of stakeholders, such as OEMs, content providers, and software

developers.19

In this context, in order to avoid barriers to entry, the remedy adopted was the

obligation to offer to OEMs two WOS versions: one with WMP unbundled and one

with WMP bundled. The effectiveness of this remedy has been criticized in many aspects.

The Commission applied similar reasoning in relation to the work group server

market. In this market more than in the media player one, competition was deemed

harmed by Microsoft's refusal to supply the required protocols and interoperability

information since they were deemed necessary to enable competitors to develop work

group servers capable of running on WOS. Maintaining the work group server market

open required that this information would be disclosed and, when necessary, licensed.

III. CONCERNS AT THE INTERFACE BETWEEN INNOVATION, EXCLUSION, AND

COMPETITION

The Microsoft III and IV decisions show that exclusionary practices centred on

innovation are likely to call for an antitrust law intervention in order to limit the anti-

competitive effects that they can generate in network markets. Such an intervention is

not devoid of problems though.

Firstly, the question arises as to when antitrust law can intervene to address behaviour

that, being centred on innovation, tends to be deemed lawful. Even more when the

behaviour is adopted by IPR owners in the exercise of their exclusive rights.20

19 The reasoning underlying the Commission's decision on tying in Microsoft IV can be clearly traced to theDG Competition discussion paper on the application of Article 82 of the Treaty to exclusionary abuses, Brussels,December 2006 (<http://ec.europa.eu/comm/competition/antitrust/others/discpaper2005.pdf>) at paras 180±181:``A company that is dominant in the tying market can through tying or bundling foreclose the tied market and canindirectly also foreclose the tying market (horizontal foreclosure). . . . The foreclosure of the tied market may allowthe dominant company to achieve larger profits in the tied market, for example through catching more of thecustomers in that market. Moreover, tying may allow the dominant company to protect or strengthen its dominantposition in the tying market''. The same argument is developed with regard to refusal to supply: ``The owner of theinput may refuse to supply in order to achieve a larger share of the profits in the downstream market. Moreover, therefusal to supply may allow the input owner to protect its position in the upstream market'' (para. 212). Similarly,the reasoning is deemed applicable to refusal to licence an IPR protected technology which is indispensable as abasis for follow-on innovation by competitors. In fact, such conduct ``may be abusive even if the licence is notsought to directly incorporate the technology in clearly identifiable new goods and services. The refusal of licensingan IPR protected technology should not impair consumers' ability to benefit from innovation brought about by thedominant undertaking's competitors'' (para. 240). In such case, therefore, the additional condition required forrefusal to license IPRs (the prevention of new goods or services described in paras 237±239) is unlikely to be stillnecessary.

20 When anticompetitive behaviours are adopted by IPR owners in ``traditional markets'' (i.e. markets devoidof the supra mentioned effects), the standard adopted to draw a line between legitimate exercise of IPRs andanticompetitive practices is to be found in the ``subject matter of rights'' which has superseded the existence/exercise dichotomy criterion. This has been recently confirmed in AstraZeneca with regard to SPCs unlawfullyextending the patent temporal protection, where the Commission has affirmed that: `` . . . AZ conduct can hardlybe described as belonging to the subject matter of the rights in question. . . . [T]he making of misleadingrepresentation is not included in the bundling of rights forming part of the subject matter of a SPC'' (Case COMP/A. 37.507/F3, AstraZeneca, § 741). Despite different outcomes, similar reasoning has been adopted in Volvo AB v.Veng, [1988] ECR 6211; (Case 238/87) Coditel v. Cine Vog Films, [1980] ECR 881; (Case 62/79)Centrafarm v.Sterling, [1974] ECR 1147 (Case 15/74). In the latter, though, the existence/exercise dichotomy is still mentioned,while it appears completely surpassed in AstraZeneca. In this decision, in fact, the acquisition of the extended patentright itself constitutes an abuse since it was the result of misleading representations.


Were the first question answered, a second and even more problematic question

would still need to be addressed. Once a line between beneficial and exclusionary

innovation has been drawn, how should antitrust remedies be designed in order not to

stifle the pro-competitive effects that innovative behaviour, although exclusionary, still

generate? Or better: How to avoid lowering the incentive to innovate further?

Remedies centred on IPR restraints are highly controversial as they are deemed to

reduce or eliminate incentives to innovate. Whenever adopted, their architecture must

thus be carefully thought out and properly designed in order to remove the anti-

competitive but maintain the pro-competitive effects.

The following sections will deal with each issue in turn.

IV. STRIKING A LINE BETWEEN ``GOOD'' AND ``DETRIMENTAL'' INNOVATION:

WHEN DO ANTI-COMPETITIVE OUTWEIGH PRO-COMPETITIVE EFFECTS?

The specific issue of assessing exclusionary conduct centred on innovation ought

to be framed within the context of the uneasy relationship between competition and

intellectual property laws. In this context, a legal standard to distinguish ``good'' from

``bad'' innovation needs to be formulated in the light of the goal pursued by intellectual

property and competition provisions.

Although competition law aims at the greatest enhancement of social wealth by

fostering innovation efficiency, followed by product efficiency and, lastly, allocative

efficiency,21 priority is to be given to advance technological and product processes as

they constitute the engine of societal development. This can happen only by preserving

competitive processes in the long run.22 Therefore, exclusionary practices not

apparently affecting consumer wealth may still provide a disincentive for production

and innovation advances through losses of competing sources of innovation, thereby

threatening or displaying the competition process. Similarly, IPRs are granted in order

to maximize the social welfare through the access versus incentive trade-off: ``charging

a price for a public good reduces access to it . . . but increases the incentive to create in

the first place, which is a possible offsetting of social benefit''.23 Intellectual property law

confers exclusive rights as a reward to innovation (they enable innovators to recoup

their investment) and incentives to further innovate. However, these rights are also

tailored in breadth, scope, and term to allow certain uses and a certain degree of access

21 Joseph F. Brodley, The Economic goals of antitrust: efficiency, consumer welfare, and technological progress, (1987) 62NYUL Rev. 1020-1021, 1042, arguing that ``In some instances exclusionary conduct may not cause an immediateinjury to consumers, but this should not prevent antitrust intervention against deliberate cost raising or otherexclusionary conduct. By lowering the return to production efficiency and inducing wasteful investment inpredatory strategies, exclusionary conduct reduces aggregate social welfare and, in the long run, consumer welfareas well''.

22 John J. Flynn, Antitrust and the Suppression of Technology in the United States and Europe: Is There a Remedy?,(1988) 66 Antitrust LJ 487, 495.

23 William M. Landes and Richard A. Posner, The Economic Structure of Intellectual Property Law, The BelknapPress of Harvard University Press, Cambridge Mass., 2003, pp. 20±21, arguing that the trade-off between accessand incentive is very important, though there is much more in the economic analysis of intellectual property law,such as the economic continuity between physical and intellectual property.


to the public so as to enable information and culture to be spread. Hence, as long as

intellectual property law maintains a balance between public benefit, on the one hand,

and scope, term and breadth of the rights conferred, on the other hand, it is consistent

with competition law.

In the light of their mutual goal (i.e. fostering and protecting the innovation

process) in the high-tech markets, intellectual property law and competition law are

likely to complement each other. The former grants exclusive rights in order to

increase innovation so as to enlarge the number of actors in the market; while the latter

seeks to maintain or encourage pluralism in the market in order to increase innovation

and output. That is to say that the same goal can be sought through different ways,

intellectual property being an ex ante tool and competition law being an ex post tool.

In order to enable competition and intellectual property law to achieve their goal,

the appropriate legal standard for assessing exclusive innovation in the software (and

high-tech) market encompasses three steps which mirror the US Court of Appeal and

European Commission reasoning. Firstly, the market power of the innovator is to be

verified under the usual criteria adopted for market power definition and assessment in

the high-tech sector. As a result, the innovator ought to have a dominant or

monopolistic position to continue the assessment. Secondly, behaviour falling under

the category of exclusionary conduct and centred on innovation must take place (that is

to say a conduct that is likely to exclude competitors and, by doing this, hamper

competition). Finally, the balance between the effects of the conduct is to be struck.

Whenever the anti-competitive outweigh the pro-competitive effects, the innovative

practice can be deemed exclusionary.

The core point is then ascertaining when the anti-competitive outweigh the pro-

competitive effects. From the Microsoft III and IV decisions it seems that this occurs

when the innovator behaviour blocks or slows down the innovation pace which both

competition and intellectual property law aim at maintaining in force.24 This is likely to

occur whenever the innovator is entrenched on what can be called the ``technological

frontier'', which is the last advancement in technology.25 Having reached this position

may enable the incumbent to block subsequent innovation: not only the subsequent

innovation of competitorsÐunable to overtake that edgeÐbut also the subsequent

innovation of the incumbent himselfÐwho does not have any further incentive to

overtake it. This determines a lowering of common welfare which may not be

immediately perceivedÐsince consumers are not harmed in the short runÐbut runs

the risk to block or severely slow down the innovation process in the long run because

24 This idea was initially formulatedÐin a time in which innovation was per se lawfulÐby James W. Brock,Structural Monopoly, Technological Performance, and Predatory Innovation: Relevant Standards under Section 2 of the ShermanAct, (1983) 21 American Business LJ 305-306. The Author also claims for a close investigation of, and appreciationfor, the nature, degree and significance of structural monopoly power. This analysis should precede and conditionthe evaluation of performances in attempting to distinguish lawful and unlawful monopoly under section 2 of theSherman Act.

25 For an analysis of this concept see Maria LillaÁ Montagnani, Predatory and Exclusionary Innovation: WhichLegal Standard for Software Integration in the Context of the Competition versus Intellectual Property Rights Clash?, (2006) 37IIC 333.


the incumbent does not fear any pressure to innovate from competitors, who are

constantly kept a step behind.26

This mechanism is likely to take place in high-tech markets in general, and in the

software market in particular, since network effects and the like make these markets

tend towards the entrenchment of the innovator on the technological frontier.27

V. DESIGNING ANTITRUST REMEDIES FOR HIGH-TECH MARKETS

Once it is established that whenever IPRs do not implement an efficient system of

rivalry anymore (due to the innovator's entrenchment on the technological frontier),

they can be questioned from a competition law point of view in order to enable

competition on the merits to take place and not to block or slow down the innovation

process.28 In this context the critical issue of tailoring remedies for the high-tech

market remains to be tackled.

As a matter of fact, it is implicit that optimal remedies should aim at stifling the

anti-competitive effects of the innovative conduct without eliminating the pro-

competitive effects that these behaviours still generate. Exclusive behaviours centred on

innovation cannot be fully prevented since they still offer a certain degree of innovation

thereby representing a step (even minimum) forward. The problem is that after a

minimum step forward no other further steps may take place due to the lack of pressure

on the innovator entrenched on the technological frontier. In such a situation an

intervention should aim at restoring the mechanism under which the innovation pace

maintains its course.

The following survey of the antitrust remedies available (I V.A) is useful to

establish the framework in which high-tech-market-oriented remedies can be designed

(I V.B). The remedies thus far adopted in the software market (mainly targeted on

IPRs) will then be addressed (I V.C).

26 This has been deemed the test carried out in Microsoft IV by FrancËois LeÂveÃque, Innovation, Leveraging andEssential Facilities: Interoperability Licensing in the EU Microsoft Case, in FrancËois LeÂveÃque and Howard Shelansky (Eds),Antitrust, Patents and Copyright, Edward Elgar, Cheltenham UKÐNorthampton MA USA, 2005, pp. 108±109,with regard to the refusal of supplying interoperability information. The Author asserts that in Microsoft IV theCommission replaced the ``new product test'', elaborated under the ``essential facilities doctrine'', with the``incentives balance test'' under which a refusal of licensing that reduces the incentives to innovate can constituteabuse when limitation of technical development and prejudice of consumers can be proved. However, the Authoralso pinpoints that even though this test is economically more appropriate, the Commission's argument todemonstrate a decrease of industry incentives (and specifically of Microsoft's incentives) was not robust enough.

27 The same result seems reached by Andreas Heinemann, Compulsory Licences and Product Integration inEuropean Competition Law Assessment of the European Commission's Microsoft Decision, (2005) 36 IIC 82.

28 Similarly Josef Drexl, Intellectual Property Rights as Constituent Elements of the Market Order, (2 September2006) Addressed to the ATRIP Congress in Parma, arguing that IPRs are to be considered in the light of theirfunction which is promoting innovation by substitution more than by imitation. Since exclusive rights should aimat avoiding imitation and fostering substitution, whenever IPRs do not achieve this goal of promoting innovationby substitution but limit themselves to avoid imitation, then something is not working anymore. The concept isfurther developed, with regard to conduct constituting abuse of dominant position in Id., Comments of the MaxPlanck Institute for Intellectual Property, Competition and Tax Law (Munich) on the DG Competition discussion paper ofDecember 2005 on the application of Article 82 of the EC treaty to exclusionary abuses, (31 March 2006) available at <http://www.ip.mpg.de/shared/data/pdf/comment1.pdf>.


A. ANTITRUST REMEDIES IN GENERAL

Very broadly speaking, on both sides of the Atlantic Ocean antitrust remedies can

be classified into two main categories, regardless of the public or private nature of the

body claiming the infringement via an administrative procedure or a legal action.29

Firstly, there are the pecuniary sanctions. These can be either damages in order to

recoup losses caused by the breach of competition law,30 or fines in order to deter

infringers from further infringements.

Secondly, there are injunctions via which infringements are brought to an end

(cease-and-desist orders) or post-remedial actions adopted in order to remove or repair

the infringement effects. Besides injunctions, prohibitive in nature, either mandatory

injunctions, such as obligations to supply (so called ``behavioural remedies''), or orders

to modify the infringer's company structure (so called ``structural remedies'') can be

adopted.31

As to the former (behavioural remedies), they are supposed to affect the behaviours

of infringers without directly altering the company structure and the relevant markets,

or the distribution of assets necessary to compete amongst rival firms. While it is not

easy, nor worthy, providing a list of behavioural remediesÐsince these depend on the

29 Generally speaking indeed, we are able to distinguish two kinds of enforcement: public and private. Theformer is initiated by public authorities, such as the Federal Trade Commission (FTC), the European Commission,and the national competition authorities (NCAs). The latter is initiated in court by private parties.

30 On level of damages and adoption of treble damages action in the US see Herbert Hovenkamp, TheAntitrust Enterprise. Principle and Execution, Harvard University Press, Cambridge MassachusettsÐLondon, England,2005, pp. 66±68, 305±307; Robert H. Lande, Why Antitrust Damage Levels Should Be Raised, (2004) Loy. ConsumerL. Rev. 329; Robert H. Lande, Are Antitrust `Treble' Damages Really Single Damages?, (1993) 54 Ohio St. L. J. 115;Edward D. Cavanagh, Detrebling Antitrust Damages: An Idea Whose Time has Come?, (1987) 61 Tul. L. Rev. 777;Steven Salop and Lawrence J. White, Treble Damages Reform: Implications of the Georgetown Project, (1986) 55 AntitrustL. J. 73; William Landes, Optimal Sanctions for Antitrust Violations, (1983) 50 U. Chi. Law Rev. 652. In the EU theEuropean Commission has recently adopted Guidelines on the method of setting fines imposed pursuant to Article 23(2)(a)of Regulation No. 1/2003, 2006/C 210/02 1.9.2006. In literature see Wouter P.J. Wils, Optimal Antitrust Fines:Theory and Practice, 29 W.Comp. 2, 183, (2006); Ivo Van Bael and Jean-Francois Bellis, as note 30 supra, at pp.1105±1127; Valentine Korah, EC Competition Law and Practice, Hart Publishing (8th Edn), Oxford and PortlandÐOregon, 2004, pp. 239±244; and Jonathan Sinclair, Damages in Private Antitrust Actions in Europe, (2002) 14 Loy.Consumer L. Rev. 547.

31 The US and EU procedural and remedial systems present several differences. In the United States, theantitrust laws are enforced by two federally-created entities, namely: the Department of Justice (DOJ), and theFTC, as well as private parties inured by antitrust violations. The FTC is entitled to only adopt cease-and-desistorders because, pursuant to Section 19(a)(1) of the Federal Trade Commission Act and similarly to the DOJ, itsmain role consists of bringing actions in court to levy fines for violation of an existing cease and desist order, or forknowing violations. On the other hand civil courts are entitled to adopt any relief, including the ``rescission orreformation of contracts, the refund of money or return of property, the payment of damages, and publicnotification respecting the rule violation or the unfair or deceptive act or practice, as the case may be.'' (Section19(b) of the Federal Trade Commission Act). In the EU, the system currently in force (Regulation 1/2003)provides the European Commission, NCAs, and civil courts with the power to apply Articles 81 and 82 of the ECTreaty. However, the Commission and NCAs have the power of issuing injunctions (cease-and-desist orders,structural and behavioural remedies) and of adopting fines; whereas civil courts can adopt damages, declareanticompetitive agreements void, and prohibit abusive behaviours. On the Modernization of EuropeanCompetition Law see Ivo Van Bael and Jean-Francois Bellis, Competition Law of the European Community, KluwerLaw International, The Hague (4th Edn), 2005, pp. 1167±1194; Wouter P.J. Wils, Principles of European AntitrustEnforcement, Hart Publishing, Oxford and Portland Oregon, 2005; and Renato Mazzini, Concurrent Proceedings inCompetition Law Procedure, Evidence and Remedies, Oxford University Press, Oxford, 2004. On the system adoptedbefore Regulation 1/2003 came into force see C.S. Kerse, E.C. Antitrust Procedure, Sweet & Maxwell, London,1998; and Valentine Korah, EC Competition Law and Practice, Hart Publishing, OxfordÐPortland Oregon, 1997.


case in hand32Ðit is worth noting that in both the US and the EU infringers can

propose obligations that they offer to comply with. Compliance is monitored through

reports that infringers must periodically hand over to authorities which can also impose

further information obligations.

As to the latter (structural remedies), they represent an extrema ratio tool since their

adoption takes place when behavioural remedies would not work and there is certainty

as to their capability to improve market structure.33 Structural remedies are measures

requiring some form of structural change on the part of the party or parties to whom

the measures are directed.34 Within the category of structural remedies different grades

of pervasiveness can be found. Besides accounting separation, authorities can impose

company separationÐunder which property would not changeÐand even the most

pervasive remedy of divestitureÐwhich, on the opposite, requires property separation

and can follow different schemes, such as functional or horizontal divestiture and full

division.

B. TAILORING ANTITRUST REMEDIES IN THE HIGH-TECH SECTOR

In the high-tech sector, remedies so far applied have been both fines and structural

and behavioural remedies.35

With regard to fines, since their goal is preventing infringers from further adopting

anti-competitive behaviours as well as deterring all market operators from implement-

ing similar conduct in turn, calculating the right level is crucial. For this reason, ``in

assessing the gravity of an infringement for the purpose of fixing the amount of the fine,

[authorities] must take into consideration not only the particular circumstances of the

case but also the context in which the infringement occurs and must ensure that its

action has the necessary deterrent effect, especially as regards those types of

infringement which are particularly harmful to the attainment of the objectives of

the community''.36 Therefore, fines that are not fine tuned will not perform the

function they are supposed to. When the goal is that of maintaining high-tech markets

open and the infringer is a company of Microsoft's size, fines alone may not be

sufficient. In the Microsoft IV Commission's decision, for example, the fine amount has

been vastly criticized since it was not deemed high enough to discourage the infringer

32 Although most of the literature on behavioural remedies is focused on merger control, nevertheless principlesapplicable to the other cases can be therein elicited. See generally Ariel Ezrachi, Behavioural Remedies in EC MergerControl Scope and Limitations, 29 W.Comp. 3, 459 (2006); and, with regard to the practice of raising rival costs, WillardK. Tom and Gregory F. Wells, Raising Rivals' Costs: the Problem of Remedies, (2003) 12 Geo. Mason L. Rev 395.

33 See the Court of Appeal statement in Microsoft, 253 F.3d § 80: ``divestiture is a remedy that is imposed onlywith great caution, in part because its long-term efficacy is rarely certain''.

34 On the structural remedies generally see Russell L. Weaver, The Rise and Decline of Structural Remedies,(2004) 41 San Diego L. Rev. 1617; and, with regard to the practice of raising rival costs, Willard K. Tom andGregory F. Wells, as note 32 supra, at pp. 395±396.

35 For a legal and economic analysis of remedies in network markets see Howard A. Shelanski and J. GregorySidak, Antitrust Divestiture in Network Industries, (2001) 68 U. Chi. L. Rev. 109-152.

36 Musique diffusion francËaise, (Case C 100-103/80) as mentioned in Ivo Van Bael and Jean-Francois Bellis, asnote 30 supra, at p. 1106.


from similar behaviour.37 A further drawback of fines lies in the companies' attitude to

recoup their losses on prices which can determine a rise in consumer prices. This is

even truer with regard to an operator entrenched in the technological frontier who is

less likely to step back because of a fine that can be easily recouped on final prices. It is

questionable, thus, that such a remedy could, at least when not coupled with others,

achieve the goal of lowering barriers to entry and of making IPRs perform their

functions of incentives to innovate.38

A second means theoretically available to protect and promote competition in the

high-tech sector is offered by structural remedies which in the United States have been

occasionally proposed or used.39 In Microsoft III, for example, Jackson J.'s decision

turned on a divestiture remedy; but parts of the decision were reversed by the Court of

Appeal. A huge debate arose over that divestiture remedy. Agreement lacked on either

the real necessity to adopt it or, if consent was reached on this point, the effectiveness of

such remedy.40 As a matter of fact, structural remedies are unlikely to be appropriate in

high-tech markets, where innovation can derive from synergies between complemen-

tary markets, and neat boundaries between markets are not easy to draw. Conversely,

such remedies may have the effect of limiting research and development to a single

product or within a sole market. Firstly, functional or horizontal divestiture (i.e.

dividing a company along product lines) would run the risk of jeopardizing innovation

since operators can be limited from fully exploiting the innovative potentialities

deriving from implementing combining functions of products belonging to adjacent

sectors. Part of the technical progress is a result of integrating functions performed by

different products. Had this not been possible due to a structural division of the high-

37 It has been noted that ``[f]or violations committed by a single offender, a necessary condition for deterrenceto work is that the expected fine, discounted for the probability of detection and punishment, exceeds the gainwhich the offender expected to obtain from the violation'' (Wouter P.J. Wils, as note 31 supra, at p. 183). Giventhis, there is no doubt that the fine imposed on Microsoft does not generate deterrent effects. In Microsoft IV, theCommission imposed a fine of around C= 500 million which is unlikely to have a strong deterrent effect and the samecan be held for CFI's fine of C= 497 million. As noted by Norman W. Hawker, Briefing Paper on the European MicrosoftCase, (23 March 2004), available at <http://www.antitrustinstitute.org/recent2/309.cfm>, concerning theCommission's fine, though the largest ever imposed, this amounts to little more than 1% of Microsoft's availablecash and less than a week's worth of revenues. The low level of fine predictability is then a further element to takeinto account in assessing such a remedy's effectiveness to perform its function (on the predictability of fines underArticle 82 EC Treaty see Robert O'Donoghue and A. Jorge Padilla, The Law and Economics of Article 82 EC, HartPublishing, Oxford and PortlandÐOregon, 2006, pp. 717±718).

38 At the opposite, favourable to private damages as the most efficient remedy: Richard J. Gilbert and MichaelL. Katz, An Economist's Guide to U.S. v. Microsoft, Competition Policy Center Working Paper No. CPC01-19,(May 2001), available at <http://repositories.cdlib.org/cgi/viewcontent.cgi?article=1007&context=iber/cpc9>.

39 The main US cases on monopolization in high-tech markets that dealt with structural remedies aresurveyed by Frederic M. Scherer, Technological Innovation and Monopolization, (March 2006), forthcoming in W. D.Collins (Ed.), Issues in Competition Law and Policy, American Bar Association, available at <http://www.ksg.har-vard.edu/m-rcbg/papers/scherer_technological_innovation.pdf>.

40 For a general survey of divesture effectiveness see Robert W. Crandall, The Failure of Structural Remedies inSherman Act Monopolization Cases, AEI-Brookings, Working Paper No. 01-05, (2001), (available at <http://www.ssrn.com>). On the specific issue of effectiveness of Jackson J's divestiture proposal see, amongst the wideliterature on the topic, Kenneth Elzinga et al., United States v. Microsoft: Remedy or Malady?, 9 Geo. Mason L. Rev633, (2001); Robert H. Lande, Why Are We So Reluctant to `Execute' Microsoft?, 1 Antitrust Source 1, (November2001); Thomas M. Lenard, Creating Competition in the Market for Operating Systems: Alternative Structural Remedies inthe Microsoft Case, 9 Geo. Mason L. Rev. 803, (2001); Stan J. Liebowitz, A Fool's Paradise: A Windows World After aForced Breakup of Microsoft, (25 February 2000), available at <http://ssrn.com/abstract=218178>.


tech sector, products we have now, such as current computers, would not be

manufactured and marketed. Secondly, full division (i.e. breaking up a company into

identical vertically integrated companies), though it appears to maintain efficiencies of

vertical integration, nevertheless, presents disruptive effects generating very high costs.

Moreover, this may divide up products which have not been fully developed yet,

thereby diminishing their feasibility.41 As for fines then, doubts arise as to structural

sanctions being effective means for promoting innovation, even more when they stand

alone. Moreover, it has been noticed that ``While courts can modify consent decrees or

lift injunctions, they cannot retroactively change the consequences of divestiture with

grave difficulty''.42 This is even truer when the markets in hand belong to the high-

technology sector where the pace of change is faster than in traditional markets.

The third means that can be applied in the high-tech sector are behavioural

remedies. They have been deemed likely to be the main (and probably most efficient)

tool to open up those markets which by reason of their structure tend towards

entrenchment.43 As a matter of fact, behavioural remedies are characterised by high

flexibility and reversibility. On the other hand, they generate uncertainty and significant

costs to the extent of being considered inferior to permanent one-off changes. However,

in changing market realities, reversible remedies may indeed be effective, as they appear

``apt to ensuring access to infrastructure and technology, preventing the creation of a gate

keeper and lowering barriers to entry''. When vertical elements and foreclosure of

technological key assets are thus involved, structural remedies, specifically divestiture,

may be less effective and even impede efficient, on-going research. Furthermore,

behavioural obligations can at times even ``provide a clear-cut remedy sharing

characteristics of structural remedies''.44 Because of their heterogeneous nature, a neat

line between structural and behavioural remedies is not easily struck.45 In fact, obligations

as those above mentioned are also considered to modify the distribution of intangible

assets amongst rival companies thereby altering the market structure.46

41 On divestiture in the software market see Thomas M. Lenard, as note 40 supra (arguing that, in Microsoft III,a hybrid structural remedyÐsuch as that elaborated by R. Craig Romaine and Steven C. Salop, Slap their Wrists?Tie their Hands? Slice Them Into Pieces? Alternative Remedies for Monopolization in the Microsoft Case, (Summer 1999) 13AntitrustÐwould be the most efficient).

42 E. Thomas Sullivan, The Jurisprudence of Antitrust Divestiture: The Path Less Travelled, 86 Minn. L. Rev. 612,(2002).

43 Ibid., at p. 565, arguing that remedies alternative to divestiture should be considered in dynamic markets,namely conduct-based remedies, because the cost of correcting market failure should not exceed theanticompetitive injury visited on consumers.

44 Ariel Ezrachi, as note 32 supra, at pp. 460±461.45 FrancËois LeÂveÃque, The Controversial Choice of Remedies to Cope with the Anti-Competitive Behavior of Microsoft,

(2000) Berkeley Program in Law & Economics, Working Paper Series Paper 34, 9, available at <ttp://repositories.cdlib.org/cgi/viewcontent.cgi?article=1055&context=blewp>. Similarly Robert O'Donoghue and A.Jorge Padilla, as note 37 supra, at pp. 718±719.

46 Howard A. Shelanski and J. Gregory Sidak, as note 35 supra, at p. 110, argue that compulsory licensing ofintellectual property and mandatory access to essential facilities are general example of redistribution of competitiveassets in the relevant market, which can be adopted in alternative to breaking companies in pieces. However, theeffects are not completely similar insofar an IPR cannot always provide the licensee with the same advantage itprovides the licensor with. Exploitation of IPRs is not only a mere fact of legal authorisation, but it also comprisesknow-how, qualified personnel and the like, which are elements not necessarily included in the licence.


C. WHEN REMEDIES TARGET IPRS

Regardless of their classification as behavioural or structural remedies, it is worth

mentioning that the target of antitrust intervention can vary from firm outputs (such as

prices, quantity, and quality) to firm organizations and structures, as well as to

intangible assets (such as IPRs and the like). In the high-tech sector in general, and in

the software market in particular, the antitrust intervention tends to target IPRs as the

main asset developed by software market operators. This is not a fixed rule though.

Pecuniary sanctions are adopted too (as in Microsoft IV) as well as remedies affecting

firms' assets different from IPRs, or conduct not related to IPRs. Yet, whenever

remedies are targeted on IPRs, various issuesÐsome of them we have already been

throughÐarise because of the incentive function that IPRs perform and the risks

related to innovation being constrained by inappropriate remedies. As a result, the

application of antitrust remedies in this sector has been even more criticized than in

other markets as accused of lowering rightholders' incentives to innovate. The tailoring

of such intervention needs thus to be carefully thought through in order to eliminate

the anti-competitive effects that exclusionary innovation can generate without stifling

its pro-competitive effects.

In detail, remedies more often adopted in the software market can be classified into

(i) disclosure obligations; (ii) mandatory licensing; and (iii) unbundling.

(i) With regard to disclosure obligations, in the Microsoft III and IV decisions they

have been adopted in relation to the new APIs in use between WOS, IE and WMP as

well as in relation to interoperability information and protocols necessary to non-

Windows group servers to communicate with Windows group server OS. In both cases

Microsoft resisted asserting that, firstly, copyright and trade secret law entitled their

foreclosure.47 And, secondly, mandatory disclosure would limit its incentives to

innovate. In Microsoft IV the defence went even beyond the mere claim of ownership and

exercise of IPRs to the extent of alleging the decrease in R&D that sharing IPRs would

cause to Microsoft whose products were alleged to result from huge investments.

Despite such a defence, in both decisions the principle emerges that a lack of disclosure

would prevent competitors from bringing their products to the same level of Microsoft's

middleware or group server OS as well as develop other and more advanced versions.

Therefore, ``incentives to innovate are much likely to vanish if competition is eliminated

in the downstream market than if the obligation to disclosure is sustained, because

Microsoft would have very little incentive to innovate on that market. . . . Furthermore,

because of increased competition on the downstream market, Microsoft would probably

feel more competitive pressure also on the upstream market''.48

47 It has been noted that in the information technology sector the combination of copyright and trade secretgives rise to a stronger software protection: the ``technology copyright'' (G. Ghidini, Intellectual Property andCompetition Law The Innovation Nexus, Edward Elgar, Cheltham UKÐNorthampton MA USA, 2006, pp. 62±67).

48 Simonetta Vezzoso, The Incentives Balance Test in the EU Microsoft Case: A More `Economics-Based' Approach?,27 ECLR 385, (2006).


(ii) In certain situations, though, disclosure is not enough: In order to develop

products competitive with those of the incumbent entrenched on the technological

frontier, competitors may also need an access to the IPRs possessed by the incumbent

which goes further the mere disclosure, and encompasses the right of reproducing the

protected materials. This tends to happen when patents are involved, yet it can also be

the case for copyrighted materials which need to be reproduced in the products of

competitors. Without an IPR licence competitors would infringe the incumbents'

IPRs. Therefore, whenever the owner refuses to ``collaborate''Ðthereby restraining

potential competition in both down and upstream marketsÐthe appropriate remedy

consists of a compulsory licence.

Although such a remedy has been even more debated than disclosure obligations,

compulsory licensing is not new to competition law nor to intellectual property law.49

While in the European and other national patent law systems compulsory licences have

been admitted whenever necessary for incremental innovation;50 in the United StatesÐ

where such a provision has not been adoptedÐcompulsory licences have still been

widely used during the 40s and 50s, and there is some evidence that this did not lead to

a decrease of investment in the American high-tech sector, rather, such mechanism

was, and still is, deemed to lead to an increase.51 With regard to the software market in

particular, whenever the innovator is entrenched on the technological frontier,

compulsory licences appear to be the only remedy to provide new incentives to

innovate.52 By sharing fundamental inputs, the monopolist is forced to keep walking

the innovation path. It has been stressed that ``[t]he issuance of a mandatory licensing or

cross-licensing scheme . . . would allow a firm to keep pace with the changing business

environment and reap the rewards from its valuable intellectual property, while

preventing the firm from wielding undue power derived from the intellectual

property''.53

Although analysing the specific conditions required to justify compulsory licensing

is outside the scope of this article, it is worth mentioning that many doctrines have been

used to justify the compulsory licensing imposed in Microsoft III and IV. Firstly, the

49 A survey of community case law on compulsory licences in Andreas Heimann, as note 27 supra, at 65.50 Gustavo Ghidini and Emanuela Arezzo, Patent and Copyright Paradigms vis-aÁ-vis Derivative Innovation: The

Case of Computer Program, (2005) 36 IIC 159.51 Frederic M. Scherer, Antitrust, Efficiency, and Progress, (1987) 62 NYULR 997.52 Philip J. Weiser, The Internet, Innovation, and Intellectual Property Policy, (2003) 103 Colum. L. Rev. 534,

arguing the necessity for a compulsory licence system whenever a de facto standard is achieved in the softwaremarket. Same opinion is expressed by Makan Delrahim, Forcing Firms to Share the Sandbox: Compulsory Licensing ofIntellectual Property Rights and Antitrust, (10 May 2001), addressed to the British Institute of International andComparative Law, London, England (transcript available at <www.usdoj.gov/atr/public/speeches/203627.htm>).The Author's favour towards compulsory licences (whenever there is harm to competition) is conditioned to anarrowly drafting though. Contra Robert P. Merges, Who Owns the Charles River Bridge? Intellectual Property andCompetition in the Software Industry, UC Berkeley Public Law and Legal Theory Working Paper No. 15, (October1999) (available at <http://ssrn.com/abstract=208089>). Arguing that standards-setting and pooling arrangementsare a more useful remedial model than compulsory licensing whenever a head-on clash between property rights andinnovation is present. In fact, compulsory licences have not proved to be such an efficient tool due to a burdensomeprocedure to go through in order for them to be granted.

53 E. Thomas Sullivan, as note 42 supra, at 606.


essential facility doctrine and its `èxceptional circumstances'',54 even though in Microsoft

IV its typical criteria appear slightly modified so to replace the traditional ``new product

or service test'' with a more economic-oriented `ìncentive-balance test''.55 Secondly,

the leverage doctrine offers an interpretative tool advocated by some scholars

notwithstanding the scepticism that it usually raises.56 Thirdly, the predatory

innovation doctrine can also give some hints to illustrate the Commission and Court

of Appeal's reasoning,57 as well as the unilateral refusal to deal doctrine.58 However,

regardless of the doctrine adopted, principles underlying the application of compulsory

licensing in the software market rest on the will to protect and promote the pace of

innovation. This can be performed by requiring that assets necessary to maintain

markets open and competitive are shared without frustrating the innovator efforts

which need to be counterbalanced by the granting of reasonable royalties.59

Disclosure obligations and mandatory licensing share the problem related to the

assessment of what must be disclosed and in case licensed. Neither the Court of Appeal

in Microsoft III, nor the Commission in Microsoft IV, thoroughly addressed this and other

crucial aspects though.60 A suggestion is likely to be found in Microsoft IV where the

Commission considered it necessary to designate a trustee for monitoring the

implementation of its orders as well as for assessing whether the information disclosed

by Microsoft was complete, accurate and timely.61 However, strong criticism has been

aroused by the decision to devolve determination of licensing conditions to third

parties.62 A second problem related to the sole mandatory licensing concerns the

54 On the later development of Essential Facility Doctrine in both the European Union and United States seeDaniel Kanter, IP and Compulsory Licensing on both sides of the Atlantic, 27 ECLR 351, (2006).

55 The shift from a new product/service test to an incentive-balance test has been stressed by StevenAnderman, Does Microsoft offer a New Paradigm for the Èxceptional Circumstances' Test and Compulsory Licenses under ECCompetition Law?, 1 CompLRev 7, (2004); Valentine Korah, The Interface Between Intellectual Property Rights andCompetition in Developed Countries, 2 Script-ed 463, (2005); FrancËois LeÂveÃque, Innovation, Leveraging and EssentialFacilities: Interoperability Licensing in the EU Microsoft Case, 28 W.Comp. 1, 71, at 75±80, (2005).

56 FrancËois LeÂveÃque, as note 55 supra, at 80, arguing that in Microsoft IV the application of the new incentive-balance test is not convincing, whereas the leverage doctrine would enable application of the compulsory licence ina more sound way.

57 Maria LillaÁ Montagnani, as note 25 supra, at 304.58 Herbert Hovenkamp, Mark D. Janis, Mark A. Lemley, Unilateral Refusals to License in the US, in FrancËois

LeÂveÃque and Howard Shelanski (Eds), Antitrust, Patents and Copyright, Edward Elgar, Cheltenham UKÐNorthampton MA USA, 2005, p. 12; Cyril Ritter, Refusal to Deal and Èssential Facilities': Does Intellectual PropertyRequire Special Deference Compared to Tangible Property?, 28 W.Comp. 3, 281, (2005).

59 John Temple Lang, The Application of the Essential Facilities Doctrine to Intellectual Property Rights underEuropean Competition Law, in FrancËois LeÂveÃque and Howard Shelanski (Eds), as note 58 supra, at p. 56.

60 Moreover, with regard to the API disclosure, it is worth pointing out that in the European Unionaccording to Art. 6 of Council Directive 91/250/EEC of 14 May 1991 on the legal protection of computerprograms (OJ L 122, 17.5.1991, 4246) decompilation is allowed to certain conditions unless ``the informationnecessary to achieve interoperability has not previously been readily available''. Although it is in Microsoft'spractice to make available some APIs to enable compatibility between its OS and independent applications, theinformation disclosed has often been considered insufficient from independent software developers. This raises thequestion as to who can lawfully decide when the supra mentioned condition of `ÀPI availability'' is met.

61 Microsoft IV decision, as note 6 supra, at 1403. Similarly did the Commission in the interim measuredecision on IMS (Commission Decision 2002/165/EC, Case COMP D3/38.044, NDC Health, OJ 2002, L59/18).

62 In both Microsoft IV and IMS decisions, determination of licence fees and conditions was devolved to anexpert without providing any guidance. In Microsoft the Trustee was furthermore devolved with wider power as tothe implementation of the Commission's order. See James Killick, IMS and Microsoft Judged in the Cold Light of IMS,1 CompLRev 32, 46, (2004).


``reasonable price'' to be set for the compulsory licences.63 On the one hand, incentives

to innovate can be compensated by royalties only when the latter are properly fine

tuned;64 on the other hand, full freedom to set royalties enables the licensor to make

contract terms too onerous for the prospective licensees. For these reasons, compulsory

licensing obligations are usually accompanied by the specification that licensing terms

must be ``fair, reasonable, and non-discriminatory''. However, such a statement lacks

content on the methodology to follow in order to comply with the obligation.65 In the

absence of any sensible guidance on this point, it could be discussed whether the

proposal of adopting market-mechanisms, such as the licensing auction which was

proposed as an alternative to the structural remedies proposed by Jackson J. in Microsoft

III, could be applied here also.66 An auction system would set the priceÐvia the

competitors' bid gameÐto which all licensees will have access to the required input.

Unlike a compulsory system that allows all perspective licensees to obtain the claimed

input at the same ``reasonable'' price, the price set via an action system could prevent

some of them from accessing the input due to their different parameters, such as

structural costs and the like. Although this may be deemed discriminatory, as long as

the bid game is considered an efficient system to set the price, the exclusion of some

prospective licensees is the equilibrium that the market would naturally reach. Hence,

compulsory licensing coupled with an auction system to set the price morphs the

reasonable price in a market price, i.e. a price that results from the game amongst

perspective licensees' bids. There remains some doubt that such a system would impair

the licensing terms that the Commission requires the compulsory licensing to follow.

(iii) Firstly, in relation to unbundling, such a remedy is less targeted on IPRs than

those previously analysed. Yet, it equally needs to be cautiously applied since, according

to the chosen design, it may be effective to the extent of altering market dynamics. In

fact, unbundling design can significantly vary, as in Microsoft III, IV, and the recent

Korean decision.67 In the US consent decree, the unbundling was agreed in order to

limit WIE visibility by requiring computer sellers to hide visible means of access to it. In

this way WIE remained present to be activated upon request by a skilled consumer or

by a third-party. On the other hand, the European Commission required Microsoft to

engage in mandatory versioning and in offering computer sellers versions of Windows

with and without WMP. The Korean Fair Trade Commission went even further by

63 On the difficulties aroused by the criterion of ``reasonable royalty'' from an economics perspective, seeFrancËois LeÂveÃque, as note 55 supra, at 86±90.

64 Damien Geradin, Limiting the Scope of Article 82 of the EC Treaty: What can the EU learn from the US SupremeCourt's Judgement in Trinko in the wave of Microsoft, IMS, and Deutsche Telekom?, 41 CMLRev. 1519, (2004).

65 Robert O'Donoghue, A. Jorge Padilla, as note 37 supra, at pp. 726±731, discussing four basic options todetermine royalties.

66 Herbert Hovenkamp, as note 31 supra, at p. 301: ``Other, more creative, Microsoft remedies never gotbeyond the discussion stage. One of these would have required Microsoft to auction off non-exclusive licences ofits Windows source code to four or five purchaser, each of which could then develop its own version of Windows,creating a competing network environment.''

67 On the decision, see Warren S. Grimes, Korea Fair Trade Commission's Microsoft Decision, FTC:WATCH#668, January 30 2006 <http://www.antitrustinstitute.org/recent2/481.pdf>.


ordering Microsoft to sell a version of its OS including neither WMP nor Windows

Messenger functionalities; by requiring Microsoft to facilitate consumer downloads of

third party media players and messenger products selected by the Commission; and by

prohibiting Microsoft from selling in Korea a version of its server software that includes

Windows Media Services.68

Secondly, unbundling is likely to be an efficient remedy only when adopted in a

real ``modularized system'': where all elements are acquirable and combinable

independently from Microsoft and OEMs' decisions, and all prices of unbundled

versionsÐas well as single elementsÐare clearly set.69 The issue of setting prices for

bundled and unbundled versions is again the core point of such remedy, since charging

the same price for both can easily perpetuate the tying in the absence of commercially

realistic reasons for consumers to not take the bundle.70 On the other hand, complete

unbundling orders may severely jeopardize innovation by restraining functionalities of

adjacent markets from being integrated in a sole product.71 The result would be similar

to that of a horizontal divestiture order. Therefore, it has been proposed that when the

result of maintaining markets open is to be accomplished, this could be rather achieved

by introducing a must-carry obligation.72 Such obligation would require the operator

that proceeded to the integrationÐsuch as MicrosoftÐto distribute both the integrated

product and the competitive products on competitors' request. In principle a must-

carry obligation may be more appropriate than an unbundling order since, similarly to

unbundling, it may serve to keep markets open but, differently from it, it does not run

the risk to stifle the innovation that integration can generate. In practice though, must-

carry obligations raise the question as to how the price for the service of carrying should

be set. Would the obligation really be adopted, a likely solution to the problem above

mentioned could be explored by analysing the principles underling the Access and

Interoperability directive, adopted within the European regulatory framework for

electronic communications, where it established mandatory access to networks of

operators with significant market power.73 The Directive seeks also to provide an

68 Strong criticism on the decision in the Statement Of Deputy Assistant Attorney General J. BruceMcdonald Regarding Korean Fair Trade Commission's Decision In Its Microsoft Case, (7 December 2005)available at <http://www.usdoj.gov/atr/public/press_releases/2005/213562.htm>.

69 On the difficulties of pricing unbundled elements see Damien Geradin and J. Gregory Sidak, European andAmerican Approaches to Antitrust Remedies and the Institutional Design of Regulation in Telecommunications, (2003),available at <www.ssrn.com> 10-12.

70 Robert O'Donoghue and A. Jorge Padilla, as note 37 supra, at pp. 732±733.71 A sound examination of bundling pro- and anti-competitive effects has been carried out by Kai-Uwe

KuÈhn, Robert Stllman and Cristina Cafarra, Economic Theories of Bundling and their Policy Implication in Abuses Cases:An Assessment in Light of the Microsoft Case, CEPR Discussion Paper No. 4756, (November 2004), available at<http://ssrn.com>.

72 Andreas Heinemann, as note 27 supra, at 81. It must be mentioned though, that the must-carry obligationstill faces the issue of pricing the service imposed on the incumbent.

73 The European new regulatory framework for electronic communications infrastructure and associatedservices is formed by one framework directive and four directives plus regulation and soft regulation. Information isavailable at <http://europa.eu.int/information_society/topics/telecoms/regulatory/new_rf/index_en.htm#Introduc-tion>. In this context, Directive 2002/19/EC (OJ L 108, 24.4.2002, 7-20) establishes rights and obligations foroperators and for undertakings seeking interconnection and/or access to their networks or associated facilities, aswell as objectives for national regulatory authorities with regard to access and interconnection.


answer to the key regulatory questions of when to provide access, to what facilities and

at what price. With regard to the price, it must be noted that although mandating access

on operators with significant market power could serveÐas well as a must-carry

obligationÐto increase competition, however, if the price of accessÐor the price

offered to carry competitive productsÐis too low, it can remove incentives to further

investments.

VI. CONCLUSION

Although in the high-tech sector in general, and in the software market in

particular, there are situations needing an antitrust intervention, indications as to whether

and how competition law may limit IPR exclusivity have yet to be fully and consistently

elaborated. Nevertheless, the recent Microsoft III and IV decisions seem to provide some

hints at least as to the goals that should be aimed at, namely: maintenance of open

markets, and lowering barriers to entry so as to protect current and potential

competition and promote innovation within the market.

To match theory with practice is not an easy task though; the crucial point is the

design of tools capable of accomplishing the above mentioned goals. Notwithstanding

the feasibility of identifying remedies in generalÐas a case by case analysis is neededÐ

some considerations can still be formulated with regard to the remedial phase.

Firstly, it should be observed that an optimal remedy does not exist. Rather, a

combination of remedies should be examined in order to establish the trade-off between

stifling a firm's incentives to innovate through an antitrust intervention on the one hand,

and opening up high-tech markets to competition by the adoption of antitrust remedies

on the other. Striking such a balance needs, however, the discussion to be less focused

on the efficiency of ex post remedies against ex ante remedies, or on the identification of

the most efficient remedy amongst those offered by competition law. On the other

hand, awareness should be reached on the blurring taking place between ex ante and ex

post remedies in both the US and EU systems. In this context the experience of the

telecoms sector is exemplificative: a combination of both tools, or hybrid means, has

been adopted to open up the sector and maintain its competitiveness.74 Hence, although

there is concern as to the regulatory tone that the antitrust intervention has been

progressively assuming in network industries (starting with the telecoms sector and

moving to the more recent Microsoft decisions),75 it is in the light of the European

telecoms sector-specific regulation that a solution could be explored. Indeed, the new

European regulatory framework for the electronic communications is shaped by

competition law principles demonstrating that even sector-specific regulation may be

market-oriented.76 A balanced solution for remedies in the software, and in general in

74 Damien Geradin and J. Gregory, as note 69 supra, at 2.75 Ibid., at 12.76 See note 73 supra.


the high-tech sector, could thus be a combination of ex ante and ex post remedies (or

hybrid remedies) whenever both are clearly inspired by competition law principles.

Secondly, even in this context the majority of remedies forming the optimal

combination should consist of economic instruments rather than tools of mere

command and control.77 While the former consist of tools used to incentivise firms to

adopt socially desirable conduct so that firms' behaviours depend on a cost-benefit

analysis;78 the latter constrain firms' behaviours contrary to social welfare but not their

incentives to do so. For example, the remedy of licensing is likely to belong to those

market-based instruments incentivising the sharing of essential inputs insofar the

benefits of licensing overcome the costs of refusing a licence. On the other hand,

disclosure of APIs has been deemed a mere regulatory tool as it does not present any

cost-benefit analysis but it is imposed by the authority.79 However, even though there

may be less incentive-based remedies, the trade-off between stifling firms' incentives to

innovate and promoting competition may still tilt towards the promotion of

competition to the extent of justifying the imposition of certain behaviours, such as

in the case of disclosure obligations. Moreover, the distinction between economic

instruments and regulatory tools does not seem to mirror the reality of high-tech

markets, where boundaries cannot be neatly drawn. More convenient would then be

carrying both trade-off and cost-benefit analyses on the whole of the remedies designed

for a case, rather than on each remedy separately.

77 FrancËois LeÂveÃque, as note 45 supra, at 9, discussing the distinction between command and controlremedies and incentives remedies and tracing it back to the policy and target chosen.

78 Ibid., at 10. The author mentions, as examples of economic instruments, ``a Pigovian tax on CO2emissions will result in pollution abatement whenever the marginal abatement cost is higher than the amount of thecharge per unit of emissions. Liability rules are another classical example of economic instruments. They provideincentives to avoid accidents. They leave firms to decide to invest in accident prevention or pay damages when theyoccur''.

79 Ibid., at 11.


Remedies to Exclusionary Innovation in the Hight-Tech Sector

Documents

Transcript of Remedies to Exclusionary Innovation in the Hight-Tech Sector