1

Why Distance Matters: The Dynamics of Offshore Location Choices

Stephan Manning Duke University, The Fuqua School of Business

1 Towerview Drive, Durham, NC 27708, USA

Email: sdm24@duke.edu

Marja Roza Erasmus University, Dep. of Strategic Management and Business Environment

P.O. Box 1738, 3000 DR Rotterdam, The Netherlands

Email: MRoza@rsm.nl

Arie Y. Lewin Duke University, The Fuqua School of Business

1 Towerview Drive, Durham, NC 27708, USA

Email: AYL3@duke.edu

Henk W. Volberda Erasmus University, Dep. of Strategic Management and Business Environment

P.O. Box 1738, 3000 DR Rotterdam, The Netherlands

Email: h.volberda@rsm.nl

Duke CIBER Working Paper

2

Why Distance Matters: The Dynamics of Offshore Location Choices

Abstract

Offshoring of technical and administrative work has become an established business practice in

recent years. One key aspect of offshoring decisions is the choice of location. Prior research

suggests that availability of expertise and low cost labor are major location choice factors,

whereas geographical distance does not matter much because of advancement of IT. By contrast,

we show that access to talent pools, external expertise and geographical proximity are important

trade-offs and part of a competitive dynamic that unfolds over time. Based on comprehensive

data of early and more recent offshore investments of U.S. and Dutch companies, we conduct a

multi-level analysis of the likelihood of nearshore vs. farshore implementations over time.

Findings suggest a co-evolutionary dynamic of location choices, changing industry and labor

market conditions. Results point to limitations of conventional theories of internationalization

and may inspire follow-up studies accounting for these multi-level dynamics.

Keywords: Offshoring, Nearshoring, Location Choice, Co-evolution, Search for Talent

Introduction

The relocation and sourcing of business functions and processes outside national borders – also

called: offshoring – has become an established business practice (Doh, 2005; Kenney et al.,

2009; Manning et al., 2008). Since the beginning of the 1990s, in particular U.S. companies have

engaged in offshoring administrative and technical work, including IT, finance & accounting,

and legal services, but also more advanced functions, including software and product

development. In more recent years, an increasing number of European companies have also

started to relocate various business functions abroad, in particular to emerging economies.

Reducing labor costs and accessing growing pools of qualified personnel outside their home

countries have been major offshoring drivers (e.g. Lewin & Couto, 2007; Lewin et al., 2009).

3

One key aspect of making offshoring decisions is the choice of the offshoring location.

The rationale for selecting particular locations has attracted a number of studies in recent years

(e.g. Doh et al., 2008; Bunyaratavey et al., 2007; A.T. Kearney, 2004). Most of these studies

stress the importance of labor cost advantages, level of education and expertise, as well as

language capabilities as location selection factors, but they also list a number of risks associated

with locations, such as political stability, wage inflation, and protection of intellectual property

(see e.g. Doh et al., 2008; A.T. Kearney, 2004; Lewin & Couto, 2007). Moreover, some recent

studies suggest that over time specialized geographical IT and knowledge service clusters have

emerged, e.g. in Russia, Eastern Europe, India and China, that attract function-specific

investments (e.g. Bresnahan et al., 2001; Dossani & Kenney, 2007; Manning et al., 2008).

Interestingly, however, most studies on offshore location strategies neglect the

importance of geographical distance. Many scholars believe that distance does not matter much

in service offshoring (e.g. Blinder, 2006). The main argument is that advanced ICT has

facilitated the modularization and reorganization of tasks and has made long-distance

coordination and communication less costly (e.g. Kenney et al., 2009; Metters & Verma, 2008).

However, empirical evidence, e.g. data collected by the Offshoring Research Network (ORN),

suggests that distance does matter in offshore investment decisions: A significant number of

companies, in particular from Western Europe, have preferred nearshore locations for their

offshore investments. Nearshoring means that captive or third-party offshore implementations

are made relatively close to the home country. By contrast, U.S. companies have mostly

offshored to farshore locations, such as India, China or the Philippines. Only recently, more

European companies seem to have shifted operations to more distant locations.

4

We seek to investigate preferences for nearshore vs. farshore locations in greater detail as

well as changes over time, based on comprehensive data of U.S. and Dutch companies, collected

by ORN. Thereby, we take a dynamic, multi-level analytical approach by looking at influences at

the macro-economic, population, firm and task level. We propose, based on our data, that a co-

evolutionary perspective (see Lewin & Volberda, 1999; Hutzschenreuter et al., 2008) helps

understand the dynamics of offshore location preferences over time (see also Manning et al.,

2008; Lewin et al., 2008). Our findings may not only stimulate future research on offshoring

strategies and location choices. They also shed light on the dynamics of internationalization in

more general. We argue that existing – rather market-oriented – theories of internationalization,

e.g. the OLI framework or the Uppsala model, are insufficient in explaining these recent

offshoring dynamics. A co-evolutionary model can be a promising alternative starting point for

analyzing more recent and future trends of internationalization.

Offshore Location Choice and the Role of Geographical Distance

The choice of location is a key dimension in making offshoring decisions (see e.g. Doh et al.,

2008; Lewin & Couto, 2007; Bunyaratavey et al., 2007). In more general, offshoring refers to the

(re-) location of business processes and functions outside the borders of the home country (e.g.

Manning et al, 2008; Kenney et al., 2009). Offshoring is often confused with outsourcing, which

denotes the transfer of tasks and processes to external providers. Offshoring, however, may

include both captive models (setting up wholly owned subsidiaries) and offshore outsourcing

(engaging external international and local providers to deliver particular services). Also,

offshoring must not be confused with market-oriented foreign investment, e.g. setting up sales

operations. Rather, offshoring is a sourcing strategy designed to set up local operations

5

supporting domestic and global activities of the firm (see Manning et al., 2008). With regard to

offshoring of technical and administrative tasks, e.g. IT, finance and accounting, software and

product development, the most important factor being ‘sourced’ offshore is qualified personnel

(see Lewin et al., 2008; Manning et al., 2008), and – if available – external capabilities and

subject-matter expertise (offshore outsourcing).

The most important offshoring drivers for U.S. and Western European companies are

saving labor costs and accessing qualified personnel. Offshore location decisions to a great

extent reflect these general drivers (see e.g. Lewin & Couto, 2007). Accordingly, labor cost

advantages, size of the talent pool available and access to subject-matter expertise are important

criteria in choosing offshore locations for particular services (see e.g. Doh et al., 2008). Along

these lines, some scholars argue that the success of India and China as preferred offshore

destinations can be mainly explained by the quantity and quality of technically skilled labor

available (see e.g. Freeman, 2006; Dossani & Kenney, 2007). However, recent studies indicate

that in particular European companies often select nearshore locations close to their home

countries rather than India or China for making offshoring implementations (see e.g. Lewin &

Couto, 2007; A.T. Kearney, 2004). Notably, a number of U.S. companies also consider Canada

and Mexico in particular as alternatives to further remote locations. We seek to understand why

this is and, yet, why recently a shift towards more remote locations can be observed.

In more general terms, the question is why and to what extent distance matters in offshore

location decisions. The role of distance has been widely neglected in the more recent offshoring

literature. Some scholars even argue that because of advances in IT, decreasing long-distance

communication costs and the emergence of specialized geographical clusters providing talent

6

and services for MNCs globally, distance should not matter much in service offshoring (e.g.

Blinder, 2006).

The more general literature on internationalization however suggests that distance can be

an important factor in any foreign investment decision. Rugman and Verbeke (2004) for example

show, based on sales data, that the majority of foreign investments is made within the region of

the home country. One important reason why many companies set up manufacturing and sales

operations close to their home countries are cost and time advantages. For example, geographical

proximity to headquarters or other central locations may reduce transportation and travel time

and costs. In the offshoring context, in particular for those tasks that require intensive training of

staff and coordination with head quarters or other locations travel costs can be expected to be an

important factor in choosing locations. Related to this, time zone differences seem to affect

coordination and communication between locations (O’Leary & Cummings, 2007). On the other

hand, strategies targeted at establishing 24h knowledge factories may favor implementations in

more remote time zones (e.g. Gupta et al., 2007).

Another equally important and historically related reason why many companies seem to

prefer geographically close destinations for foreign investment is greater familiarity with

economic, cultural and institutional conditions in countries within the region (e.g. Rugman &

Verbeke, 2004). Most prominently, Johanson and Vahlne (1977) argue that investment in

contexts characterized by ‘psychic proximity’ facilitate learning that is necessary to expand the

global footprint. Psychic proximity relates to the accessibility of information about local market

conditions. Availability of location information in the context of offshoring can reduce search

costs, e.g. costs involved in finding talent or qualified service providers, which may facilitate and

speed up the implementation of offshore projects. Reasons why geographical proximity often

7

correlates with psychic proximity (or greater familiarity of context conditions) are e.g. the

embeddedness of companies in local and regional networks, trade relationships between

countries and prior foreign investments.

Similarly, IB scholars have argued, based on Hofstede’s concept of cultural distance

(Hofstede, 1980), that companies are more likely to invest in host countries that share norms and

values with home countries (e.g. Kogut & Singh, 1988; for a critical view, Shenkar et al., 2008).

Again, geographical proximity is often correlated with cultural proximity. One important factor

here is the use of the same or a similar language which facilitates intra- and interfirm

communication and coordination. In fact, similar norms and language have been listed as

primary reasons in previous offshoring studies for the preference of many Western European

companies to select locations in other European countries (e.g. A.T. Kearney, 2004). Related to

this, Xu & Shenkar (2002) argue that similar regulative, normative and cognitive institutions (see

in general, Scott, 2001) – in short: institutional proximity – may affect location choice. Along

these lines one could argue that the fact that many European countries belong to the European

Union and therefore adhere to the same or a similar legal system reduces institutional distance in

favor of nearshore investments.

Most of these theoretical approaches, however, have two major weaknesses. On the one

hand, they may not explain differences in location preferences. For example, according to ORN

data, U.S. companies seem to have a much higher preference for India and other remote locations

than most European companies (see Figure 1). Some European companies, in turn, e.g. Spanish

companies, also seem to prefer further remote locations, e.g. in Latin America. What’s more,

companies from the same home country often have different preferences for nearshore vs.

farshore locations (see in more detail below). On the other hand, established approaches seem to

8

insufficiently explain changes in location preferences over time. As we demonstrate in more

detail below, ORN data suggests that European firms have a greater preference for farshore

locations in recent years. Figure 2, for example, shows, based on ORN data, how the distribution

of offshore investments in India by country of origin has shifted over time from U.S. to European

countries. The Uppsala school of internationalization (Johanson & Vahlne, 1977) would argue

that European firms have learned over time to establish subsidiaries in locations formerly

characterized by great psychic distance. This however neither explains why most U.S. companies

offshored to India from early on, nor does it explain why in recent years even European

companies with no or little offshoring experience are more likely to go farshore than nearshore.

----------------------------------

Insert Figure 1,2 Here

----------------------------------

We argue that a multi-level analytical approach is needed to come to grips with the dynamics of

offshore location choice. As proposed by Hitt et al. (2007), Volberda and Lewin (2003) and

others, multi-level approaches serve to simultaneously take into account influences on strategic

decisions or managerial behavior coming from multiple systemic contexts – e.g. the firm itself,

as well as macro-economic, industry, and project implementation-level factors. Also, they invite

thinking about interactions between these levels, including intended and unintended

consequences of firm-level decisions on the very contexts in which firms operate. This basic idea

has been further developed in particular by co-evolution theory which seeks to explain firm-level

adaptation-selection within changing competitive environments (Lewin & Volberda, 1999). In

the following, we outline our analytical framework in more detail. We develop hypotheses to

explain the preference for nearshore vs. farshore implementations based on explanatory factors at

9

different levels. These factors will be derived mainly from existing offshoring studies as well as

the more general IB and internationalization literature.

Explaining Nearshore Investments: A Dynamic Multi-Level Framework

To come to grips with preferences for nearshoring we apply a multi-level analytical framework.

Multi-level analyses are suitable where a phenomenon, such as a strategic firm-level location

decision, cannot be easily explained just by factors at one level, e.g. firm-level preferences,

national or industry context factors, but only by the combination and interaction between these

levels (Hitt et al., 2007). In line with recent papers on the offshoring phenomenon (e.g. Manning

et al., 2008, Lewin et al., 2009; Kenney et al., 2009) we see firm-level offshoring decisions in

general – and location decisions in particular – co-evolving with industry and macro-level trends

as well as home and host country economic conditions. The propositions we develop in the

following, however, are not meant to be exclusive, but they are designed to help understand part

of the dynamics of factors influencing location decisions in general and nearshore vs. farshore

location choices in particular. Further studies are needed to analyze those as well as other

interacting factors in greater detail.

Macro-level effects: Offshoring decisions are influenced by factors that cannot be simply related

to firm, industry or country-level trends, but that relate to trends affecting strategic decisions

across firms, industries and countries. In the context of offshoring, the advancement of ICT can

be named as one of the key factors facilitating and accelerating recent offshoring trends (Kenney

et al., 2009; Metters & Verma, 2008). In terms of location choice, the argument has been made

that advanced ICT reduces communication and coordination costs and therefore makes

10

geographical distance less relevant (Blinder, 2006). In other words, advancement of IT should

make, ceteris paribus, nearshore location decisions less likely over time. In a similar vain,

offshoring has become an established business practice, promoted by firm- and industry level

experience, but also by a broader discourse, stimulated by business press, consulting and

academia (e.g. Levy, 2005). India in particular has attracted much attention in the offshoring

discourse in recent years (e.g. Dossani & Kenney, 2007). But also China has become a much-

celebrated location for offshore investments (see e.g. Huang & Khanna, 2003). The

establishment of offshoring as an accepted business practice across industries and in particular

the fast-growing popularity of India and China should also make, ceteris paribus, nearshore

location decisions less likely over time. We therefore propose:

H1 (Macro-level effect): The more recent an implementation is (independent of company

experience), the less likely is it a nearshore implementation.

Country-level effects: One of the most important drivers for companies to make offshore

investments is access to lower-cost personnel. At the same time, we have argued, based on

previous studies, that companies prefer to make investments closer to their home countries for a

combination of reasons, e.g. search costs, coordination costs, cultural and institutional proximity.

Initially, therefore, we predict companies to make use of nearshore resources, such as qualified

low-cost personnel if available. However, in the longer term, the very offshoring motivation to

access personnel and to lower labor costs (see e.g. Lewin & Couto, 2007) will favor farshore

investments. Therefore, we argue that, ceteris paribus, the more labor resources are available

farshore vs. nearshore the more likely will a company make a farshore investment. In addition,

we argue that the lower labor costs are farshore vs. nearshore, compared to the home country, the

11

less likely will a company make a nearshore rather than a farshore investment. Both arguments

have been supported by numerous economists, e.g. Freeman (2006). Importantly, these economic

conditions may change over time. Research has shown that growing competition for talent and

wage inflation may drive companies to shift operations to other ‘second-tier’ locations. While we

cannot directly predict effects at this time, we will consider the growing competition for talent at

particular locations, e.g. hotspots, as indicators for the changing availability of lower-cost

qualified personnel over time.

H2a (Country-level effect): The bigger the pool of qualified, lower-cost personnel

nearshore (ceteris paribus), the more likely is a nearshore implementation.

H2b (Country-level effect): The bigger the pool of qualified personnel available in

farshore vs. nearshore locations, the less likely is a nearshore implementation.

H2c (Country-level effect): The lower the cost of labor in farshore vs. nearshore locations

compared to the home country, the less likely is a nearshore implementation.

Population-level effects: Offshoring location decisions are also affected by behavior of other

firms, e.g. of the same industry or national origin. With regard to offshoring, clearly companies

of the same national origin show similar patterns of decision-making (see above). Part of this

similarity stems from a common domestic economic and institutional environment. Part of it,

however, relates to imitating behavior. Case studies indicate that pioneer companies entering a

location from a particular country attract followers of the same national origin. On the one hand,

this has to do with the development of trust in local business conditions. On the other hand,

companies often actively shape the local investment environment, e.g. by building institutional

links with local administrations and universities (e.g. Manning et al., 2008). As a consequence,

12

they ‘customize’ local service capabilities (Athreye et al., 2005), attracting future investment

from companies from the same business system. From the point of view of IB, this can be

explained by the reduction of psychic and institutional distance – e.g. the increasing familiarity

of locations among firms over time (Xu & Shenkar, 2002; Johanson & Vahlne, 1977). As,

according to theory, companies tend to be initially more familiar with nearshore locations, this

dynamic is likely to positively affect the likelihood of farshore implementations.

In a similar vain, we propose that experience with locations related to offshoring

particular functions, e.g. software, product development etc., may increase the likelihood of

companies offshoring this function to a particular location. This can be also called the

specialization effect. For example, Bangalore, India, has clearly attracted mainly IT and software

development-related investment, because, over time, specialized local service capabilities have

developed that attract companies across industries (Athreye et al., 2005; Dossani & Kenney,

2007). Again, we assume that this population or industry effect is stronger for farshore than it is

for nearshore investments. In other words, when it comes to nearshore implementations,

companies can be expected to be more knowledgeable about providers and qualified personnel

based on personal experience and ‘pioneers’ are not needed so much to make nearshore

investments attractive. In sum, we predict that population experience with farshore investments

as well as specialization effects may decrease the likelihood of a nearshore implemention.

H3 (Population-level effect): The more implementations have been made farshore

relative to nearshore across firms up to a certain point in time, the less likely is a

nearshore implementation at this time.

13

Firm-level effects: Strategy research suggests that in a competitive space firms differ in terms of

strategic choices (e.g. Child, 1972), because of different resource endowments, decision-making

processes, absorptive capacity, strategic goals and managerial discretion etc. Also they respond

to economic and institutional challenges, e.g. shortage of qualified personnel or EU enlargement,

in different ways (Oliver, 1991). We therefore propose that the very factors companies perceive

to be important when making location decisions for particular offshore projects matter.

With regard to nearshore implementations, clearly the importance of geographical

proximity in a location decision can be predicted to play a key role. Importantly, the very fact

that companies invest nearshore must be separated from the importance geographical proximity

has in a particular decision. This is because nearshore investments can be potentially made for a

variety of reasons, e.g. previous investments (e.g. manufacturing), availability of special

expertise etc. Proximity itself, by contrast, is assumed to highly correlate with low coordination

costs and familiarity of the location. In this context, perceived language capabilities and cultural

proximity can be expected to be associated factors positively contributing to nearshore rather

than farshore investments (e.g. Doh et al., 2009; A.T. Kearney, 2004). These considerations,

according to the literature, need to be traded off against the perceived importance of saving labor

costs and availability of personnel in a potential offshore location. As labor costs in India and

China (and other, second tier, Asian and Latin American locations) are significantly lower than

in Europe (including Eastern Europe) and Canada, we propose that the importance of low labor

costs and availability of labor as firm-level location factors have a negative effect on the

likelihood of nearshore implementations.

In line with the Uppsala school of management (Johanson & Vahlne, 1977), but also with

the resource-based view and its derivates, we further propose that firm-level offshoring

14

experience and – although not directly measured – advanced capabilities in coordinating offshore

operations may influence location decisions. Similar to our argument for the relevance of

population experience, we argue that firm experience with farshore investments is positively

related to further farshore rather than nearshore investments. This is because, in accordance with

Vahlne & Johanson (1977), we assume for now that offshoring experience with a particular

location reduces psychic distance and also reduces search and coordination costs. This effect

seems to be stronger and more relevant with regard to farshore than it is with regard to nearshore

investments. All these considerations can be summarized in the following propositions:

H4a (Firm-level effect): The more important labor cost savings are in a firm’s offshore

location decision, the less likely is a nearshore implementation.

H4b (Firm-level effect): The more important talent availability is in a firm’s offshore

location decision, the less likely is a nearshore implementation.

H4c (Firm-level effect): The more important geographical proximity is in a firm’s

offshore location decision, the more likely is a nearshore implementation.

H4d (Firm-level effect): The more important cultural proximity is in a firm’s offshore

location decision, the more likely is a nearshore implementation.

H4e (Firm-level effect): The more important language capability is in a firm’s offshore

location decision, the more likely is a nearshore implementation.

H4f (Firm-level effect): The more experience a company has with farshore locations, the

less likely is a nearshore implementation.

In the model developed below we further explore the effect of additional factors we treat as

controls, e.g. firm size, but we also explore the effect of additional firm-level drivers and risks in

15

extended models, e.g. managerial control and client acceptance. Previous research suggests that

these factors matter in location choice (e.g. Lewin & Couto, 2007). However, since they are not

core to the basic model and since they are expected to vary a lot by firm, we do not predict any

effects at this time. However, we will explore and discuss their effects later on and invite future

research to further investigate their impact on nearshore vs. farshore decisions.

Task-level effects: Finally, offshoring research indicates that location preferences very much

depend on characteristics of the task being offshored (e.g. Doh et al., 2009; Hutzschenreuter et

al., 2008). In particular the degree of commoditization of tasks has been stressed as an important

indicator not only for their general ‘offshorability’, i.e. their ‘separability’ from other tasks

(Blinder, 2006), but for the degree of coordination and local firm-specific investment needed to

provide these tasks offshore. High degree of commoditization means that knowledge about a task

is widely spread and diffused across companies and industries, and a potentially large number of

providers and captive centers is able to perform this task. Also, knowledge about performing this

task can be easily ‘transfer’ across locations and organizations. We therefore predict that the

more commoditized tasks are the less important becomes geographical proximity, hence the less

likely is a nearshore investment.

In addition, we recognize that, according to a number of studies (e.g. Dossani & Kenney,

2007; Patibandla & Petersen, 2002; Athreye et al., 2005), India has become the most important

provider of IT and software development services. China, in turn, is specializing primarily in

providing engineering and product development services (A. T. Kearney, 2004; Lewin & Couto,

2007). The concentration of IT, software and product development services in India and China is

also facilitated by the high availability of qualified scientists and engineers in these countries as

16

well as the high degree of digitalization of these services, in particular IT and software, which

eases the transfer and remote coordination of these tasks (e.g. Leonardi & Bailey, 2008). We

therefore predict that offshore investments related to IT, software and product development are

more likely to be made farshore than nearshore. However, we also explore to what extent

administrative functions, e.g. finance and accounting, call centers etc., are more likely to be

nearshored, but, based on previous research we cannot make predictions at this point. From a

theoretical perspective, we focus on the degree of standardization as a predictor for the

(decreasing) relevance of geographical proximity.

H5a (Task-level effect): The more commoditized a task is, the less likely is a nearshore

implementation.

H5b (Task-level effect): The more a task is related to IT, software and product

development, the less likely is a nearshore implementation.

In addition, we seek to explore whether the choice of delivery model (captive or outsourced) is

correlated with nearshore vs. farshore preference. In recent years, specialized external services

providers have become important players in the offshore space, both locally and internationally.

Today, service providers do not just provide IT and business process outsourcing services, but

increasingly develop expertise in product development and analytical services (Couto et al.,

2008). In a decision-making process, the choice of delivery model is often coupled with location

choice. However, previous literature or existing theory does not allow us to make any prediction

about this correlation. We therefore include delivery model merely as a control variable and

discuss our findings later on to promote further research on this aspect.

17

Data and Methodology

We test the effect of the proposed multi-level factors on companies’ decisions to locate business

functions to nearshore rather than farshore locations based on 1,598 offshore implementations by

275 U.S. firms (1,187 implementations) and 88 Dutch firms (411 implementations). The data is

drawn from the ORN database, which is based on an annual survey and which contains

comprehensive data on offshoring strategies, drivers, risks, outcomes and concrete offshore

implementations (both captive and outsourced) of currently 1,322 U.S. and European companies

of all sizes across industries and functions (Lewin & Couto, 2007). 48% of these companies are

actually offshoring, 17% are considering offshoring and 35% are not considering yet. In this

study we solely focus on companies that are currently offshoring. Industries in the ORN database

as well as in our subsample include e.g. manufacturing, software, finance and insurance, and

professional services. Functions offshored include IT, administrative services (e.g. HR, legal,

finance & accounting), call centers, software and product development, marketing and sales, and

procurement. Importantly, offshore implementations reported in the ORN database go back to

the 1980s and early 1990s. However, 99% of implementations reported were made after 1990;

the vast majority was launched in the last five years.

We selected the U.S. and the Dutch sample for this particular study (see Table 1). There

are several reasons this: First, the U.S. and the Netherlands are geographically positioned in

different parts of the world giving them different access to potential nearshore and farshore labor

markets attracting offshore investments. In this respect, the Netherlands represents a typical

Western European country whose companies engage in offshoring. The difference between U.S.

and Dutch (or other Western European) companies in terms of their geographical position makes

them interesting candidates for studying the role of distance in selecting offshore destinations.

18

Second, however, U.S. and Dutch companies are similar in terms of the distribution of functions

being offshored, years of offshoring experience, and delivery models selected (see Table 2-4). In

fact, Dutch companies on average are among the more experienced offshoring firms in Europe.

The similarity between U.S. and Dutch companies in these respects helps control for extraneous

variation (Eisenhardt, 1989). However, the samples do differ in terms of the distribution of firms

by size. The majority of Dutch companies are small and midsize, while the U.S. sample contains

a large amount of larger companies (see Table 5). Not least for this reason, we control for size in

all the regression models below. Third, the U.S. and the Dutch samples are the largest samples in

the ORN database. The high number of offshore implementations allows for a fine-grained

regression analysis and an analysis by subsamples (see in detail below).

-----------------------------

Insert Tables 1-5

-----------------------------

We use a logit model to examine the effect of various multi-level factors on the likelihood of

choosing nearshore vs. farshore locations (dependent variable). Similar approaches have been

taken to study the likelihood of external delivery models vs. captive models (e.g.

Hutzschenreuter et al., 2008) or the likelihood of offshoring product development rather than

other functions (e.g. Lewin et al., 2009). We focus on the likelihood of nearshoring. Nearshore

location is defined as Canada, Mexico and Central America for U.S. companies, and Eastern and

Western Europe for Dutch companies. For both Dutch and U.S. companies, India and China are

the most important farshore locations (see Figure 1). We test the effect of multi-level factors on

the likelihood of nearshoring in different subsamples. The most basic sample includes all U.S.

and Dutch implementations over time. Here we test the effect of multi-level factors (see below)

19

across time and countries of origin. In a second step, we divide the basic sample into two sub-

samples: the U.S. and the Dutch sample, as well as implementations before 2004 and since 2004.

This time split allows for a comparison of multi-level effects on nearshoring vs. farshoring

decisions over time. The year 2004 was chosen mainly for statistical reasons as it divides the

samples roughly into two equally large subsamples. However, it also allows us to separately

analyze dynamics in the most recent years. In a third step, we further divided the samples before

2004 and since 2004 by country which allows us to compare location dynamics by time and

country at the same time.

-----------------------------

Insert Tables 6,7

-----------------------------

Independent variables are listed and explained in Table 6. We test Hypothesis H1 (macro level:

increasing tendency to farshore) mainly by using a 2004 time split dummy (YEARSPLIT) in the

basic sample model. YEARSPLIT is 1 for all implementations since 2004, 0 for implementations

prior to 2004. However, indirectly we contribute to a better understanding of changing

preferences over time by looking at different effects in the subsamples (see below). For testing

H2a and H2b (country level), we, on the one hand, test the effect of national origin of the firm

(i_NL), on the other hand, we use two proxy variables to measure the effect of host country

economic conditions on their likelihood of being selected: COUNTRY_W – the delta between

the log of GDP/capita in the home and host country as a proxy for wage differential, and

COUNTRY_T – the log of the number of students enrolled in tertiary education as a proxy for

size of labor market. Similar measures have been applied by other scholars (e.g. Doh et al.,

2009). Hypothesis H3 (population-level effect) is measured with REL_EXP – the ratio of prior

20

nearshore vs. farshore implementations made across firms of the same national origin. A similar

measure for imitation effects has been used e.g. by Hutzschenreuter et al. (2008).

Hypotheses H4a through H4e (Firm-level location factors) are tested using the variables

LOC_TALENT (importance of talent availability), LOC_LABC (importance of labor costs),

LOC_ PROXG (importance of geographical proximity), LOC_PROXC (importance of cultural

proximity) and LOC_LANG (importance of language capability). All these variables are

standard Likert-scale variables (1 to 5) based on the following ORN survey question: “For each

implementation, please indicate your level of agreement with the following reasons for choosing

this particular country as an offshore location.” (1 – strongly disagree; 5 – strongly agree). In

extended regression models we included other location factors based on this question, e.g.

LOC_MARKET (market proximity), LOC_COLM (collocating with existing manufacturing)

and LOC_CUSTOMER (serving local customers). Another important location factor that is

intended to contribute to the explanatory power of country level effects is LOC_NHOT

(avoidance of hotspots). We included this factor because it indicates to what extent companies

choose a location because it promises to provide better access to qualified, yet lower-cost

personnel relative to other locations. H4f on the effect of firm-level experience is tested using

two simple dummy variables – FARPREV, NEARPREV – which are 1 if the respective

company has made previous farshore or nearshore investment respectively.

Hypotheses H5a and H5b (task-level factors) are tested in the following way: To test the

effect of task commoditization on the relevance of geographical distance (H5a), we use the

variable COMM_TODAY which measures the degree of commoditization of a function, e.g. IT.

This measure is generated from the ORN service provider survey (see also Couto et al., 2008). In

this survey, respondents are asked the following question: “In your opinion, for each class of

21

services that your company provides, how commoditized had this service become? (1 – very low,

5 – very high)”. We took the average response for each function (n= 280 responses) as a proxy

for the degree of commoditization of tasks. To test H5b (effect of software, IT, product

development function), we take SOFTPDIT as a dummy to indicate if the respective function

offshored belongs to the group of IT, software development or product development functions in

the ORN survey. Other surveys, e.g. Doh et al. (2009), use similar measures to characterize the

nature of tasks.

In addition, we use various control variables: EMP_HOME is used as an indicator for

size of the company. It contains the log of the number of employees working for the respective

company domestically. We realize that this size measure can be criticized in that it does not

account for trends towards temporary employment. Also, it does not count employees working at

outsourcers for the company. However, it is an indicator of the ‘resource base’ under

‘ownership’ control of the company. Also, it indicates to what extent a company uses overhead

implying the need for administrative efficiency in order to reduce costs (e.g. Lewin & Couto,

2007). Finally, the control variable i_OUT helps measure the impact of choice of the service

delivery model of preference for nearshore vs. farshore locations. i_OUT is 1 if an

implementation involves either a local, domestic or international service provider, where, in any

case, the task is performed offshore. i_OUT is 0 if the company established a wholly owned

captive unit to perform the task. In addition, we used function dummies (i_IT, i_CC, i_ADMIN

etc.) in certain models to indicate whether a particular function was being offshored.

22

Results

The findings confirm in a nutshell that geographical distance to potential offshore labor markets

matters in the selection of offshore locations (Tables 8-10). Dutch companies, in accordance with

Hypothesis H2a, are much more likely than U.S. companies to select nearshore locations close to

their home country (Model 1, 2, 3), because of their proximity to a larger nearshore labor market.

Availability of qualified personnel is in fact partly positively related to nearshore investments in

the Dutch sample (Model 14). However, the likelihood to offshore to remote locations increases

over time (Model 1), although only some models as well as the descriptive statistic of location

distribution confirm the significance of this trend both for U.S. and Dutch companies. In other

words, the macro-level Hypotheses H1 (tendency towards farshore) can be partially confirmed.

Country-level factors are tested in the extended Models 10-15. Results suggest that, as predicted,

low labor costs and availability of qualified personnel positively influence the preference for

farshore rather than nearshore locations. Hypotheses H2b and H2c can be confirmed. By

contrast, population tendency towards farshore (Hypotheses H3) does not positively affect the

preference for farshore decisions. In other words, population behavior does not have a significant

affect on firms’ location preferences. Hypotheses H3 cannot be confirmed.

-----------------------------

Insert Tables 8-11

-----------------------------

Rather, findings indicate that the firm-level importance of location factors as well as prior

farshoring experience have a significant influence on location decisions (Hypotheses H4a

through H4f). Most striking is the trade-off between the importance of availability of talent and

labor costs on the one hand and the importance of geographical proximity on the other hand (see

23

most models). Findings confirm that the more important availability of talent and low labor costs

are in an offshoring decision, the less likely are nearshore implementations (H4a, H4b). On the

other hand, the more important geographical proximity is, ceteris paribus, the more likely is a

nearshore implementation (H4c). Interestingly, cultural proximity and language capability do not

seem to influence nearshore decisions. Hypotheses H4d and H4e cannot be confirmed. On the

contrary, after 2003, companies, in particular from the U.S., even associate language capability

negatively with nearshore locations (Model 9). In other words, lack of language capability –

beside other reasons – can be even listed as a reason why many U.S. companies do not choose

nearshore locations, in particular Mexico and Central America. Cultural proximity only initially

seems to have supported nearshore investments – in the Dutch sample (Model 6). Later on, it did

not play a significant role.

Findings suggest however that firm-level experience with farshore destinations matters in

location decisions (H4f). Across companies, but especially in the Dutch sample, prior farshore

investments positively influence subsequent farshore investments. On the other hand, previous

nearshore investments do not seem to matter a lot. In other words, inexperienced companies are

more likely to invest nearshore than farshore. This finding supports the Uppsala model of

internationalization (Johanson & Vahlne, 1977), according companies, on the one hand, prefer to

invest in locations characterized by low psychic distance, on the other hand, learn how to reduce

psychic distance to further ‘remote’ locations over time. Interesting however are parallel changes

in the firm-level importance of other location factors. Findings suggest that for all companies,

but in particular for Dutch companies (see also Table 11), low labor costs has become

increasingly important as a location factor, which increases the likelihood of farshore

implementations. For U.S. companies, by contrast, reducing labor costs has always been an

24

important offshoring driver (Table 10, 11). Dutch companies, however, also regard availability

of labor as increasingly important - supporting farshoring (Model 7). Factors like proximity to

markets or supporting local customers used to be important for Dutch companies, but are not

anymore (see Table 11). Yet, these factors do not seem to influence nearshore vs. farshore

decisions significantly (see Model 10-12). Interestingly, for U.S. companies, avoiding hotspots

has become an important criterion for choosing nearshore locations (Model 9). For Dutch firms

this seems to be the opposite (Model 7), although the effect is not significant.

Task-level factors (Hypotheses H5a, H5b) play only a minor role in deciding about

nearshoring vs. farshoring, one exception being IT and procurement in favor of farshoring.

Overall, H5a and H5b cannot be confirmed. Quite interesting, however, is the effect of the

control variables firm size and service delivery model. In particular in the U.S. sample, larger

firms show an increasing tendency to nearshore (Model 3, 5, 9). At the same time, external

delivery models are positively related to nearshoring, in particular in the U.S. sample in more

recent years (Model 3, 9, 15). In the following we discuss some of these findings in greater

detail, paving the way for a range of potential follow up studies. We also list a number of

limitations that need to be addressed in the future.

Discussion

Empirical findings of this study suggest a co-evolutionary dynamic of offshoring location

choices. When U.S and Dutch companies started relocating business functions and processes in

the early 1990s, they faced different initial conditions in their location decisions resulting from

their geographical position. Dutch companies, on the one hand, had the option, in particular after

1990, to tap into a rather large nearshore labor market – first mainly Western Europe, then

25

Eastern Europe, thereby reconciling the benefits of geographical proximity with the opportunity

to hire lower cost talent. U.S. companies, on the other hand, were incented to explore farshore

locations, in particular India, right away because the size of the nearshore labor market

(compared to Western and Eastern Europe) was rather small. Also, most U.S. companies have

been strongly driven by the opportunity to lower labor costs from the very beginning, reinforcing

their orientation towards farshore locations, despite their geographical distance.

The U.S. experience in India arguably sparked an increasing interest in the media and in

consulting, promoting an offshoring agenda that until today stimulates a wide range of debates

and studies. Offshoring has become associated with the ‘global race for talent’ (Manning et al.,

2008) and the opportunity to increase efficiencies globally. This discursive development (see

also Metters & Verma, 2008) might have contributed to the significant increase of ‘awareness’ of

availability of talent and labor cost advantages as offshoring location factors among Dutch

companies. Moreover, Dutch companies are likely to have experienced tightening labor markets

in Eastern Europe. Case studies suggest that many European companies did not think ‘long-term’

when investing in Eastern European locations. Driven by their immediate opportunity to hire

lower cost people nearshore, they in fact contributed to increasing wage inflation and

competition for talent and ‘used up’ the nearshore labor market. Only then alternative options,

e.g. India, came to managerial attention. Another dynamic might be the enlargement of the EU in

2004. Many companies see this affecting opportunities to benefit from labor arbitrage effects.

Consequently, more Dutch firms are now offshoring to India and other offshore locations.

U.S. firms, however, seem to spread out to alternative options. At least two factors seem to play

a role here: As more European companies source talent and external expertise in India, problems

of wage inflation and employee turnover accelerate (e.g. Mayer-Ahuja & Feuerstein, 2007). This

26

might be one reason why more U.S. companies now consider nearshoring – not only because of

geographical proximity, but as an option to avoid hotspot locations. At the same time, however,

specialized, India- and China-based service providers have emerged over time who are

increasingly internationalizing their operations (Couto et al., 2008; Manning et al., 2008). In

order to better serve their longer term U.S. clients, many service providers have by now set up

operations close to the customer, either in the U.S. directly or in nearshore destinations. This has

arguably promoted recent decisions by U.S. companies to invest nearshore. ORN research

suggests that in particular large companies make increasing use of international – rather than

local – service providers. Plus, in particular Mexico and Central America are dominated today by

full-service providers, such as Accenture, IBM and Wipro.

In more abstract terms, findings suggest that sourcing strategies and offshoring location

choices by U.S. and Western European companies are systemically related to each other within a

fast changing global competitive environment. The global search and competition for (lower-

cost) talent is shaped by strategic responses of MNCs – as well as other players, e.g. external

service providers – to changing competitive conditions as well as unintended consequences of

actions at multiple systemic levels. Established theories of internationalization, e.g. the OLI

framework (Dunning, 1981), the Uppsala Model (Johanson & Vahlne, 1977), but also Porter’s

work (Porter, 1990), seem to have underestimated these dynamics. While Porter for example

would argue that MNCs can develop a competitive advantage based on their domestic experience

or – in this context – because of their initial geographical proximity (or distance) to offshore

labor markets, this argumentation falls short of recognizing the ongoing parallel, intersecting

trajectories at the firm, industry and national economy level (see e.g. Volberda & Lewin, 2003;

Lewin & Volberda, 1999). We propose, based on our findings, that future research on offshoring

27

(location) strategies should be more time-sensitive, e.g. by applying concepts like ‘window of

opportunity’ (see e.g. Tyre & Orlikowski, 1994).

Future studies also need to overcome limitations of this study. We focused solely on

companies from the U.S. and Netherlands. Although they are certainly important contributors to

the global offshoring dynamic, other firms might show other strategic patterns. For example,

more than U.S. and Dutch companies, UK and Spanish companies are expected to rely on

historical colonial relationships, e.g. with India and Latin America respectively. Also, we

abstracted away from the fact that in particular companies from the U.S. might differ in terms of

their headquarter proximity to potential offshore locations. Companies at the East coast, for

example, might regard Ireland as a nearshore destination, while Californian companies might be

more oriented towards China and Mexico from the very beginning. Also, as mentioned earlier,

company size might make a significant difference in location choice patterns. Since many U.S.

companies are larger than most Dutch companies, this might influence nearshore vs. farshore

preferences. Finally, one current limitation of ORN data is the lack of differentiation between

‘contact location’ and ‘task location’. In particular the rise of international service providers is

likely to change location choice patterns significantly which, so far, is only partly reflected in

ORN data.

In addition, our findings may stimulate various follow-up research questions explorating

the dynamic nature of internalization in general and offshoring location choices in particular.

Possible follow-up research questions could include to what extent companies observe

competitive responses, and how they, in turn, respond to these observations? In other words,

examining a moderating effect of absorptive capacity on offshoring location choices could add to

the explanatory power of the model. Other interesting research venues might include the role

28

nation states play. To what extent for example are domestic and nearshore policy-makers aware

of these dynamics and how do they respond? And how does migration of talent (see e.g. OECD,

2008) affect location choice patterns? How is talent different, in this respect, from other ‘location

factors’? We suggest, based on our exploratory study, that a co-evolutionary multi-level

framework should be used to better understand these dynamic phenomena.

References

A.T.Kearney. 2004. Making offshore decisions: A.T.Kearney's 2004 offshore location

attractiveness index, Chicago et al.

Athreye SS. 2005. The Indian software industry and its evolving service capability. Industrial

and Corporate Change 14: 393-418

Blinder AS. 2006. Offshoring: the next industrial revolution? Foreign Affairs 85: 113-28.

Bresnahan T, Gambardella A, Saxenian A. 2001. 'Old economy' inputs for 'new economy'

outcomes: cluster formation in the new Silicon Valleys. Industrial and Corporate Change

10: 835-60.

Bunyaratavej K, Hahn ED, Doh JP. 2007. International offshoring of services: a parity study.

Journal of International Management 13: 7-21.

Child J. 1972. Organizational structure, environment and performance: The role of strategic

choice. Sociology 6: 1-22

Couto V, Mani M, Sehgal V, Lewin AY, Manning S, Russell JW. 2008. Offshoring 2.0:

Contracting knowledge and innovation to expand global capabilities, 2007 Service

Provider Survey Report by Booz & Co. & Duke University.

Doh JP. 2005. Offshore outsourcing: implications for international business and strategic

management theory and practice. Journal of Management Studies 42: 695-704

Doh J, Bunyaratavej K, Hahn ED. 2009. Separable but not equal: the location determinants of

discrete services offshoring activities. Journal of International Business Studies

Forthcoming.

Dossani R, Kenney M. 2007. The next wave of globalization: relocating service provision to

India. World Development 35: 772-91.

29

Eisenhardt KM. 1989. Building theories from case study research. Academy of Management

Review 14: 532-50

Dunning, JH 1980. Toward an Eclectic Theory of International Production: Some Empirical

Tests. Journal of International Business Studies 11: 9-31.

Freeman R. 2006. Does globalization of the scientific/engineering workforce threaten U.S.

economic leadership? NBER Innovation Policy & the Economy Working Paper.

Gupta A, Seshasai S, Mukherji S, Ganguly A. 2007. Offshoring: The transition from economic

drivers toward strategic global partnership and 24-hour knowledge factory. Journal of

Economic Commerce in Organizations 5: 1-23.

Hitt MA, Beamish PW, Jackson SE, Mathieu JE. 2007. Building theoretical and empirical

bridges across levels: multilevel research in management. Academy of Management

Journal 50: 1385-99.

Hofstede G. 1980. Culture's Consequences: International Differences in Work-related Values.

Beverly Hills: Sage Publications.

Huang Y, Khanna T. 2003. Can India overtake China? Foreign Policy 137 (July/August): 74-81.

Hutzschenreuter, T, Lewin, AY, Dresel, S. 2008. Governance mode for offshoring of support

functions: Comparison of US and German companies. Paper presented at the 2008

Academy of International Business Conference.

Johanson J, Vahlne J-E. 1977. The internationalization process of the firm - a model of

knowledge development and increasing foreign market commitments. Journal of

International Business Studies 8: 23-32.

Kenney M, Massini S, Murtha TP. 2009. Offshoring administrative and technical work: do we

need a new theory for this? Journal of International Business Studies Forthcoming.

Kogut B, Singh H. 1988. The effect of national culture on the choice of entry mode. Journal of

International Business Studies 19: 411-32.

Leonardi PM, Bailey DE. 2008. Transformational technologies and the creation of new work

practices: Making implicit knowledge explicit in task-based offshoring. MIS Quarterly

32: 411-36.

Levy D. 2005. Offshoring in the new political economy. Journal of Management Studies 42:

685-693.

30

Lewin AY, Couto V. 2007. Next Generation Offshoring: The Globalization of Innovation, Duke

University CIBER/Booz Allen Hamilton Report, Durham.

Lewin AY, Massini S, Peeters C. 2009. Why are companies offshoring innovation? The

emerging global race for talent. Journal of International Business Studies Forthcoming.

Lewin AY, Volberda HW. 1999. Prolegomena on coevolution: A framework for research on

strategy and new organizational forms. Organization Science 10: 519-34.

Manning S. 2008. Customizing clusters: On the role of multinational corporations in the

formation of science and engineering clusters in emerging economies. Economic

Development Quarterly 22: 316-23.

Manning S, Massini S, Lewin AY. 2008. A dynamic perspective on next-generation offshoring:

the global sourcing of science & engineering skills. Academy of Management

Perspectives 22: 35-54.

Mayer-Ahuja N, Feuerstein P. 2007. IT-labour goes offshore: regulating and managing attrition

in Bangalore. SOFI Working Paper.

Metters R, Verma R. 2008. History of offshoring knowledge services. Journal of Operations

Management 26: 141-7

O'Leary MB, Cummings NJ. 2007. The spatial, temporal, and configurational characteristics of

geographic dispersion in teams. MIS Quarterly 31: 433-52.

OECD. 2008. The Global Competition for Talent: Mobility of the Highly Skilled. Paris: OECD

Report.

Oliver C. 1991. Strategic responses to institutional processes. Academy of Management Review

16: 145-79.

Patibandla M, Petersen B. 2002. Role of transnational corporations in the evolution of a high-

tech industry: the case of India’s software industry. World Development 30: 1561-77

Porter ME. 1990. The competitive advantage of nations. New York: Free Press.

Rugman AM, Verbeke A. 2004. A perspective on regional and global strategies of multinational

enterprises. Journal of International Business Studies 35: 3-18

Scott WR. 2001. Institutions and organizations. Thousand Oaks: Sage.

Tyre MJ, Orlikowski WJ. 1994. Windows of opportunity: temporal patterns of technological

adaptation in organizations. Organization Science 5: 98-118

31

Shenkar O, Luo Y, Yeheskel O. 2008. From "distance" to "friction": substituting metaphors and

redirecting intercultural research. Academy of Management Review 33: 905-23.

Volberda HW, Lewin AY. 2003. Co-evolutionary dynamics within and between firms: from

evolution to co-evolution. Journal of Management Studies 40: 2111-36

Xu D, Shenkar O. 2002. Institutional distance and the multinational enterprise. Academy of

Management Review 27: 608-18.

32

Tables and Figures Figure 1: Distribution of location choices by national origin of firm

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

BE DE NL ESP UK USA

( p )

Source: Duke University/Archstone Consulting Offshoring Research Network 2005 Survey and Duke University/Booz Allen Hamilton Offshoring Research Network 2006 Survey and Duke University/The Conference Board Offshoring Research Network 2007/8 Survey

Per

cent

of t

otal

num

ber

of im

plem

enta

tions

N= 193 190

OtherOther AsiaPhilippines

ChinaIndia

Latin AmericaMexico

CanadaEastern EuropeWestern Europe

411 174 103 1187

Figure 2: Distribution of offshore implementations in India by national origin of firm*

11

0%

10%

20%

30%

40%

50%

60%

Pre-2002 2002-2004 2005-2007

SpainBelgiumGermanyScandUKNLUS

Source: Duke University/Archstone Consulting Offshoring Research Network 2005 Survey and Duke University/Booz Allen Hamilton Offshoring Research Network 2006 Survey and Duke University/The Conference Board Offshoring Research Network 2007/8 Survey

Perc

ent o

f tot

al n

umbe

r of

impl

emen

tatio

ns

Year in which offshore implementation was made

*n=776 implementations in ORN database

33

Table 1: Offshoring Status of U.S. and Dutch Companies in ORN Sample

No. of firms

No. of implementati

ons No. of firms

No. of implementa

tions No. of firms

No. of implementatio

nsCurrently offshoring 301 1,271 102 453 403 1,724Considering offshoring 53 182 54 108 107 290Not considering offshoring 73 73 208 208 281 281

Total 427 1,526 364 769 791 2,295

US Netherlands Total

Table 2: Distribution of Functional Implementations in U.S. and Dutch Sample

US Netherlands TotalProduct Development 22% 20% 21%Software Development 20% 16% 18%Administrative Services 17% 13% 15%Information Technology 16% 14% 16%Call Centers 14% 14% 14%Marketing & Sales 5% 13% 8%Procurement 5% 10% 7%Total 100% 100% 100% Table 3: Distribution of Service Delivery Models in U.S. and Dutch Sample

US Netherlands TotalCaptive 46% 58% 49%Outsourced 54% 42% 51%Total 100% 100% 100% Table 4: Distribution of Location Choices Over Time in U.S. and Dutch Sample*

Total Nearshore % Total Nearshore % Total Nearshore %US 373 45 12% 461 38 8% 1,187 135 11%Netherlands 124 48 39% 119 35 29% 411 130 32%Total 497 93 19% 580 73 13% 1,598 265 17%

Total SamplePre-2004 2004-present

*Some firms excluded because launch year data not available

34

Table 5: Distribution of Firms by Size in U.S. and Dutch Sample*

Small (<500

employees)

Midsize (500-

20,000 employe

es)

Large (>20,000

employees) TotalNo. of firms % of total

No. of firms % of total No. of firms % of total

No. of firms % of total

US 69 28% 90 36% 88 36% 247 100%Netherlands 49 67% 24 33% 0 0% 73 100%Total 118 37% 114 36% 88 28% 320 100% *Some firms excluded because size data not available Table 6: List of Variables

Dependent variable Explanation / Hypothesis Format NEARSHORE Nearshore implementation Dutch: 1 for Western or

Eastern Europe; 0 otherwise U.S.: 1 for Canada, Mexico or Central America; 0 otherwise

Independent variables YEARSPLIT (After2003) H1: Implementation after 2003

(50% split ; sample separation by time)

1;0

i_NL H2a: Dutch company (also used for sample split)

1;0

COUNTRY_T H2b: Number of students in tertiary education system

Log(number)

COUNTRY_W H2b: Wage differential measured by log(GDP/Capita home country) – log(GDP/Capita host country)

Number

REL_EXP H3: Population location choice trends (nearshore/farshore)

No. of past farshore / nearshore implementations

LOC_TALENT, LOC_LABC, LOC_PROXG, LOC_LANG, LOC_PROXC

H4: Core location factors (e.g. importance of geographical prox., cultural prox., language comp., labor costs, talent)

5-Point Likert scale (1… not important at all; 5… very important)

LOC_MARKET, LOC_COLMAN, LOC_CUSTOMER, LOC_NHOTSPOT

H4: Additional location factors (e.g. importance of market prox., co-location with manufacturing, serving local customers, avoiding hotspots)

5-Point Likert scale (1… not important at all; 5… very important)

FARPREV, NEARPREV H4f: Firm-level experience with prior farshore, nearshore investments

1;0

COMM_TODAY (COMM_FUTURE)

H5a: Level of commoditization of task (today and in the future)

Based on 5-Point Likert scale (1…very low; 5… very high)

i_SOFTPDIT H5b: IT, Software or Product development function

1;0

i_IT, i_CC, i_ADMIN, i_PROC, i_MS

Control: Function Dummies (IT, Administrative Services, Procurement, Marketing & Sales)

1;0

EMP_HOME Control: Number of employees domestically (size proxy)

Log (number)

i_OUT Control: Dummy for outsourced (rather than captive) operation (service delivery model)

1;0

35

Table 7: Correlations between independent variables

Variable Obs Mean Std. Dev 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 181 loc_talent 1169 3.69 1.03 12 loc_labcost 1194 4.03 1.07 0.17* 13 loc_proxgeog 1147 2.89 1.33 0.1* -0.06 14 loc_lang 1389 3.13 1.32 0.35* 0.13* 0.37* 15 loc_proxcult 1156 2.99 1.17 0.22* -0.02 0.65* 0.53* 16 loc_nhotspot 1140 2.76 1.08 0.3* 0.12* 0.41* 0.35* 0.41* 17 loc_market 1134 3.01 1.36 0.15* -0.07 0.39* 0.24* 0.35* 0.39* 18 loc_colman 1363 2.70 1.43 0.07 -0.01 0.48* 0.26* 0.35* 0.28* 0.45* 19 loc_customer 1134 2.91 1.36 0.15* -0.1* 0.45* 0.31* 0.39* 0.36* 0.79* 0.55* 1

10 loc_provloc 1372 2.87 1.31 0.44* -0.03* 0.25* 0.29* 0.33* 0.34* 0.26* 0.26* 0.31* 111 i_Out 1399 0.54 0.50 0.04 0.11* -0.11* 0 -0.09* -0.01 -0.24* -0.14* -0.19* 0.23* 112 emp_home 1767 7.29 3.17 0.13* 0.13* 0.12* 0.12* 0.11* 0.06 -0.05 0.04 -0.02 0.07 -0.06 113 country_w 1802 2.60 1.42 0.12* 0.44* -0.28* 0.07 -0.25* -0.06 -0.21* -0.09* -0.24* -0.04 0.1* 0.06 114 country_t 1775 15.20 1.51 0.19* 0.23* -0.21* -0.02 -0.21* -0.09* -0.1* -0.05 -0.12* -0.01 0.04 0.06 0.61* 115 comm_today 1819 3.11 0.46 0.03 0.1* 0.14* 0.14* 0.18* 0.14* 0 0 0.03 0.06 0.1* 0.21* -0.02 -0.15* 116 comm_future 1819 3.53 0.45 0.02 0.09* 0.15* 0.16* 0.19* 0.14* 0.01 0.02 0.04 0.08* 0.11* 0.22* -0.02 -0.15* 0.97* 117 farprev 1081 0.51 0.50 -0.08 0.02 0.01 0.02 -0.04 0.02 0.01 0 0 0 -0.01 0.21* -0.02 -0.04 0 0.01 118 nearprev 1081 0.18 0.38 -0.13* -0.18* 0.08 -0.01 0.01 0.03 0.13* 0.05 0.12* 0.01 -0.01 -0.06 -0.24* -0.23* 0.05 0.06 0.26* 1

Significance level: * <1%

36

Table 8: Regression models (Dependent variable: Likelihood of Nearshoring)

1 2 3 4 5 6 7 8 9Independent Variables Coded Name Total Total Total NL US NL NL US US

Pre-2004 2004+ Pre-2004 2004+ Pre-2004 2004+# Obs 584 263.00 321.00 143.00 441.00 70.00 51.00 177.00 247.00LR Chi-2 215.23 102.58 128.26 69.31 93.20 41.95 30.74 45.23 56.89Prob > Chi2 0.0000 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00Log Likelihood -169.009 -91.151 -65.970 -62.916 -98.729 -26.513 -19.181 -47.631 -36.023R-Squared 0.3890 0.3601 0.4929 0.3552 0.3207 0.4417 0.4449 0.3220 0.4412

Constant -3.550*** (1.238) -4.678*** (1.816) -6.250*** (2.239) 1.977 (1.664) -6.190*** (1.964) 0.121 (2.834) 2.830 (2.736) -6.517** (2.780) -8.301*** (3.142)

SubsamplingSurvey Country i_NL 3.118*** (0.449) 3.259*** (0.704) 4.287*** (0.871) n/a n/a n/a n/a n/a n/aAfter 2003 yearsplit03 -0.478* (0.291) n/a n/a -0.453 (0.510) -0.596+ (0.387) n/a n/a n/a n/a

Location FactorsTalent Availability loc_talent -0.587*** (0.170) -0.564** (0.239) -0.512** (0.260) -0.514* (0.278) -0.719*** (0.245) -0.629 (0.478) -0.943+ (0.608) -1.332*** (0.384) -0.238 (0.390)Low Labor Cost loc_labc -0.449*** (0.132) -0.198 (0.205) -0.808*** (0.224) -0.459** (0.214) -0.327+ (0.203) -0.424 (0.404) -1.086** (0.432) -0.003 (0.321) -0.696* (0.364)Geo Proximity loc_proxg 1.133*** (0.166) 1.185*** (0.245) 1.385*** (0.296) 1.167*** (0.257) 1.286*** (0.298) 1.328*** (0.428) 0.874* (0.475) 1.436*** (0.487) 1.647*** (0.526)MatcLanguage loc_lang 0.049 (0.159) 0.156 (0.228) -0.414* (0.251) -0.137 (0.263) 0.191 (0.229) -0.687 (0.511) 0.547 (0.539) 0.733* (0.387) -0.724* (0.382)Cultural Proximity loc_proxc 0.040 (0.174) 0.234 (0.238) -0.136 (0.309) -0.097 (0.276) 0.158 (0.278) 0.719* (0.433) 0.570 (0.678) 0.280 (0.418) -0.083 (0.466)Avoiding Hotspots loc_nhot 0.036 (0.162) 0.038 (0.229) 0.238 (0.286) -0.297 (0.303) 0.180 (0.218) 0.391 (0.654) -0.593 (0.539) -0.093 (0.304) 0.729+ (0.464)

Company FactorsService Delivery Model i_Out 0.366 (0.313) 0.004 (0.426) 0.896+ (0.567) -0.200 (0.548) 0.466 (0.435) 0.432 (1.039) 1.297 (1.148) 0.261 (0.566) 1.460+ (0.968)Size of Firm emp_home 0.132** (0.064) 0.080 (0.084) 0.354*** (0.132) 0.027 (0.157) 0.175** (0.077) -0.466 (0.356) -0.059 (0.332) 0.063 (0.092) 0.335** (0.154)

Function VariablesIT i_IT -0.677 (0.601) -1.401* (0.828) 0.620 (1.008) -1.763+ (1.121) -0.255 (0.857)Call Center i_CC -0.070 (0.586) -0.649 (0.792) 1.410 (1.018) -0.238 (0.873) 0.051 (0.884)Product Dev or Softwarei_PDSoft -0.121 (0.550) -0.699 (0.744) 1.377 (0.981) -0.044 (0.724) -0.067 (0.888)Administrative i_Admin -0.454 (0.657) -1.278 (0.922) 1.516 (1.154) 1.264 (1.499) -0.551 (0.904)Procurement i_Proc -1.046 (0.735) -1.277 (0.936) -0.012 (1.369) -2.195* (1.171) -0.142 (1.040)Marketing Services i_MS -0.252 (0.677) -1.503* (0.885) 2.744** (1.303) -0.512 (0.911) -0.705 (1.366)Software or PD or IT SoftPDIT 0.690 (0.963) -1.542 (1.430) -0.166 (0.585) -0.422 (0.682)

Standard Errors in brackets. Significance levels: *** <1%, ** < 5%, * < 10%, +<15%.

37

Table 9: Extended regression models (Dependent variable: Likelihood of Nearshoring)

10 11 12 13 14 15Independent Variables Coded Name Total NL US Total NL US

# Obs 496.00 103.00 393.00 526.00 105.00 421.00LR Chi-2 274.17 96.47 137.21 272.07 115.02 130.33Prob > Chi2 0.00 0.00 0.00 0.00 0.00 0.00Log Likelihood -88.526 -22.337 -58.491818 -101.494 -14.465 -71.477R-Squared 0.6076 0.6835 0.5398 0.5727 0.7990 0.4769

Constant 9.76*** (3.22) 23.245*** (8.97) 5.208 (3.962) -4.412 (5.381) -36.352** (17.636) 4.937 (8.293)

SubsamplingSurvey Country i_NL 2.75*** (0.63) n/a n/a 3.355*** (0.623) n/a n/aAfter 2003 yearsplit03 0.28 (0.44) 1.587 (1.36) 0.024 (0.545) 0.285 (0.409) 3.837** (1.874) -0.169 (0.498)

Location FactorsTalent Availability loc_talent -0.76*** (0.28) -1.097 (0.78) -0.728* (0.380) -0.428* (0.231) -2.385 *(1.302) -0.226 (0.294)Low Labor Cost loc_labc -0.02 (0.21) -0.016 (0.50) 0.121 (0.281) -0.355** (0.173) -1.958** (0.892) -0.235 (0.242)Geo Proximity loc_proxg 1.37*** (0.25) 2.220*** (0.85) 1.491*** (0.380) 1.325*** (0.222) 4.063*** (1.499) 1.402*** (0.329)MatcLanguage loc_lang -0.14 (0.24) -0.724 (0.70) -0.021 (0.323) 0.082 (0.214) -2.085* (1.180) 0.216 (0.278)Cultural Proximity loc_proxc -0.09 (0.28) 0.283 (0.61) -0.112 (0.384) -0.135 (0.238) -0.745 (0.985) -0.044 (0.334)Avoiding Hotspots loc_nhot -0.04 (0.24) -0.822 (0.86) 0.056 (0.297) -0.048 (0.215) 1.922 (1.462) -0.013 (0.255)Market Proximity loc_market 0.13 (0.25) 0.385 (0.73) -0.027 (0.339)Collocating with existing Maloc_colm -0.16 (0.18) -0.575 (0.52) -0.090 (0.225)Customers loc_customer -0.18 (0.24) 0.305 (0.72) -0.183 (0.357)Best Provider Location loc_provloc 0.29 (0.22) -0.156 (0.54) 0.471+ (0.297)

Company FactorsService Delivery Model i_Out 0.41 (0.48) 2.067 (1.54) 0.093 (0.096) 0.767* (0.426) 1.455 (1.386) 0.843+ (0.527)Size of Firm emp_home 0.11 (0.09) 0.487 (0.37) 0.372 (0.615) 0.155** (0.074) 0.648 (0.506) 0.144 (0.083)

Market FactorsLocation Wage Level country_w -0.80*** (0.22) -1.534** (0.61) -0.699*** (0.263) -1.043** (0.409) -7.068** (2.893) -0.043 (0.580)Advanced Education Enrollmcountry_t -0.74*** (0.15) -0.948*** (0.34) -0.728*** (0.186) 0.185 (0.447) 4.300** (1.942) -0.877 (0.699)Working Population country_l

ExperienceRelative Exp rel_expPrev Farshore (dummy) farprev -1.333*** (0.459) -3.669** (1.482) -0.679 (0.567)Prev Nearshore (dummy) nearprev 0.572 (0.472) 0.613 (1.563) 0.198 (0.595)

Function VariablesSoftware or PD or IT SoftPDIT -0.09 (0.48) -0.625 (1.61) 0.297 (0.567) 0.340 (0.367) 0.331 (1.703) 0.498 (0.490)

Commoditization of Function (Today) 2.68 (2.16) 10.893+ (7.32) -0.065 (2.698)Commoditization of Function (Future) -3.07 (2.30) -12.668+ (8.03) -0.024 (2.812)

Standard Errors in brackets. Significance levels: *** <1%, ** < 5%, * < 10%, +<15%.

38

Table 10: Strategic Offshoring Drivers of U.S. and Dutch Companies

US Netherlands TotalLabor cost savings 91% 61% 82%Growth strategy 70% 78% 72%Other cost savings 68% 68% 68%Access to qualified personnel 69% 57% 65%Part of a larger global strategy 58% 74% 63%Competitive pressure 60% 52% 58%Increasing speed to market 48% 58% 51%Business process redesign 47% 54% 49%Improved service levels 45% 49% 46%Exploit country-specific advantages 40% 52% 45%Exploit location-specific advantages 38% 51% 43%Domestic shortage of qualified personnel 31% 51% 40%Accepted industry practice 38% 30% 36%Access to new markets 24% 42% 29%Differentiation strategy 27% 30% 28%Enhancing system redundancy 25% 23% 24% Percentage of respondents rating driver as “important” or “very important” (4 or 5 on Likert Scale) Table 11: Location Selection Factors of U.S. and Dutch Companies

Pre-2004

2004-present Total Pre-2004

2004-present Total

Low cost of labor 84% 80% 81% 47% 69% 59%High level of expertise 78% 72% 75% 58% 64% 59%Talent pool available 71% 69% 69% 58% 60% 54%Low costs (besides labor costs) 73% 62% 66% 47% 65% 56%Matches language requirements 62% 55% 57% 31% 33% 31%Access to local market 52% 39% 44% 59% 31% 45%Collocating with existing BP facility offsho 64% 49% 51% 25% 20% 22%Supporting existing customers locally 53% 39% 44% 56% 29% 40%Collocating with existing manufacturing p 48% 38% 43% 30% 26% 27%Geographical proximity 56% 41% 45% 24% 25% 25%Location of the best service provider 36% 45% 40% 40% 41% 37%Cultural proximity 47% 41% 41% 41% 23% 32%Quality of infrastructure 32% 37% 34% 37% 37% 35%Political stability in host country 40% 32% 33% 38% 25% 31%Government incentives 35% 30% 32% 28% 30% 27%Avoiding "hot spots" 35% 29% 30% 15% 18% 16%Other 28% 30% 29% 1% 8% 5%

US Netherlands

Percentage of respondents rating driver as “important” or “very important” (4 or 5 on Likert Scale)