ArticlePDF Available

Resource, Strategies, Location Determinants, And Host Country Location Choice By Emerging Market Firms

Article

Resource, Strategies, Location Determinants, And Host Country Location Choice By Emerging Market Firms

Abstract and Figures

The extant literature had studied the determinants of the firms’ location decisions with help of host country characteristics and distances between home and host countries. Firm resources and its internationalization strategies had found limited attention in this literature. To address this gap, the research question in this dissertation was whether and how firms’ resources and internationalization strategies impacted the international location decisions of emerging market firms. To explore the research question, data were hand-collected from Indian software firms on their location decisions taken between April 2000 and March 2009. To analyze the multi-level longitudinal dataset, hierarchical linear modeling was used. The results showed that the internationalization strategies, namely market-seeking or labor-seeking had direct impact on firms’ location decision. This direct relationship was moderated by firm resource which, in case of Indian software firms, was the appraisal at CMMI level-5. Indian software firms located in developed countries with a market-seeking strategy and in emerging markets with a labor-seeking strategy. However, software firms with resource such as CMMI level-5 appraisal, when in a labor-seeking mode, were more likely to locate in a developed country over emerging market than firms without the appraisal. Software firms with CMMI level-5 appraisal, when in market-seeking mode, were more likely to locate in a developed country over an emerging market than firms without the appraisal. It was concluded that the internationalization strategies and resources of companies predicted their location choices, over and above the variables studied in the theoretical field of location determinants.
Content may be subject to copyright.
Florida International University
Digital Commons @ FIU
FIU Electronic Theses and Dissertations
12-9-2009
Resource, Strategies, Location Determinants, And
Host Country Location Choice By Emerging
Market Firms
Naveen K. Jain
Florida International University, njain001@fiu.edu
This document is brought to you for free and open access by Digital Commons @ FIU. It has been accepted for inclusion in FIU Electronic Theses
and Dissertations by an authorized administrator of Digital Commons @ FIU. For more information, please contact dcc@fiu.edu.
Recommended Citation
Jain, Naveen K., "Resource, Strategies, Location Determinants, And Host Country Location Choice By Emerging Market Firms"
(2009). FIU Electronic Theses and Dissertations. Paper 147.
http://digitalcommons.fiu.edu/etd/147
FLORIDA INTERNATIONAL UNIVERSITY
Miami, Florida
RESOURCES, STRATEGIES, LOCATION DETERMINANTS, AND HOST COUNTRY LOCATION
CHOICE BY EMERGING MARKET FIRMS
A dissertation submitted in partial fulfillment of the
requirements for the degree of
DOCTOR OF PHILOSOPHY
in
BUSINESS ADMINISTRATION
by
Naveen Kumar Jain
2010
ii
To: Dean Joyce Elam
College of Business Administration
This dissertation, written by Naveen Kumar Jain, and entitled Resources, Strategies, Location
Determinants, and Host Country Location Choice by Emerging Market Firms, having been approved in
respect to style and intellectual content, is referred to you for judgment.
We have read this dissertation and recommend that it be approved.
____________________________________
Juan Sanchez
____________________________________
William Newburry
____________________________________
Kaushik Dutta
___________________________________
Torben Pedersen
___________________________________
Sumit K. Kundu, Major Professor
Date of Defense: December 9, 2009
This dissertation of Naveen Kumar Jain is approved.
__________________________________
Dean Joyce Elam
College of Business Administration
__________________________________
Dean George Walker
University Graduate School
Florida International University, 2010
iii
DEDICATION
I dedicate this thesis to my parents, Lal Chand Jain and Sheela Jain; my sister Niti Jain; and my wife,
Shefali Jain. Without their patience, understanding, support, and most of all love, the completion of this
work would not have been possible.
iv
ACKNOWLEDGMENTS
I wish to thank the members of my committee for their support, patience, and good humor. Their gentle but
firm direction has been most appreciated. Dr. Juan Sanchez, Dr. William Newburry, Dr. Torben Pedersen
were particularly helpful in guiding me towards theory-building and qualitative methodology. Dr. Kaushik
Dutta provided me software industry specific guidance that helped me in data collection. Dr. James Jaccard
helped me with methodology. Finally, I would like to thank my major professor, Dr. Sumit K. Kundu.
From the beginning, he had confidence in my abilities to not only complete a degree, but to complete it
with excellence.
I have found my coursework throughout the Curriculum and Instruction program to be stimulating
and thoughtful, providing me with the tools with which to explore both past and present ideas and issues.
v
ABSTRACT OF THE DISSERTATION
RESOURCES, STRATEGIES, LOCATION DETERMINANTS,
AND HOST COUNTRY LOCATION CHOICE BY EMERGING MARKET FIRMS
by
Naveen Kumar Jain
Florida International University, 2010
Miami, Florida
Professor Sumit K. Kundu, Major Professor
The extant literature had studied the determinants of the firms’ location decisions with help of host country
characteristics and distances between home and host countries. Firm resources and its internationalization
strategies had found limited attention in this literature. To address this gap, the research question in this
dissertation was whether and how firms’ resources and internationalization strategies impacted the
international location decisions of emerging market firms.
To explore the research question, data were hand-collected from Indian software firms on their
location decisions taken between April 2000 and March 2009. To analyze the multi-level longitudinal
dataset, hierarchical linear modeling was used. The results showed that the internationalization strategies,
namely market-seeking or labor-seeking had direct impact on firms’ location decision. This direct
relationship was moderated by firm resource which, in case of Indian software firms, was the appraisal at
CMMI level-5. Indian software firms located in developed countries with a market-seeking strategy and in
emerging markets with a labor-seeking strategy. However, software firms with resource such as CMMI
level-5 appraisal, when in a labor-seeking mode, were more likely to locate in a developed country over
emerging market than firms without the appraisal. Software firms with CMMI level-5 appraisal, when in
market-seeking mode, were more likely to locate in a developed country over an emerging market than
firms without the appraisal.
It was concluded that the internationalization strategies and resources of companies predicted their
location choices, over and above the variables studied in the theoretical field of location determinants.
vi
TABLE OF CONTENTS
CHAPTER PAGE
I. INTRODUCTION ............................................................................................................................1
II. LITERATURE REVIEW: EMERGING MARKETS ....................................................................10
III. LITERATURE REVIEW: RESOURCE-BASED VIEW...............................................................17
IV. LITERATURE REVIEW: INTERNATIONALIZATION STRATEGIES ....................................29
V. LITERATURE REVIEW: DETERMINANTS OF HOST COUNTRY
LOCATION CHOICE ....................................................................................................................34
VI. MODEL AND HYPOTHESES ......................................................................................................50
Model..............................................................................................................................................51
Internationalization Strategies and Location Choice of Indian Software Firms..............................53
Electronic Knowledge Sharing Database as a Resource and Location Choice...............................56
Low Cost of Software Production as a Resource and Location Choice..........................................58
Process Improvement Implementation as a Resource and Location Choice...................................60
VII. VARIABLES AND DATA COLLECTION...................................................................................62
VIII. METHODOLOGY, RESULT, AND DISCUSSION......................................................................84
Result ..............................................................................................................................................96
Discussion.....................................................................................................................................109
IX. LIMITATIONS AND FUTURE RESEARCH.............................................................................113
X. CONCLUSION............................................................................................................................. 116
LIST OF REFERENCES ............................................................................................................................118
APPENDIX.................................................................................................................................................142
VITA ...........................................................................................................................................................145
vii
LIST OF TABLES
TABLE PAGE
1. Resource Types in Extant Literature ...............................................................................................19
2. Research on Location Determinants ............................................................................................... 34
3. Variables related to Internationalization Strategies......................................................................... 64
4. Macroeconomic Determinants ........................................................................................................67
5. Institutions.......................................................................................................................................69
6. Cultural Distance ............................................................................................................................72
7. International Experience and TMT ................................................................................................. 76
8. Data-source for Dependent and Independent Variables..................................................................77
9. Data-source for Firm-specific Control Variables............................................................................76
10. Data-source for Host Country-specific Control Variables..............................................................76
11. Level of Firm-specific Data Collection...........................................................................................78
12. Level of Host Country-specific Data Collection.............................................................................81
13. Correlation Table ............................................................................................................................86
14. Collinearity Statistics ......................................................................................................................90
15. List of Variables in HLM................................................................................................................92
16. Descriptive Statistics....................................................................................................................... 93
17. Correlation Table of Variables in HLM.......................................................................................... 94
18. Collinearity Statistics of Variables in HLM....................................................................................95
19. Level 1 Control Variables...............................................................................................................97
20. Selected Level 1 Control Variables.................................................................................................98
21. Level 2 Control Variables.............................................................................................................100
22. HLM Results.................................................................................................................................105
23. Significant Hypotheses .................................................................................................................107
1
CHAPTER I
INTRODUCTION
As multinational firms engage in two-way knowledge flows between parents and subsidiaries in order to
enhance their competitive advantage, subsidiary location choice is gaining managerial attention (Dunning,
1998, 2009). Multinational firms no longer consider a location merely as a financial risk reduction tool
(Dunning, 2009; Rugman, 1979), but use it as an asset sourcing means capable of enhancing their
competitive advantage by creating a coherent set of locations that fit well with each other (Dunning, 2009;
Ghemawat, 2001; Ricart, Enright, Ghemawat, Hart, & Khanna, 2004). Location, as a factor, affects
multinationals’ market valuations (Pantzalis, 2001). It is not surprising that the multinational managers,
equipped with information made available to them by advanced communication and technological tools
(Ricart et al., 2004), are making sophisticated decisions about slicing the activities of firms more finely and
spreading them more thinly to international locations more apt to undertake these activities (Buckley &
Ghauri, 2004; Ricart et al., 2004). Although location choice seems to be an important strategic decision, it
still remains an under-studied area in field of International Business (IB) (Buckley, Devinney, & Louviere,
2007; Cantwell, 2009; Dunning, 1998; Dunning, 2009; Mudambi & Navarra, 2003).
Location, though a central point of discussion in IB literature in the 1960’s (for example, Vernon’s
Product Life Cycle theory in 1966), lost ground as scholars shifted their attention to the firm and its
strategies in the 1970’s and 1980’s. Location reemerged as a factor to consider in the 1990’s in a changed
world order that includes increasing globalization, increased competition, interconnectedness of activities,
alliance capitalism, technological breakthroughs, and use of intellectual property as a central resource for
internationalization (Dunning, 2009). Realizing this, scholars began to study country effects and showed
them to positively contribute towards the growth and profitability of a multinational’s affiliate above and
beyond the effects of firm, business unit and industry (Brouthers, 1998; Christman, Day, & Yip, 1999;
Makino, Isobe, & Chan, 2004, Tong, Alessandri, Reuer, & Chintakananda, 2008).
For an emerging market firm (EMF) which, in general, possesses limited resources compared to
its developed country brethrens, choice of host country locations may be part of an important strategy
because the country and industry effects are shown to be even more salient than the firm and business unit
2
effects in case of EMFs (Makino, Isobe, & Chan, 2004). Consequently, an optimum host country location
may enhance an EMF’s resource base, competitive advantage and profitability (Perez-Batres & Eden, 2008;
Vermeulen & Barkema, 2002). Thus, it becomes important to study the location choices of EMFs which
have been internationalizing at a rapid pace; and are posed to internationalize further, as institutional
reforms unfold in emerging markets (Child & Tse, 2001). In recent times, outward FDI from developing
and transition economies reached a record high of US$ 304 billion (UNCTAD, 2008). At the same time,
international sales from emerging markets aggregated to approximately US$ 1.9 trillion with worldwide
employment increasing to 6 million people (UNCTAD, 2006). Despite these advances by EMFs,
knowledge about management outside North America is still lacking in both quality and quantity (Tsui,
2007; Werner, 2002).
The research question, namely location choices of EMFs, raised in this dissertation assumes
importance because of manifold reasons. First, scholars have started to investigate only recently as to why
and when firms from developing and newly-industrialized countries invest in other countries (Chen &
Chen, 1998; Lecraw, 1993), and how their location determinants differ from those of the developed country
firms (Kimura & Lee, 1998). However, consensus on these issues is still elusive and requires systematic
conceptual and empirical investigations (Makino, Lau, & Yeh, 2002). Second, the theories developed in
North America vary in degree of their applicability to emerging markets and require significant adjustments
before they can be used to study and interpret the processes of EMFs due to differences in social, cultural,
political, legal and economic systems in the two sets of countries (Hofstede, 2007; Li, 2004; Shenkar &
Von Glinow, 1994). Third, the resource advantages and internationalization processes of EMFs differ from
those of developed country firms (Curevo-Cazurra, 2008; Dunning, 2000).
The impact of a firm’s resources on its international location decision does not seem to be well-
explored, though there exists a well-developed literature on host country location choice; and though, firm
resources have long been shown to contribute towards the internationalization of both traditional and
virtual firms initially or subsequently in a host country (Bogner, Thomas, & McGee, 1996; Dunning, 1980;
Hitt, Bierman, Shimizu, & Kochhar, 2001; Hymer, 1976; Kotha, Rindova, & Rothaermel, 2001). The
relationship between firm’s resources and internationalization has been observed for EMFs as well
3
(Chittoor, Sarkar, Ray, & Aulakh, 2009). However, we know only a little about the relationship between
specific resource types and their impact on firms’ internationalization (Hitt, Bierman, Uhlenbruck, &
Shimizu, 2006). One such relationship between resources and internationalization that we are aware of is
that different types of resources impact differently the international performance of a multinational, with
human resources positively moderating the relationship between internationalization and firm performance,
while foreign government relational resources negatively moderating the relationship in case of services
firms (Hitt et al., 2006).
The resource-based view of the firm states that firms are heterogeneously endowed with different
resources. Since different types of resources impart different exploitative abilities to firms and since
different resources are fungible to varying degrees when applied to a different environment (Miller &
Shamsie, 1996; Ricart et al. 2004), it is highly likely that a diverse set of firm’s resources may differentially
affect its international location choices because a host country may have a set of economic, institutional,
social, and competitive environments that are different from the home country of the firm. Given
differences in factor endowments, demand conditions and competition levels between home and host
country; a firm’s resources can become either advantageous or disadvantageous in a host country (Curevo-
Cazurra et al., 2007; Caves, 1996; Dunning, 1995; Itaki, 1991; Teece, 1977; Zaheer, 1995) because the
advantages provided by resources are relative to the competitive environment in which the firm operates
(Amit & Schoemaker, 1993; Miller & Shamsie, 1996; Tallman, 1991). This indicates that resources may be
important determinants of firms’ location decisions. For example, Japanese firms that developed resources
such as engineering and sourcing capabilities stayed in Taiwan and Singapore when wage hikes occurred in
these two countries, while others that did not possess these resources had to relocate to Malaysia and
Thailand (Song, 2002). Similarly, the U.S. firms that had higher R&D intensity and a greater proportion of
overseas sales located in Israel while others that did not have these specific resources chose to locate in
other parts of Middle-East (Fiegenbaum, Shaver, & Yeung, 1997).
Moreover, a firm’s resources lead it to adopt certain strategies (Prahalad & Hamel, 1990) and an
optimum combination of resources and strategies generates sustainable competitive advantage for a firm
(Barney & Arikan, 2001). A firm endowed with different resources may adopt different international
4
strategies to match the environment in a country and be more competitive there (Kraatz & Zajac, 2001).
Thus, a firm may locate in different host countries with different international strategies. Hence, it is
probable that a firm’s international strategy, besides its resources impacts its location decision. To
understand the impact of an EMF’s resource-types and international strategies on its international location
decision, this dissertation makes a preliminary effort to identify, segment and categorize various types of
resources possessed by EMFs and in the process, makes yet another contribution to the present literature.
The dissertation makes several other contributions as discussed below.
First, the study seeks to fill a gap in the IB literature by combining the literature on MNC’s
resources and strategies with the literature on IB location choice. Scholars such as Werner (2002) and
Ricart et al. (2004) identify MNC’s strategies as an area of future research in IB. Peng (2001) reviewed the
use of resource-based view, one of the key developments in the area of strategy, in the field of IB between
years 1991- 2000 and found five IB areas, namely strategic alliance, market entry, international
entrepreneurship, emerging market strategies, and multinational management, where resource-based view
has been used. Peng’s (2001) study highlights the lack of work in integrating the strategy with location
literature. In fact, in my limited literature review, I came across only a few scholarly pieces, for example,
Xu and Shenkar (2002) and Makino et al. (2002) that examine the joint impact of resource type and
internationalization strategies on location choice of emerging market multinational firms. Henisz and
Macher (2004) suggest that the levels of firms’ technological capabilities affect their host country location
decisions. Yang, Jiang, Kang, and Ke (2009) conduct case studies of a Japanese (Matsushita) and a Chinese
(Haier) firm to study the joint impact of firm’s resources, industrial characteristics, and institutional
environments on internationalization of these two firms. In another study, Tseng, Tansuhaj, Hallagan, and
McCullough (2007) study how knowledge and property-based resources lead to different patterns of growth
in multinationality. Thus, it is not puzzling that combining strategy with IB is a potential area of future
research, at the least in location literature. Adopting the same approach as past scholars (had) who
combined strategy with IB field with help of resource-based view (Peng, 2001), this dissertation also uses
resource-based view to combine the two fields. Combining IB and strategy fields in the context of
emerging markets has been recognized as a prospective research area (Peng, 2001).
5
Second, the question of host country location choice, given an EMF’s resource base and possible
strategies, assumes both scholarly and managerial significance because the performance of EMFs in host
country depends upon how well they are able to transfer their competences from the home base (Child,
Chung, & Davies, 2003). A hostile foreign environment may dissipate the already scant resources of an
EMF and render it non-competitive in the market-place.
Third, by combining country factors with firm resources and strategies, the dissertation makes an
attempt to integrate different levels of analyses in line with the suggestions made to study location related
research by adopting a multilevel approach (Enright, 2002; Ricart et al., 2004). In general, multilevel
research is recommended in order to enrich the future research in field of IB (Arregle, Hebert, & Beamish,
2006; Hitt, Tihanyi, Miller, & Connelly, 2006) and management at large (Hitt, Beamish, Jackson, &
Mathieu, 2007).
Fourth, a multilevel approach has potential to generate a comprehensive model of location choice.
A comprehensive model would enhance our understanding of the location literature because there are
several competing theories explaining a number of plausible determinants considered by firms in making a
location decisions, for example Agglomeration externalities (Almeida, 1996; Porter, 2000; Saxenian, 1996;
Smith & Florida, 1994), Institutional and Political Distance (Delios & Henisz, 2000; Henisz, 2000;
Hillman, 2005; Nigh, 1985; Tallman, 1988;), Internationalization Process (IP) model (Johanson & Vahlne,
1977; Johanson & Weidersheim-Paul, 1975), Network Linkages (Bandelj, 2002; Chen, Chen, & Ku, 2004;
Filatotchev, Strange, Piesse, & Lien, 2007; Johanson & Vahlne, 1990), OLI paradigm (Dunning, 1988),
and Oligopolistic Reaction (Flowers, 1976; Knickerbocker, 1973; Yu & Ito, 1988). Besides these theories,
the firm resources (Dunning, 1983; Hymer, 1976) and firm strategies (Dunning, 1983) are also said to
impact the internationalization process of multinational firms. However, there is no framework which
makes an attempt to simultaneously study these competing but plausible explanations.
Fifth, a comprehensive model that is able to combine at least some of these different viewpoints to
study the host country location choice by the emerging market multinational firms will have beneficial
implications for managers because these above-mentioned competing viewpoints may guide them to
choose different host country locations. In the process, the managers may end up making a sub-optimal
6
location choice and may detrimentally spread the firm’s resources too thin. For example, if managers use
the theories put forth in the field of economic geography to guide the firm’s location decision, they may
like to augment the EMF’s limited resources (when compared with their developed country counterparts)
by venturing out to those geographic locations that provide network externalities irrespective of the
distance between their home country and the geographic clusters in the host country. At the same time, an
examination grounded in Network Linkages may suggest them that the firms should follow the IP model, as
ethnic community which is a very important source of network linkage, largely, remains concentrated in
areas bordering the home country of firms (Chen & Chen, 1998). On the contrary, an analysis of location
decisions of firms based on Institutional Distance may provide mixed support, contingent upon the extent
of government regulations in the industry of internationalizing firms (Canal-Garcia & Guillen, 2008). The
IP-model and the OLI paradigm will also expect managers to make different country choices and may
create a dilemma in reconciling the guidance of the two models (Curevo-Cazurra, 2008). No wonder,
Buckley et. al (2007) found that the managers appeared to follow fairly rational rules and fundamental
operational factors in creating sets of location alternatives but their final choice aligned more with country-
specific factors and less with traditional theoretical models.
Sixth, there have been calls by scholars to conduct further empirical research in the resource-based
view (Armstrong & Shimizu, 2007). This empirical dissertation indirectly fills that gap by drawing upon
the resource-based view to understand the location choice of EMFs.
Last but not the least; the dissertation indirectly examines, by studying the location patterns of
EMFs, whether an internationalizing EMF is constrained by “distance” while selecting a host country.
Distance is said to be one of the important location determinants considered by multinationals while
making their location decisions (Johanson & Vahlne, 1977). EMFs are said to be less risk-averse and less
guided by psychic, cultural, geographic and economic distances (Bonaglia, Goldstein, & Mathews, 2007;
Luo & Tung, 2007). A cursory look at the location choices of many EMFs from China, Mexico, Brazil, or
India suggests, though, a mixed support for the distance paradigm, indicating a need for further
investigations to explain the location choices of EMFs. There are only limited empirical studies exploring
this issue in context of EMFs which are in their initial stage of internationalization process, and hence
7
provide an interesting natural setting to test the distance paradigm proposed by Johanson and Vahlne
(1977). This insight would gain further importance as Dunning (2009) observes that location is now a more
complex decision than the one proposed in IP model.
To explore the research question raised here, an extensive review of literature on emerging
markets, resource-based view, internationalization strategies, and determinants of location choice is
undertaken to study the relationship between an emerging market firm’s resources and strategies, and its
international location choice. The dissertation combines these theories by employing the eclectic paradigm
of international production (Dunning & Lundan, 2008). This framework has been employed to study
similar research questions such as the choice of location by multilatinas (Cazzura, 2008). The eclectic
paradigm of international production combines MNC’s various international strategies such as market-
seeking, asset-seeking, efficiency-seeking, and resource-seeking with the OLI paradigm to explain reasons
for international production. The emphasis of the OLI paradigm, though, is to explain why a firm becomes
multinational. The paradigm explicates that when a firm internalizes its ownership-specific advantages
(which are both resources of a multinational firm and its transaction-specific advantages) and interacts
them with the location-specific advantages of a host country, it becomes multinational (Dunning & Lundan,
2008). Because the eclectic paradigm of international production incorporates the interaction of all
independent variables used in this dissertation, namely firm’s unique resources (which mean only asset-
specific and not transaction-specific advantages in the dissertation), its international strategies, and location
determinants, the framework is suitable to study this dissertation topic.
Another choice, to analyze the research question, may be the contingency based approach. As per
contingency theorists such as Lawrence and Lorsch (1967), Donaldson (2001), Drazin and Van de ven
(1985), and Miller (1981); a fit among resources, strategies and country selection is necessary to result in a
superior firm performance. Contingency theory emphasizes the role of firm resources, managerial
strategies, and environment towards firm performance. Though, the roles of environment, firm strategies
and a fit between the two have been validated in context of emerging markets (Child et. al, 2003); the
emphasis of the contingency approach is on superior firm performance as a dependent variable. Since the
8
dependent variable in this dissertation is host country location, the eclectic paradigm of international
production is better suited to explain the model proposed in the dissertation.
IP model is yet another frequently used model to explain the host country location decisions. But
its emphasis is more on distance as a location determinant. As the dissertation incorporates various levels of
analysis and makes an attempt to present a comprehensive model, the choice of IP model is not very
suitable for the research question raised in the dissertation.
Figure 1 summarizes the conceptual model proposed in the dissertation. In the model, it can be
observed that resource types of EMFs determine its internationalization strategies, which in turn, affect
location decision of EMFs, after controlling for other host country location determinants. The direct
relationship between the internationalization strategies and international location choice is proposed to be
moderated by the resource types of EMFs.
Figure 1: Conceptual Model
Resources
Relational Resource (Non-
market, Ethnic, Business)
Process-related Resource
Market Knowledge Resource
Natural-asset based Resource
Internationalization
Strategies
Market-seeking
Asset-seeking
Resource-seeking
Efficiency-seeking
Opportunity-seeking
Location Determinants (Control Variables)
Distance (Psychic, Cultural, Geographic,
Economic)
Institutions (regulatory, political,
societal)
Experiential learning and managers’
background
Agglomeration (network externalities,
imitation, oligopolistic reaction)
Macroeconomic factors (labor cost,
infrastructure, tax, exchange rate)
Customer following
Availabilit
y
of natural assets
Location Choice
DC (developed country)
EM (emerging markets)
LDC (least developed country)
9
The dissertation draws upon survey-based and archival data from Indian software firms to examine the
proposed model. The emphasis by Indian government on education has created a large supply of highly-
skilled labor force available at rather cost-competitive rates. Some Indian firms, especially the software
firms, have used this country-created asset to internationalize by developing process-based resources. The
high quality of the talent has led these firms to offer low-cost process-related solutions to their customers
and have helped the former go international. So, the dissertation studies how process-related resources and
the internationalization strategies of firms affect their location choices and empirically examines this
relationship in context of Indian software firms.
Given the overall model, I carry out a review of the literature that has bearing upon the model. The
literature review is divided into chapters 2-5. In chapter 2, a review of the literature on emerging markets
and EMFs is conducted. In chapter 3, a literature on Resource-based-view is detailed. In chapter 4, a review
of internationalization strategies is carried out; and chapter 5 makes an effort to explain the scholarly work
on host country location determinants.
10
CHAPTER II
LITERATURE REVIEW: EMERGING MARKETS
A review of literature on emerging market is necessitated as the context of the dissertation is set in
an emerging market, namely India. As western theories have limited applications in emerging markets
(Shenkar & Von Glinow, 1994), it is important to review this literature to understand the nuances of
emerging markets and how these affect EMFs.
The term “emerging markets” has been brought to surface by the World Bank in 1980’s (Wikipedia).
As per the literature, the defining characteristics of emerging markets are as follows:
low-income but rapid economic growth coming from liberalization of their economies (Luo, 2002;
Meyer, 2004), offering local firms a high-risk but high-return business environment (Makino,
Beamish, & Zhao, 2004).;
evolving regulatory, political and societal institutions on an already existing base of some
institutions that promote commerce (Gelbuda, Meyer, & Delios, 2008; Khanna & Palepu, 1997);
an economic environment, which is hostile (i.e. importance and deterrence of environmental
factors), dynamic (i.e. predictability and variability of environmental factors), and complex (i.e.
diversity and heterogeneity of environmental factors) (Luo & Peng, 1999; Tan & Litschert, 1994);
volatile economic policies with a gradual move towards paring down of government intervention
in the economy with an aim to move towards free market economic system;
unstable political scenario that has increased industrial and strategic uncertainties for local and
foreign firms (Hoskison, Eden, Lau, & Wright, 2000; Luo, 2002; May, Stewart, & Sweo, 2000;
Peng & Luo, 2000) and slowed inward FDI (Petersen & Pedersen, 1999);
weak legal framework resulting in high measurement and enforcement costs leading to high
transaction costs (Choi, Lee, & Kim, 1999; Hoskisson et al., 2000; Khanna & Palepu, 1997; Xin &
Pearce, 1996);
coexistence of both market mechanism (i.e. allocation of resources by market forces) and
redistributive mechanism (i.e. allocation of resources by the government) (Zhou, 2000);
missing well-defined property rights resulting in less competitive firms (Makino et. al, 2004);
11
rampant corruption, bribery, rent-shifting and opportunism (Hoskisson et al., 2000; Nelson, Tilley,
& Walker, 1998;), making enforcement of laws a big concern even when laws have been enacted
and government connections are pursued (Choi et. al, 1999; Hoskisson et al., 2000; Khanna &
Palepu, 1997); and hence turning the competitive environment dysfunctional in emerging markets
(Li & Zhang, 2007);
less competitive and smaller market-size with fewer resource-endowed firms (Aulakh, Kotabe, &
Teegen, 2000; Ghemawat & Khanna, 1998; Sol & Kogan, 2007);
fewer location advantages based on created assets such as infrastructure and human capital
(Hoskisson et al. 2000, Narula & Dunning, 2000; Meyer, 2004) but more location advantages
based on natural assets (Dawar & Frost, 1999; Aulakh et al., 2000);
less developed intermediaries like thin capital markets, shortages of skilled labor, underdeveloped
factor and product markets, and infrastructural bottlenecks resulting in high financial and
transformation costs for firms (Hong, 2004; Liu & Li, 2002; Khanna & Palepu, 1997; Meyer,
2004; Tan & Litschert, 1994; Wan, 2005; Wright, Filatotchev, Hoskisson, & Peng, 2005; Xin &
Pearce, 1996);
less mature financial markets making firms escape with inadequate disclosures, weak corporate
governance and control (Khanna & Palepu, 2000), and leading to a longer sustenance of poorly
performing firms (Chacar & Vissa, 2005) than in efficient capital markets of developed countries
where poorly performing firms are weeded out (Wan, 2005);
stronger societal influences and relationship-oriented culture (Hong, 2004; Hoskisson et al., 2000;
Luo & Peng, 1999); making managerial networking important in emerging markets but using it to
overcome fundamental threats such as extortion or expropriation rather than to obtain customers,
market information or secure credits as in developed markets (Henisz & Zelner, 1999; Xin &
Pearce, 1996);
General suspicion of stakeholders such as government, suppliers and customers towards foreign
firms (Hoskisson et al., 2000) requiring foreign firms to have a long presence (Child, 1997; Luo &
Peng, 1999) or local community involvement (Gifford & Kestler, 2008) to gain legitimacy.
12
Based on these characteristics; it may be suggested that the emerging markets lie somewhere in
between the least developed economies and the developed economies. Hoskisson et al. (2000) identify a
total of sixty-four emerging markets (appendix 1) categorized into two groups – fifty one developing
economies in Africa, Asia, Latin America, Middle East (the same as identified by International Finance
Corporation - IFC); and thirteen transition economies in erstwhile USSR block and China. However, all
these countries do not form a homogeneous set. There are differences among emerging markets due to
different pace of their economic and institutional reforms (Wright et al. 2005), and social cultures. For
example, institutions are changing slowly in Gulf Cooperation Council economies (Kshetri & Ajami,
2008). Consequently, even the gains from the economic and institutional changes have not been uniform
across emerging markets (Hoskisson et al., 2000). For example, inward FDI and organizational learning in
different emerging markets have varied from a high level in countries such as China, India to a low level in
countries such as Baltic States and Russia (Hitt, Ahlstrom, Dacin, Levitas, & Svobodina, 2004; Jansson &
Sandberg, 2008).
Asian countries of Hong-Kong and Singapore, which are normally termed as Newly Industrialized
Countries (NICs) are categorized as developed countries in the dissertation. NICs have high created asset
environments or advanced institutional environments (Wan, 2005), which may have led to differences
between EMFs and NIC firms. For example, the developed institutions in NICs have resulted in fewer
business groups, but more market-oriented and less diversified firms (Chakrabarti, Singh, & Mahmood,
2007; Guillen, 2000; Lee & Slater, 2007; Pananond, 2007; Wan, 2005). NIC firms depend more on ethnic
ties than EMFs in their internationalization efforts (Chen, 2003; Mathews, 2002). Hence, it is appropriate to
club NICs together with developed countries than emerging markets.
The literature on emerging markets can be divided into two phases with a first wave developed around
the early 1980’s by scholars such as Fagre and Wells (1982), Kumar & McLeod (1981), Lall (1983), and
Lecraw (1983), among others. These scholars used the term developing countries, though, and not
emerging markets. The second phase began in late 1990’s and continues to date. In between these two
phases, there were occasional scholarly pieces in the early 1990’s (Cantwell, 1989; Erramilli, 1992; Hu,
13
1995; Kumar, 1994; Lecraw, 1993; Lee & Beamish, 1995; Tan & Litschert, 1994; Tolentino, 1993;
Woodward & Rolfe, 1993).
Fagre and Wells (1982) linked various characteristics of a USA, European or Japanese transnational
corporation (TNC) such as size, intrafirm transfers, advertising, R&D intensity, and product diversity to
their extent of equity ownership in their affiliates in the developing countries. Lecraw (1983) extended the
literature by suggesting that the host developing countries too possessed some bargaining power vis-à-vis
TNCs and conducted an empirical study of USA, European and Japanese firms’ investment in five ASEAN
(Indonesia, Thailand, Malaysia, Singapore, Philippines) host countries. He proposed that the characteristics
of the host country and those of a TNC influence the TNC’s ownership percentage in its subsidiaries
(Lecraw, 1983). He found that the bargaining power of these host countries came from the possession of
scarce resources or the ability to control access to their markets. On the contrary, a TNC gained advantage
over the host country and entered through an FDI route if the TNC had a big brand name, high marketing or
management skills, technological products or could provide its subsidiary with inexpensive financial
capital. Lecraw (1983) found a J-shaped relationship between a TNC’s equity ownership in its affiliate and
the TNC’s perception of its affiliate’s success. Lall (1983) proposed that developing country firms
internationalized by exporting their products to other developing countries and the least developed
countries, but did not export to developed countries. These exporting firms possessed low cost advantage
and had developed efficient mass production technologies (Lall, 1983).
The first phase of the literature on internationalization of EMFs developed in pre-globalization era, but
the second phase of the literature development coincided with the globalization era which represents a
different phenomenon altogether, and which also changed the determinants of host country selection
(Dunning, 1998). Globalization era saw many countries liberalizing their economies, firms forming cross-
border alliances, and information flowing without asymmetry (Dunning, 2009). No wonder, scholars such
as Luo and Tung (2007) suggest that the analysis of the internationalization of EMFs in globalization era
requires a new perspective different from the pre-globalization era since the emerging market
multinationals of today are much less path-dependent and much more risk-taking than the third world
MNCs of 1980’s, though still sharing some basic strengths like cost advantages and weaknesses like
14
limited knowledge of international markets. The internationalization of EMFs in the globalization era is
driven less by cost factors; and more by search for markets and technological innovations to compete
successfully globally. These firms utilize pull factors and connections to accelerate internationalization
(Mathews, 2006). Moreover, the determinants of host country location choice seem to be changing with
some factors like population becoming more salient in 2000 than in 1980 (Flores & Aguilera, 2007).
Consequently, this dissertation utilizes more the literature in the second phase written after mid 90’s.
The literature witnessed a spurt of scholarly interest in emerging markets (especially in China) in the
second phase. There have been many dedicated sessions on emerging markets in leading journals such as
Academy of Management Journal (2000), Journal of Management Studies (2005), Journal of International
Business Studies (2007) and Journal of International Management (2007), to name a few. Most of the
studies in these dedicated volumes have used institutional theory, transaction cost economics or resource
based view or their combination to understand the nuances of emerging markets and how these
idiosyncrasies affect the economic, institutional, societal and resource environments in emerging markets
and their impact on development, operations and survival of local firms and foreign firms (Luo & Peng,
1999; Peng & Heath, 1996; Shenkar & Von Glinow, 1994; Suhomlinova, 2006). Further detailing is given
in the following paragraphs.
The constrained local environment replete with institutional voids and high transaction costs in
emerging markets has led some EMFs to develop unique hybrid organizational structures such as business
groups and interfirm networks (Hoskisson et al., 2000; Khanna & Palepu, 1997; Peng & Heath, 1996).
Other ownership types in emerging markets are state-owned-enterprises (SOEs), privatized firms,
entrepreneurial-led firms and foreign entrants (Peng, 2003). Business groups, themselves, can be of three
types, namely family-owned, widely-held, and state-owned (Curevo-Cazurra, 2006).
Business groups and network-based structures help these EMFs overcome weaker legal institutions by
promoting trust-based transactions among member firms or using network resources to enforce a contract,
thereby arresting the fear of opportunistic attempt in a transaction (Guillen, 2000; Khanna & Palepu, 1997).
Other usages of membership are access to intangible and financial resources residing within the group or
network (Chang & Hong, 2000; Khanna & Palepu, 1997; Makhija, 2004) or with its foreign partners who
15
team with the business group or network to access the governments or market-specific knowledge in
emerging markets (Lee & Beamish, 1995; Leff, 1978; Khanna & Palepu, 2000; Meyer & Nguyen, 2005).
These groups or networks develop reputational capital and become attractive partners for foreign
multinationals desirous of launching their operations in an emerging market (Lu & Ma, 2008; Pan & Chi,
1999), thereby gaining access to the superior technological and managerial skills of their multinational
partners (Leff, 1978; Luo & Tung, 2007; Khanna & Palepu, 2000). Foreign firms, despite contributing
critical resources, are willing to maintain a balance in management control, in order to enhance the survival
of the venture in emerging economies (Karhunen, Lofgren, & Kosonen, 2008; Lee, Chen, & Kao, 1998;
Steensma & Lyles, 2000).
But with more market-oriented economic and institutional reforms sweeping through in emerging
markets, many business group or network affiliated EMFs face obsolescence in their resource set as what
was important earlier may not be so important under the more market-oriented environment (Wright et al.,
2005). Business groups are observed to loose their advantage over other competitors (Hoskisson et al.,
2000), and even as joint venture partners procuring scarce resources for foreign multinational firms
(Khanna & Rivkin, 2006; Meyer, Estrik, Bhaumik, & Peng, 2008). The impact of managerial non-market
resources on business performance is contingent upon intensity of competition, growth rate and structural
uncertainty in the business environment, with a weaker influence on performance under competitive and
growing environment with decreasing structural uncertainty (Li, Poppo, & Zhou, 2008; Peng, 2003; Peng
& Luo, 2000) because networking benefits firms with market expansion and competitive positioning, but
does not enhance their internal operations (Park & Luo, 2001).
A recent study by Li and Zhang (2007) illustrates that the relationship between firm’s political
networking and new venture performance grew weaker under strong institutional environment; but the
firm’s functional experience in sales/marketing, finance, administration, R&D and manufacturing was
positively related to the new venture performance in high-tech industry in emerging market in a low
dysfunctional environment characterized by high property rights protection and ability to enforce contracts
(Li & Zhang, 2007) Another study by Li, Zhou, and Shao (2009) shows that the political ties impede
16
whereas business ties strengthen the positive relationship between foreign firm’s differentiation strategy
and profitability in China.
In the changed scenario, SOEs, networks and business groups, among others, are likely to persist
(Carney, 2008; Chang, 2006) but are transforming their personalized non-market resources into more
transparent and formal ties aimed at transferring to them the technological and managerial knowledge from
their multinational partners (Pananond, 2007); or are undergoing restructuring (Hoskisson, Johnson,
Tihanyi, & White, 2005; Kedia, Mukerjee, & Lahiri, 2006); or creating strategic flexibility (Uhlenbruck,
Meyer, & Hitt, 2003) in order to attain higher performance in the evolving environment. In fact, outward
FDI from China, though in its initial stage, is dominated by SOEs (Morck, Yeung, & Zhao, 2008).
The economic and institutional reforms in emerging markets have provided proper incentives to
many other EMFs to bring changes in their corporate culture that has enabled these firms to undertake
equity-raising with more corporate disclosures; put in effort to improve industry-specific technological
capabilities, build brands, launch more value-added and improved products suitable for the conditions
prevalent in emerging markets (Aulakh & Kotabe, 2008; Hooley, Cox, Shipley, Fahy, Beracs, & Kolos,
1996; Hoskisson et al., 2000; Newman, 2000; Pananond, 2007), enhance their contracting abilities (Toulan,
2002), and acquire managerial knowledge more appropriate for the changes taking place in the environment
(Soulsby & Clark, 1996).
From the literature review, it can be concluded that emerging markets have a different
institutional, economic, societal and resource environments than a developed country. However, EMFs, in
general, are making endeavors to develop their resources even under the constrained environment of their
home countries. EMFs are no longer depending solely on their non-market resources to tap the evolving
opportunities presented to them by the growing home country economies. But the institutional constraints
in emerging markets impact the resource-base of EMFs (North, 1990). The literature review in this chapter
presents a setting for studying the resource base of EMFs. Hence, in the next chapter, a literature review of
Resource-based-view is carried out in order to examine the resource-base of EMFs.
17
CHAPTER III
LITERATURE REVIEW: RESOURCE-BASED VIEW
The resources of a firm have gained prominent position in IB field. Early IB scholars such as
Caves (1971), Dunning (1983), and Hymer (1976) considered resources of a firm as drivers of its FDI.
These scholars, though, used a different term, namely ownership-specific advantage, instead of resource.
Dunning (1983) distinguishes between two kinds of ownership-specific advantages of a multinational firm,
namely asset-specific ownership advantages (OA) and transaction-specific ownership advantages (OT). The
model in the dissertation is concerned with EMF’s asset-specific ownership advantages, which are similar
to the term ‘resources’ in the resource-based view of firm. A literature review of resource-based view may
help us understand and categorize EMFs’ resource types, which form an important constituent of the
conceptual model presented in the dissertation.
The resource-based view of firm states that a firm’s internal resources, and not its external
environment, give the firm a sustained competitive advantage in the marketplace. Firm resources include
all assets, capabilities, organizational processes, firm attributes, information, knowledge etc. controlled by
the firm that enable it to conceive of and implement strategies that improve its efficiency and effectiveness
(Barney, 1991). Resources are heterogeneously distributed among different firms and provide competitive
advantage to those firms that have a portfolio of valuable, rare, inimitable and non-substitutable resources
(Amit & Schoemaker, 1993; Barney, 1991; Crook, Ketchen, Combs, & Todd, 2008; Grant, 1991; Peteraf,
1993; Wernerfelt, 1984). These advantage-generating resources cannot be purchased in factor market
(Dierickx & Cool, 1989). Resources such as technological abilities, reputation and brand name,
management skills, size, and international experience etc. have long been recognized as one of the key
drivers of a firm’s internationalization (Caves, 1971; Dunning, 1983; Hymer, 1976; Yeung, 1994) because
resources help an entrant firm to overcome disadvantages (Beamish, 1984), knowledge gaps (Petersen,
Pedersen, & Lyles, 2008) and liability of foreignness (Zaheer, 1995; Zaheer & Mosakowski, 1997) in a
host country.
It is highly likely that resource-base of firms from different countries may differ from each other
because firms develop unique resources to meet the idiosyncratic requirements of their home countries’
18
heterogeneous resource environments (Miller & Shamsie, 1996; Priem & Butler, 2001; Wan, 2005) that
determine the relative competitive advantages of home country firms (Chan, Isobe, & Makino, 2008; Fagre
& Wells, 1982). Moreover, the heterogeneity in economic and institutional conditions, that shape the
transformational and transactional capabilities of firms (North, 1990), determines whether the local firms
would be motivated to invest in path-breaking inventions and develop technological resources. This
suggests that the developed country firms may have different resource-base than EMFs because the two
sets of countries differ in their country-specific advantages.
The resource, economic and institutional environments in developed countries are such that the
local firms invest in path-breaking technologies and develop tangible property-based and intangible
knowledge-based resources (Khanna & Palepu, 1997). Such firms grow big and erect high entry barriers on
strength of efficient process and innovative product transformational technologies since firms in the
developed country environment are made available cheap finances and skilled labor (Wan, 2005).
On the contrary, EMFs have limited incentive to develop innovative path-breaking technologies or
develop superior managerial skills as property rights regime is weaker in emerging markets, restricting the
growth of these firms (Peng & Heath, 1996; Wan, 2005). The unpredictable government policies also
stymie the growth of firms (Khanna & Palepu, 1997). EMFs incur high transaction cost to access resources
at home (Chittoor & Ray, 2007; Xin & Pearce, 1996). Lack of infrastructure in emerging markets hampers
the brand-building efforts of EMFs (Khanna & Palepu, 1997), which do not possess international brands
and are perceived as producers of low-quality goods by consumers in developed countries (Cordell, 1993).
Thus, EMFs, in general, do not possess as strong physical, value-generating, or proprietary knowledge
resources such as size, experience, proprietary innovative technologies, superior human capital, reputation
or world-recognized brands as are possessed by the developed country firms (Barney & Arikan, 2001;
Craig & Douglas, 1997; Erramilli, Agarwal, & Kim, 1997; Sol & Kogan, 2007).
The distinctions between the resource-base of two sets of firms may be better understood if an
attempt is made to categorize the various types of firm resources. The extant literature provides some
categorizations of firm’s resources. Some of the resource types are given in Table 1:
19
Table 1: Resource Types in Extant Literature.
Resource Type Author and
Year
Explanation of Resource Type
Tangible / intangible
resources
Wernerfelt,
1984
Tangible resources: financial and production-related assets
like machine capacity;
Intangible resources: know-how, reputation, networks,
organizational culture, patents and licenses.
Human /
organizational /
physical resources
Barney, 1991 Human resources: training, experience and relationships of
managers and workers inside a firm;
Organizational resources: reporting system, planning,
coordination etc.;
Physical resources: plant & equipment, geographic location;
physical technology.
Physical / financial /
human / technological
/ reputation /
organizational
resources
Grant, 1991 Financial resources: Cash, ability to raise finances;
Technological resources: patents;
Reputation: brand reputation.
Contained / system
resources
Black &
Boal, 1994
Contained resource: an identified simple network of resource
factors that can be monetarily valued and traded;
System resource: socially created, complex network of firm
resource factors which make the monetary valuation and
tradability of a system resource implausible
Property-based /
knowledge-based
resources
Miller &
Shamsie,
1996
Property based resource: appropriable, controllable, specific
and well-defined resource which is protected from imitation
by virtue of property rights. Examples are organizational
slack, internally generated profits and externally raised
20
finances, patents and licenses.
Knowledge-based resource: protected from imitation by virtue
of knowledge barriers Examples are technological, managerial
and marketing skills.
Advantageous /
disadvantageous /
complementary
resources
Curevo-
Cazurra,
Maloney, &
Manrakhan,
2007
Disadvantageous resources: core-rigidities
Market / non-market
resources
Hirschman,
1958;
Kobrin, 1980
Non-market resource: helps a firm regulate its non-market
environment that includes relationship with actors such as
government institutions and community, which provide orders
to markets, firms and other types of institutions and
organizations to repair their failures and function them
effectively and efficiently; but does not include firm’s buying
and selling relationships;
Market resource: helps a firm consummate its market-related
tasks such as buying and selling
Customer assets /
channel assets / input
assets / process assets
/ market knowledge
assets
Verdin &
Williamson,
1994;
Markides &
Williamson,
1996
Customer assets: brand recognition, customer loyalty and
installed base.
Channel assets: established channel access, distributor loyalty
and pipeline stock.
Input assets: knowledge of imperfect factor markets, loyalty of
suppliers and financial capacity.
Process assets: proprietary technology, product or market-
specific function experience and organization systems
Market knowledge assets: accumulated information on the
21
goals and behavior of competitors, price elasticity of demand or
market response to business cycle.
It can be inferred that the above-tabulated resource types are more suitable for categorizing the
resources of developed country firms, though some of the above-mentioned resource typologies can be
used to categorize some (but not all) resource-types of EMFs. For example, non-market resources can be
used to label the resources of EMFs that network with governments to create asymmetry-based competitive
advantages (Hoskisson et al., 2000; Miller, 2003). Similarly, market knowledge assets can describe the
resources of EMFs that apply, to internationalize, the knowledge of how business gets done in emerging
markets (Sol & Kogan, 2007). However, there are other resources of EMFs, which solicit a new set of
resource typologies, more amenable to describe the resources of EMFs. For example, a survey by
UNCTAD (2006) found that the competitive advantages of EMFs stem from relationship management,
cultural affinity and organizational structure. Aulakh et al. (2000) suggest that EMFs can excel in
commodities since emerging markets possess plentiful natural resources and low cost labor. Mathews
(2006) proposes that EMFs develop resources by learning through their linkage with developed country
firms and leveraging those resources and linkages to internationalize.
Based on the EMFs’ resource description in the literature, this dissertation uses some existing and
develops some new resource typologies to propose the following four resource types for EMFs:
Relational resources,
Market knowledge resources,
Process-related resources, and
Natural asset based resources.
These typologies are in line with the six generic strategies suggested for the internationalization of
EMFs (Craig & Douglas, 1997) namely low-cost commodity, component manufacturing, private label
manufacturing, low-cost leader, first generation technology, and specialized niche. A detailed description
of each of these resources and how these help in internationalization of EMFs follows.
22
1) Relational Resources: A firm’s relationship with external constituents has been acknowledged as a
resource (Dyer & Singh 1998). Relational resources generate reciprocity-based trust, information sharing
and joint problem-solving among partners (Uzzi, 1997). In fact, FDI by EMFs is proposed as management
of important network relations (Chen, 2003). There are following three types of relationships observed in
case of EMFs.
with home and host governments, and with local and foreign financial institutions and banks. This
has been termed “Non-market Resource” in the dissertation;
with ethnic population termed “Ethnic Relational Resource”;
with foreign technology partners in OEM or joint venture relationship termed “Business
Relational Resource”.
(i) Non-market Resources: These resources provide the firms with an adaptive ability to move
beyond the institutional constraints and play a more active role in constrained environments (Oliver, 1991),
and have been found to positively influence the firm performance (Peng & Luo, 2000), including new
venture performance even in high-tech industry under weak institutional regimes of emerging markets (Li
& Zhang, 2007). The firms in emerging markets create sustainable competitive advantage by erecting
asymmetry barriers and institutional barriers (Farashahi & Hafsi, 2009; Miller, 2003; Wan, 2005) through
developing non-market resources that provide them access to the scant physical resources and restrict
others from accessing the same resources (Boddewyn, 1988; Brewer, 1993; Khanna & Palepu, 1997;
Malik, 2008; Wan, 2005). Non-market resources are drawn from cultivation of network resources by
managers of a focal firm by cooperating and exchanging favors with managers of other local firms and
governmental authorities (Acquaah & Bryan, 2007; Hong, 2004; Luo, 2001; Park & Luo, 2001; Xin &
Pearce, 1996).
Non-market resources of firms engender them an ability to manage institutional idiosyncrasies and
help in their internationalization (Henisz, 2003) because different institutional environments in host
countries may render a resource incapable of providing the competitive advantage to a multinational (Black
& Boal, 1994; Brouthers, Brouthers, & Werner, 2008; Oliver, 1997). Multinational firms are finding it
hard to harmonize or coordinate the different institutional environments in different countries because their
23
objectives may differ from the host country objectives (Makhija, 1993); resulting in increase in cross-
border transaction costs whenever these firms locate in institutionally dissimilar environments (Dunning,
2009) and poor performance by foreign affiliates (Delios & Henisz, 2000).
Though, government’s intervention in emerging markets is going down, there are some industries
such as infrastructure development where government intervention is still very high (Henisz, 2003) and
investment policies for foreign firms have been volatile (Doh & Ramamurti, 2003). The development of
infrastructure is marred with unclear regulations, expropriation hazards (Henisz & Zelner, 2001), and lack
of credible promises by emerging market governments (Doh & Ramamurti, 2003). Such a scenario offers
EMFs a strategic opportunity to invest in infrastructure sector in other emerging markets (Canal-Garcia &
Guillen, 2008; Ghemawat & Khanna, 1998) because EMFs can utilize their non-market resources to gain
an edge in constrained host country environments. The developed country firms have been reluctant to
invest in infrastructure sectors in emerging markets (Ramamurti, 2004) since a failure to understand the
institutional environment results in unforeseen costs and political hazards in global projects for the firms
(Henisz, 2000; Orr & Scott, 2008). In fact, emerging market infrastructure firms are among the largest
foreign investors in infrastructure sectors in other emerging markets, outperforming their developed nation
counterparts (Curevo-Cazurra & Genc, 2008). For example, Chilean firms have been investing in
infrastructure sectors in other Latin American countries where economic and institutional liberalization
occurred later, and were more successful than their developed country counterparts, since they could apply
the liberalization know-how lessons learnt in Chile to other Latin American countries where similar
conditions unfolded later (Sol & Kogan, 2007).
(ii) Ethnic Relational Resources: The other kind of relational resource stems from the ethnic
relationships which have played catalytic roles in creating home-grown multinationals in emerging markets
(Gillespie, Riddle, Sayre, & Sturges, 1999; Ramamurti, 2004), because EMFs, in general, have limited
knowledge of other international markets (Luo & Tung, 2007). Ethnic community can help a new entrant
understand the local rules of the land and provide access to key government actors (Chen & Chen, 1998).
For example, Thailand’s eminent business groups namely Charoen Pokphand and Siam Cement
internationalized on strength of their ethnic relationships (Pananond, 2007). Additionally, ethnic
24
community residing in a host country can open doors to some key customers in the host country. For
example, software industry in India grew partially on strength of diaspora in the USA and Europe (Kapur &
Ramamurti, 2001). Moreover, ethnic community in a host country can itself act as key customers. Jollibee
Group of Philippines draws on the ethnic community to open its restaurants internationally. Similarly,
Televisa from Mexico has launched its international operations by developing products that cater to the
taste of people with common ethnicity.
(iii) Business Relational Resources: Another type of relational resource arises from business
relationships of local EMFs either under OEM or contract supplier relationships with the developed country
firms that set up local operations with market-seeking strategy and are either forced by the host country
governments to use local components (Luo, 2002) or voluntarily seek such relationships to enhance their
competitiveness (Swamidass & Kotabe, 1992). Consequently, OEM relationships develop between the
local EMF and the developed country multinationals (Luo & Tung, 2007). Some other firms undertake
contract manufacturing for supplying goods to the developed country multinationals under private labels
(Dawar & Frost, 1999). Such contract-based operations have given an opportunity to these EMFs to learn
from their foreign partners and have led to, over a period of time, their expertise-development in
management and mass production with product-quality at par with the standards expected by multinational
firms (Lall, 1983; Li, Lin & Arya, 2008; Luo & Tung, 2007; Meyer, 2004). These local firms then use their
relationships to gain foothold in other markets (Elango & Pattanaik, 2007; Luo & Tung, 2007; Yiu et al.,
2007) by using their partners’ complementary skills (Elango & Pattanaik, 2007; Khanna & Palepu, 1997,
2000; Khanna & Yafeh, 2005; Pananond, 2001; Peng, 2003; Peng & Zhao, 2005; Yeung, 2000). For
example, Haier, a multinational EMF from China started as a private-label supplier to Liebherr of Germany
(Liu & Li, 2002), and later stepped up its international efforts by launching its own brand. Vertical
relational ties resulted in increased manufacturing productivity for Argentinean furniture firms and helped
them access global markets (Mesquita & Lazzarini, 2008). Many other EMFs get an opportunity to expand
abroad as sub-contractor to an existing multinational or local global customer when the latter launch their
operations into a new international market (Mathews, 2006). For example, multinational banks from
emerging markets are likely to follow their clients while internationalizing (Petrou, 2007).
25
(2) Market knowledge Resources: In the evolving environment of emerging markets, these EMFs
have adopted flexibility in their strategies and carried out dynamic alteration in their resources (Uhlenbruck
et. al, 2003) to develop suitable products for local consumers who are less-demanding but have distinct,
though infrequent changes, in customer tastes (Aulakh et al., 2000; Dawar & Frost, 1999; Luo, 2001),
requiring EMFs to not introduce as many innovative products in emerging markets as are required by the
developed country firms in their home markets (Aulakh et al., 2000). Further, EMFs have acquired
marketing skills to match the infrastructure challenges and rudimentary distribution channels present in
their countries (Aulakh et al., 2000; Dawar & Frost, 1999; Gomez, 1997; Lall, 1983; Sol & Kogan, 2007).
Equipped with this knowledge of their local markets, EMFs are able to project how operations are
conducted in other emerging markets and the least developed countries (Lall, 1984) and hence, face lower
knowledge gap when they internationalize in markets with similar economic, institutional and resource
environments (Lee & Beamish, 1995). EMFs turn the disadvantage of operating in a turbulent institutional
and economic environment replete with poorer regulatory quality, lower control over corruption,
underdeveloped infrastructure, and less-innovative products into an advantage when they invest in other
emerging markets and the least developed countries where institutional and economic environments are
similar (Cazzura & Genc, 2008; Dawar & Frost, 1999; Lecraw, 1993; Luo & Tung, 2007; Sol & Kogan,
2007). For example, business and commercial practices in countries such as China (for example: unique
terms of payment, higher price sensitivity, and personnel direct marketing) are different from those in
developed countries (Luo, 2001) and may give Chinese firms an advantage in other emerging markets.
Developed country firms, on the contrary, cannot work successfully with their traditional
strategies to tap the low-income strata in emerging markets (London & Hart, 2004; Ricart et al., 2004) as
business culture specificity impedes exploitation of existing capabilities which are rendered irrelevant and
redundant under different external environments of emerging markets (Dunning, 1995; Itaki, 1991; Luo,
2002; Miller & Shamsie, 1996; Priem & Butler, 2001; Wan, 2005). Consequently, Korean firms reported
higher stability of and satisfaction with their joint ventures’ performance than did the developed country
firms (Lee & Beamish, 1995). Telecommunication firms from the USA avoided markets with high
uncertainty in resource environment (Dowell & Killaly, 2008).
26
(3) Process-related Resources: Some emerging markets have provided stimulating business
environment in some industries. The motivating business environment in these industries accrues from
either the government’s investment in education that has created skilled labor-force at affordable rates in
these emerging markets; or creation of a suitable policy environment and infrastructure that have attracted
developed country multinational firms in secondary industrial sectors to turn these emerging markets into
manufacturing base with efficiency-seeking FDI. Multinationals have begun to produce standardized
intermediary goods in these emerging markets (Dunning, 1988; Makino, Beamish, & Zhao, 2004; Meyer,
1998; Uhlenbruck & Castro, 2000; Vernon, 1966). At the same time, the governments have provided
financial supports to the incumbent firms to become internationally competitive (Buckley et al., 2007; Wan,
2005).
Some of the progressive local firms have capitalized on the availability of skilled but cheap labor-
force in their home country by developing low-cost process technological alternatives to mature, labor-
intensive and standardized technologies from the developed countries (Meyer, 2004).
The initial resources of these EMFs have been further augmented through the technological,
international market access, and managerial knowledge spillovers from local presence of developed country
multinationals in their industries (Banga, 2006; Buckley, Clegg, & Wang, 2002; Chang & Xu, 2008;
Meyer, 2004; Tian, 2007; Wei & Liu, 2006). Besides spillovers, productivity gains for the progressive
domestic firms have also been observed from the location of foreign firms in emerging markets (Aitken &
Harrison, 1999; Chang & Xu, 2008; Meyer, 2004). The productivity gains have the potential to enhance the
international competitiveness (Porter, 1990) of EMFs in their industries. For example, several multilatinas
became multinationals when their countries undertook structural reforms (Cazzura, 2008).
As a result, some EMFs have become internationally competitive in standardized, mass-
production based process technologies by offering low-cost solutions (Dawar & Frost, 1999; Kumar &
McLeod, 1981; Lecraw, 1983; Wells, 1983). Thus, EMFs have become internationally competitive in
contract manufacturing, building materials, software-coding, and breweries; despite their smaller sizes,
distance from key consumer markets and lack of resources (Klein & Wocke, 2007; Mathews, 2006).
27
Many of these EMFs are observed to augment their resources by investing in developed countries
(Chen & Chen, 1998; Child & Rodrigues, 2005; Chittoor et al., 2009; Lecraw, 1993; Lee & Slater, 2007;
Luo & Tung, 2007; Klein & Wocke, 2007; Makino et al., 2002; Mathews, 2006) and are able to outgrow
the investment development path of their countries (Narula & Dunning, 2000) with an aim to excel in
development of product based technologies, besides process-based technologies. But the product-based
technological breakthroughs by them are more likely to be incremental and competence-enhancing than the
competence-destroying type (Tushman & Anderson, 1986).
(4) Natural Asset based Resource: Some emerging markets are endowed with plentiful natural
resources. Some firms in these nations can either compete worldwide in those natural commodities or
transform those natural resources into some value-added products at relatively cheaper rates because of
availability of low-cost labor (Aulakh et al., 2000; Dawar & Frost, 1999), as a nation can be competitive
worldwide in those goods that utilize in their production process the abundant factors of that country
(Makino et al., 2004).
EMFs have gained foothold in natural commodities and natural resource based value-added
products (Aulakh et al., 2000). Examples include Mittal Steel, Tata Steel and Indonesian pulp and paper
manufacturing firms (Kedia et al., 2006; Luo & Tung, 2007; Mathews, 2006).
It can be concluded from the literature review in this chapter that the resource base of EMFs is
different from those of developed country firms, though many resources of EMFs are derived from their
relationships with or presence of developed country firms in emerging markets. The difference in resource
base can be attributed to the differences in institutional and resource environments between developed
countries and emerging markets. It can further be understood that despite possessing a resource base which
is different from the resources that have been considered as drivers of multinationality of firms, EMFs can
still internationalize with help of their resource portfolios. As the dissertation draws data from the survey of
Indian software firms which have internationalized based on their process-related resources, it examines the
impact of process-related resources on the location choice of these firms.
Resources endow a firm with sustained competitive advantage when these resources are used to
craft value-creating strategies (Armstrong & Shimizu, 2007; Barney & Arikan, 2001; Hitt, Bierman,
28
Shimizu, & Kochhar, 2001; Newbert, 2007) that are matched with the external environment (Miller &
Shamsie, 1996) and are not simultaneously being implemented by current or potential competitors (Barney,
1991). Thus, it is the interaction of firm’s resources and its strategies that produce positive firm
performance (Hitt et al., 2001).
Since strategies are crafted in an environment of high uncertainty and complexity by managers
who have bounded-rationality (Gavetti & Rivkin, 2007) and discretion in resource deployment and
development (Amit & Schoemaker, 1993), the conditions of firm heterogeneity in possession and usage of
resources will result in different strategies by different firms. For example, various resource types give rise
to different alliance strategies for firms (Eisenhardt & Schoonhoven, 1996; Hoffmann, 2007; Park, Chen, &
Gallagher, 2002; Villalonga & McGahan 2005); different firm diversification strategies (Chatterjee &
Wernerfelt, 1991); different strategic paths (Teece, Pisano, & Shuen, 1997); vertical integration decisions
(Argyres, 1996); early or late entry (Schoenecker & Cooper, 1998); entry mode choice (Anand & Delios,
2002); outsourcing decision (Almor & Hashai, 2004; Jacobides & Hitt 2005); and different product launch
strategies (Hsieh, Tsai, & Hultink, 2006). Thus, different resource types are likely to generate different
types of strategies.
In the next chapter, I review the literature on firms’ internationalization strategy or motive. By
internationalization strategy, the dissertation means the motivations of firms for going abroad (Dunning,
1983). The term internationalization strategy, herein, does not purport to mean the multinational’s control
and structure related strategies, namely multidomestic, international, global and transnational strategies as
proposed by Bartlett and Ghoshal (1989).
29
CHAPTER IV
LITERATURE REVIEW: INTERNATIONALIZATION STRATEGIES
Internationalization strategies of a firm may affect its host country location choices because a firm
may not be able to implement all of its potential strategies in a host country, given the institutional,
resource, economic, and societal make-up of the host country. A literature review of internationalization
strategy may help us understand the strategy types a firm may have for its international market and at the
same time, it may help us select the internationalization strategies that can form part of our model and be
further examined.
Extant literature has identified many generic strategies of firms proposed by scholars such as
Hambrick (1983), Miles and Snow (1978), and Porter (1980) among others. In the context of international
business, scholars such as Cantwell (1989) and Dunning (1980) stand out for proposing strategies adopted
by multinational firms.
Miles and Snow (1978) used the firms’ strategies to categorize the former as defenders,
prospectors, analyzers and reactors. Defender firms prosper through stability, reliability and efficiency;
Prospector firms prosper through stimulating and meeting new product-market opportunities; Analyzer
firms prosper by being more innovative than defenders but less innovative than prospectors; Reactors keep
vacillating in their environment and fail to prosper (Miles & Snow, 1978). The strategies suggested by
Miles and Snow (1978) have been used by a limited set of studies in IB. Analyzer orientation is
demonstrated to be the best suited for Chinese markets, while prospector and defender strategies result in
poor financial performance for market-seeking subsidiaries of foreign firms (Luo & Park, 2001).
Porter (1980) proposed cost-based and differentiation-based strategies. Another dimension on
which generic strategy of a firm can be described is its focused or broad market scope (Porter, 1980). The
performance impact of generic strategies was shown to be contingent upon the environmental factors with
cost strategy working better in stable markets, while differentiation strategy working better in volatile
environment (Miller, 1988). These generic strategies have been studied in context of EMFs by Aulakh et al.
(2000) wherein they researched the relationship between the generic strategies of, choice of host country
location by, and performance of Latin American firms. They found that the cost-based strategies in
30
developed countries while the differentiation-based strategies in developing countries were associated with
better performance for Latin American firms. Li et al. (2009) demonstrate that low cost or differentiation-
based strategies are associated with foreign firms’ profitability in China. A recent meta-analysis showed
that the cost and differentiation paradigm of competitive strategy should be enhanced since its linkage with
the performance has not been established (Campbell-Hunt, 2000).
Hambrick (1983) suggested that firms can follow four generic strategy types, namely cost
efficiency, asset parsimony, differentiation, and scale/scope. Firms following cost efficiency strive to lower
the cost incurred per unit of output; whereas firms following asset parsimony strive to use fewer resources
per unit of output. Differentiation and scale/scope strategy types are the same as defined by Porter (1980).
The strategy types proposed by Hambrick for mature industries have been used by few scholars in IB.
Specifically keeping a multinational firm in view, Dunning (1983) proposed typologies for the
international strategies (he calls them motives though) of multinational firms. These motives are market-
seeking, resource-seeking, efficiency-seeking and/or asset-seeking (Dunning, 1983) with chief aim of all
these motives being to increase the firm’s competitiveness. Resource-seeking firms invest abroad to gain
access to (i) cheaper physical natural raw materials such as minerals, agriculture, or (ii) cheaper unskilled
or semi-skilled labor, or (iii) technological or management or marketing expertise. Market-seeking firms
invest abroad to gain access to the customers and markets in host country. Efficiency-seeking firms invest
abroad to make optimum use of location-specific advantages with an aim to achieve efficiency through
economies of scale and scope. Asset-seeking firms invest abroad to gain access to the assets of foreign
firms. Dunning’s four typologies have been extensively used in research works in context of emerging
markets, as shown in the following paragraph.
Developed country firms have entered emerging markets with market-seeking or efficiency-
seeking motives as standardized intermediate products can now be sourced from wherever it is cheapest to
produce these products (Dunning, 1980; Narula & Dunning, 2000; Makino et al., 2004; Porter, 2000; Sethi
et al., 2003; Uhlenbruck & Castro, 2000). FDI by developed country firms driven by efficiency-seeking
motives has been more successful than market-seeking motives in emerging markets (Meyer, 1998).
Curevo-Cazurra (2007) suggested that EMFs from Latin America went international to set up production,
31
marketing subsidiaries or a combination of the two. EMFs that have location specific advantages in home
country are more likely to start with marketing subsidiaries first. EMFs that have location-specific
advantages in host country (like natural resources or acquisition target) are more likely to start with
production subsidiaries and also those firms that can easily transfer their products and technologies abroad
are more likely to go for production subsidiaries first (Cazzura, 2007). Though EMFs have been
internationalizing with these four motives or strategies (Deng, 2003), asset-seeking motive and market-
seeking motive have been widely accepted for EMFs (Mathews, 2006; Luo & Tung, 2007; Peng & Wang,
2000). Firms from emerging markets are securing valuable organizational resources in international
markets to enhance competitiveness (Cantwell, 1989; Child & Rodrigues, 2005; Hoskisson et al., 2005; Lee
& Slater, 2007; Luo & Tung, 2007; Mathews, 2006; Narula & Dunning, 2000).
Dunning, later added three other motives for multinationals’ activities. These motives are escape
investments, support investments, and passive investments (Dunning & Lundan, 2008). Escape investments
refer to the FDI made to escape restrictive legislation or macro-organizational policies by home
governments. Examples include round-tripping of investments between China and Hong-Kong (Dunning &
Lundan, 2008). Support investments are FDIs acting as cost-centers with an aim to support the activities of
rest of the firm. Examples include purchasing outfits of companies like Sears and Wal-Mart (Dunning &
Lundan, 2008). Passive investments are portfolio investments with equity infusion but no direct
management control. Examples are petro-dollar investments by Middle-Eastern firms (Dunning & Lundan,
2008). As these three motives are not the mainstream motives of multinational firms, these three motives
have found limited use in IB research.
Cantwell (1989) suggested that a multinational firm can go abroad either with asset-exploitation or
asset-augmentation strategies. Asset-exploitation is transfer of firm’s proprietary assets abroad and asset-
augmentation is acquisition of strategic assets such as marketing, technological or management skills by
firm. Cantwell’s strategies have also been used in context of emerging markets, but their emphasis is more
on technological acquisition, as shown in the following references. The technically advanced firms are
increasingly dispersing their activities geographically in order to augment their technical assets (Cantwell,
Dunning & Janney, 2004; Cantwell & Janney, 1999, Kummerle, 1996). EMFs internationalize to gain
32
managerial and technical knowledge (Aulakh, 2007; Lecraw, 1993). EMFs may start internationalization
with asset-exploitation mode, but soon follow it with asset-augmentation mode (Klein & Wocke, 2007).
Multinational’s asset exploitation and asset building are inversely associated with environmental
complexity and industrial uncertainty (Luo, 2002). Business culture specificity impedes exploitation but not
capability building (Luo, 2002). For example, the success of firms like Haier, Lenovo and South African
Breweries in developed countries like the USA supports their operations in other countries through
technological and reputational spin-offs (Liu, 2007; Liu & Li, 2002; Klein & Wocke, 2007).
Hitt, Hoskisson, and Kim (1997) suggest that firms diversify internationally with various motives
such as economies of scale, access to new resources, location advantages, cost reduction, and knowledge
acquisition. Nachum and Zaheer (2005) proposed that multinational firms have market-seeking, efficiency-
seeking, resource-seeking, export-seeking and knowledge-seeking motives to go international.
Specific to internationalization of EMFs, Craig and Douglas (1997) proposed six generic
strategies, namely low-cost commodity, component manufacturing, private label manufacturing, low-cost
leader, first generation technology, and specialized niche.
Further, Dawar and Frost (1999) proposed typologies for the international strategies of EMFs.
They suggested that EMFs may be defender, dodger, extender and/or contender based on a two-
dimensional matrix with dimensions as ‘pressures to globalize in its industry” and ‘extent to which it can
transfer its resources abroad’. An extender firm can go to analogous markets with similar economic,
institutional and/or societal environments; a contender firm can go to any market; a dodger firm is able to
gain competitive advantage in its domestic market against the international rivals due to possession of some
assets that find value in unique domestic economic, institutional or social environment; and a defender firm
has to concentrate on defending its domestic market share against the onslaught of international rivals
(Dawar & Frost, 1990). On the contrary, Luo and Tan (1998) suggest defensive strategies for local firms
but analyzer strategies for foreign firms in China.
Luo and Tung (2007) propose that EMFs internationalize for strategic-asset seeking, opportunity-
seeking and market-seeking purposes. Opportunity-seeking is another perspective added to the existing
literature by these scholars. Emerging market and emerging transition market firms face institutional voids
33
and trade constraints at home such as quota restrictions and/or anti-dumping penalties, and firms from these
countries internationalize to circumvent the constraints at home by seizing opportunities offered by other
countries. For example, China's FDI, at its infant stage though, is biased towards tax havens and proximate
Southeast Asian countries (Morck et al., 2008).
For the purpose of this dissertation, I propose to use the four typologies proposed by Dunning
(1983) and augment them with opportunity-seeking motive proposed by Luo and Tung (2007). The four
typologies offered by Dunning (1983) supplemented by Luo and Tung’s (2007) opportunity-seeking
strategy offer a richer explanation than other existing typologies of internationalization strategies because
the former provides a comprehensive coverage of potential strategies that can be followed by EMFs in their
internationalization effort. Dunning’s (1983) typologies capture parsimoniously the international strategies
of multinational firms put forth by various scholars in IB field. It is further believed that these five
typologies, namely market-seeking, asset-seeking, efficiency-seeking, resource-seeking, and opportunity-
seeking would help to understand better the research question in the dissertation. As the empirical
examination in the dissertation draws data from Indian software firms, the dissertation focuses on market-
seeking and labor-seeking strategies because these two strategies have been used by Indian software firms
in international markets.
Firms’ strategies affect the internationalization paths of multinationals (Makino et al., 2002). A
firm evaluates all possible FDI locations based on traditional determinants and selects the country that best
fits its strategy (Sethi et al, 2003). In next chapter, a review of the literature on determinants of host country
location choice is carried out.
34
CHAPTER V
LITERATURE REVIEW: DETERMINANTS OF HOST COUNTRY LOCATION CHOICE
A literature review of determinants of host country location choice is required to study the model
proposed in the dissertation. The review will provide a better understanding of the determinants that have
been examined by the scholars and will make the model comprehensive enough to consider all
determinants already examined by the scholars.
FDI determinants are now complex and multi-dimensional (Sethi et al., 2003) but very important
variables as they impact a firm’s profitability (Vermeulen & Barkema, 2003). There have been umpteen
research studies to understand the determinants of host country location choice. These studies have been
conducted using longitudinal data gathered from firms from different industries including manufacturing
(Yu & Ito, 1988; Rose & Ito, 2008) and services (Kundu & Contractor, 1999), different countries (but
mainly developed countries), and different time frames to increase the external validity of the research
outcomes. Some of these studies have investigated the determinants affecting FDI from a specific home
country into several host countries or inward FDI into a specific host country from several home countries,
whereas others have used datasets that includes multiple home and host countries. A tabular synopsis of
some of the research studies on this topic follows in Table 2.
Table 2: Research on Location Determinants
Authors Year
Published
Sample
Countries
Industry
examined
Factors affecting FDI
location decisions
Johanson &
Weidersheim-Paul
1975 Sweden -ve effect of psychic
distance i.e. firms locate in
proximate countries in initial
years.
Flowers 1976 European and
Canadian firms
investment in
USA
+ve effect of oligopolistic
reaction
35
Johanson & Vahlne 1977 Sweden -ve effect of psychic
distance
Root & Ahmed 1978 Inward FDI in
developing
countries
+ve effect of per capita
GNP, low corporate tax and
political stability.
Nigh 1986 USA Banking
Engwall & Wallenstal 1988 Sweden Banking
Terpestra & Yu 1988 USA Advertising
agencies
+ve effect of customer
presence, oligopolistic
reactions, market size,
international experience,
geographic proximity.
Yu & Ito 1988 Inward FDI in
the USA
Tire and textile
industries
+ve effect of oligopolistic
reaction
Dixit 1989 -ve effect of exchange rate
on FDI location
Benito & Gripsrud 1992 Norway Manufacturing No effect of cultural distance
on location
Erramilli 1992 USA Services Moderating impact of
international experience on
cultural distance and FDI
location
Li & Guisinger 1992 USA, Europe,
Japan
Services +ve effect of market size and
oligopolisitc reaction, but –
ve effect of cultural distance.
Woodward & Rolfe 1993 Export-
oriented units
+ve effect of per capita
GDP, exchange rates, length
36
of income tax holidays,
presence of export-
processing zones, political
stability and manufacturing
concentration; -ve effect of
wage rates, transportation
cost and inflation rates.
Kumar 1994 USA Export-units +ve effect of infrastructure,
size of export-processing
zones, availability of skilled
manpower, while -ve effect
of wage rates, openness of
country
Hennart & Park 1994 Japanese
investment in
the USA
Mimetic effect on location
decision
Mariotti & Piscitello 1995 Inward FDI in
Italy
-ve effect of information
cost assymertry between
foreign and local firms on
foreign firm’s location
decisions
Huchzermeier &
Cohen
1996 -ve effect of corporate tax
and exchange rate on
location decisions
Kogut & Chang 1996 Inward FDI in
the USA
Electronics
industry
+ve effect of prior entry in
and –ve effect of real
exchange rate movements of
37
the host country
Grosse & Trevino 1996 -ve relationship between
cultural, geographic distance
and imports from the USA.
O’Grady & Lane 1996 Canada Retailing Psychic proximity does not
guarantee success in host
country. Out of 32 firms
investigated, only 7 were
successfully functioning in
the USA.
Shaver, Mitchell &
Yeung
1997 Manufacturing +ve effect of experience in
host country on FDI survival
Mudambi 1998 +ve impact of prior host
country investment on future
investment.
Yamori 1998 Japan Financial
institution
+ve effect of manufacturing
FDI and market size.
Dow 2000 Australia Exporters Geographic distance but not
Psychic distance is
significant predictor
Nachum 2000 Inward FDI in
the USA
Financial
Services
+ve effect of agglomeration
benefits
Shaver & Flyer 2000 Inward FDI in
the USA
+ve effect of agglomeration
on laggards than on leaders
Carpenter &
Fredrickson
2001 USA +ve effect of TMT’s
international experience,
educational and tenure
38
heterogeneity on global
strategic posture.
Henisz & Delios 2001 Japan MNEs
-ve effect of political
hazards of host countries.
Ito & Rose 2002 Tire +ve effect of oligopolistic
reaction and international
experience.
Song 2002 Japan MNEs Moderating effect of
subsidiary capabilities on –
ve effect of wages.
Stare 2002 Slovenia,
Czech
Republic,
Hungary
Services -ve impact of institutional
and cultural distance.
Zhao, Delios, Yang 2002 Japanese
investment in
China
+ve effect of regional
development, transportation
infrastructure, market size
and trade
Zhao & Zhu 2000 Inward FDI in
China
+ve effect of market
potential, cost factors and
infrastructure adequacy.
Sethi, Guisigner,
Phelan, Berg
2003 USA
investment in
East Asia
+ve effect of low wages and
market size.
Globerman & Shapiro 2003 USA MNEs +ve impact of governance
infrastructure including
regulation, property rights,
39
legal systems.
MacCarthy &
Atthirawong
2003 Labor costs, political
stability, infrastructure,
economic factors affect
international plant location
decisions
Henisz & Macher 2004 Semiconductor +ve impact of technological
sophistication of host
country, but –ve impact of
political hazard.
Trevino & Mixon 2004 Inward FDI in
Latin America
-ve effect of institutional
distance
Bianchi & Ostale 2005 Inward FDI in
Chile
Retailers +ve effect of institutional
embeddedness.
Chang & Park 2005 Korean firms
investment in
China
Inverted U-shaped
relationship between
network externalities and
collocation
Gimeno, Hoskisson,
Beal, & Wan
2005 USA Telecom Mimetic effect on location
decision
Kim 2005 USA Auto-parts Geographic proximity to
customers is preferred to re-
location to Mexico where
wages are low.
Tihanyi, Griffith &
Russell
2005 Meta-analysis No significant relationship
between cultural distance
and international
40
diversification, but –ve
effect for high-tech
industries.
Alcacer 2006 Cellular
handset
+ve agglomeration effects
on R&D subsidiary and for
less capable firms
Rothaermel, Kotha, &
Steensma
2006 Outward FDI
from USA
Internet firms -ve relationship with country
risk, cultural distance,
uncertainty avoidance; +ve
relationship with
individualism and
masculinity. Market size
moderated the relationship.
Alcacer & Chung 2007 Inward FDI in
the USA
+ve effect of industrial
agglomeration on less
capable firms; while +ve
effect of academic
agglomeration on more
capable firms.
Bhardwaj, Dietz &
Beamish
2007 43 nations Cultural dimensions, namely
uncertainty avoidance has –
ve effect, but trust has
moderating effect on FDI
Filatotchev, Strange,
Piesse, & Lien,
2007 Taiwanese
firms
investment in
China
Ownership structure of firms
has influence on location
decision.
41
Flores & Aguilera 2007 Top 100 US
MNCs
+ve effect of market
affluence, infrastructure,
similarity in political and
legal systems, trust levels;
but –ve effect of cultural
distance and wage levels.
Hutzschenreuter,
Pedersen, & Volberda
2007 +ve impact of managerial
intentionality and firm
experience; while
moderating impact of
institutional forces on
internationalization path.
Petrou 2007 Banks +ve effect of customer
presence on developing
country banks’ location
decisions; while +ve effect
of market size on developed
country banks’ decisions.
Pusterla & Resmini 2007 Inward FDI in
Hungary,
Bulgaria,
Romania,
Poland
+ve effect of market size and
agglomeration; while –ve
effect of per capita wages.
Tsang & Yip 2007 Singapore MNEs +ve impact of economic
distance on FDI hazard rates
Cuervo-Cazurra 2008 Latin America MNEs Initially expanded to
economically closer
42
markets.
Coeurderoy & Murray 2008 Britain,
Germany
New-tech
firms
+ve impact of intellectual
property protection;
moderating impact of home
country regulatory regime.
Dowell & Killaly 2008 USA Telecom firms -ve effect of frequency and
amplitude of market
resource variation of a host
country. Experience in the
host country moderates the
relationship.
Ellis 2008 China Exporters Psychic distance moderates
the relationship between
market size and entry
sequence.
Hutzschenreuter &
Voll
2008 Germany MNEs -ve effect of expansion
moves with high levels of
added cultural distance per
unit of time and irregular
moves in culturally distant
nations on profitability.
Nachum, Zaheer, &
Gross
2008 USA MNEs +ve impact of geographic
proximity to centers of
knowledge, markets and
resources
Rose & Ito 2008 Japan Automobile Bandwagon effect is not
universal in oligopolistic
43
industries
Wang & Schaan 2008 Japan MNEs Non-linear relationship
between cultural distance
and performance
It can be observed from the above table that the literature on location determinant has been
developing for the past thirty years and results still pose a potential to enrich it further. It can be further
interpreted from the table and literature review that the host country location determinants may broadly be
classified into the following seven categories:
(i) four types of distances namely psychic distance, cultural distance, geographic distance, and
economic distance;
(ii) three types of institutional environment namely regulatory institutions, political institutions, and
societal institutions;
(iii) experiential learning and managers’ background;
(iv) agglomeration;
(v) macroeconomic factors;
(vi) customer or partner following; and
(vii) availability of natural resources.
A further review of each of these categories is given below.
(i) Distance: Distance in the field of international business may be of four types, namely psychic
distance, cultural distance, geographic distance and economic distance. Ghemawat (2001) proposed another
type of distance, namely administrative distance in his CAGE distance model. But administrative distance
has found only limited use in the literature on “location determinants”. Hence, administrative distance is
not used in the dissertation.
Psychic Distance: The term psychic distance is coined by Beckerman (1956), but used formally in
IB by Johanson and Weidersheim-Paul (1975) and Johanson and Vahlne (1977). Psychic distance includes
differences in language, business practices, political systems, religion etc. (Johanson & Vahlne, 1990).
44
Psychic distance may be one of the reasons as to why the biggest multinational firms still derive majority of
their revenue from one (Rugman & Verbeke, 2004) or two regions (Dunning, Fujita, & Yakova, 2007).
Johanson and Weidersheim-Paul (1975) and Johanson and Vahlne (1977) studied the location
choice of Swedish firms in their initial stage of internationalization. They suggested that the process of
internationalization is fraught with uncertainties of various kinds which can be reduced only by knowledge
acquired by conducting overseas operations. These uncertainties deter the host country location choice of
an internationalizing firm such that a firm starts with countries that are closer to its home country in terms
of geographic and psychic distance.
Results of studies on psychic distance and sequence of foreign market entry have been
inconclusive. Dow and Karunaratna (2006) summarize that there is a debate about psychic distance as to
“what constitutes it”, whether it should be measured using objective measures or using perceptions of
individuals and propose a construct called “psychic distance stimuli” to hopefully end this debate. Dow and
Karunaratna (2006) split psychic distance into (i) psychic distance stimuli using educational, cultural and
religious distance, and former colonial ties and degree of democracy, and (ii) perceived psychic distance as
perceived by managers since the managers select the host country location. In a recent study on 924
international entry sequence of 73 Chinese exporters, Ellis (2008) proposed that psychic distance cannot be
the cause of foreign direct investment which is a big financial investment with long-term investment. The
author showed psychic distance to act as a moderator between the relationship of foreign market size and
location choice, challenging the direct effect of psychic distance (Ellis, 2008).
Cultural distance: This factor has been shown to directly influence the FDI rate (Grosse &
Trevino, 1996; Li & Guisinger, 1992) or moderate the relationship between market size and FDI location
(Flores & Aguilera, 2007). Cultural distance punctuates organization’s learning in different countries and
cultural blocks (Barkema, Bell, & Pennings, 1996) and was significant for successful international joint
venture operations of Dutch firms in developing countries but not in developed countries (Barkema,
Shenkar, Vermeulen, & Bell, 1997). Relational linkages based on ethnic ties and cultural similarity
promoted the international growth of Taiwanese and Singaporean firms in China and South East Asia by
minimizing transaction and coordination costs (Hsing, 1996; Yeung, 2000). Individual dimensions of
45
culture affect differently the international locations of firms. Firms prefer to locate in those countries that
have cultures marked with low levels of uncertainty avoidance and high levels of trust (Bhardwaj, Dietz, &
Beamish, 2007).
A study using data of 2404 expansion moves by 91 German multinationals whose international
expansion was tracked for 5 to 20 years examined that the firms making expansion moves involving a high
level of added cultural distance per unit of time and the firms expanding in culturally distant countries in
irregular fashion exhibit less profitability (Hutzschenreuter & Voll, 2008). Similarly, a foreign direct
investment data of Japanese firms in 53 countries covering 36 years of their foreign expansion moves
illustrated that cultural distance and profitability had inverted U-shaped relationship (Wang & Schaan,
2008).
However, a recent meta-analysis found only the moderating effect of cultural distance on
international diversification depending upon the level of technological intensity of an industry and only for
recent samples which may indicate that firms are diversifying into culturally distant countries (Tihanyi,
Griffith, & Russell, 2005). Managers create framework to understand a different culture and can manage
those cultural distances where they understand the bilateral cultural differences (Chapman, Gajewska-De,
Clegg, & Buckley, 2008).
Geographic Distance: Outward FDI of Korean firms was deterred by the geographic distance in
initial stages of internationalization (Erramilli, Srivastava, & Kim, 1999). Geographic distance was found
to be a factor affecting inward FDI in Mexico (Thomas & Grosse, 2001). A country’s proximity to global
distribution of knowledge, markets and resources may promote FDI in that country (Grosse & Trevino,
1996; Nachum et al., 2008), but the result was found to be more pronounced for smaller firms than larger
firms (Nachum et al., 2008).
Economic Distance: Ghemawat (2001) proposed economic distance which includes difference in
wages and technological capabilities of the two countries. Several multilatinas began their international
journeys with economically closer markets (Curevo-Cazurra, 2008a). But another study found that FDI
hazard rates are lower in countries with more or less economic development than in countries with the
similar level of economic development (Tsang & Yip, 2007).
46
(ii) Institutional environment: Firms value those institutional environments that make them exploit
their competitive advantages in host countries (Dunning, 1998). Multinational firms are advised to locate in
countries with low institutional differences as firms need to conform to the local institutions (Trevino &
Mixon, 2004). Institutional compatibility in the location portfolio of a multinational firm increases its
ability to benefit from knowledge flows among its various nodes (Dunning, 2009; Kostova, 1999; Kostova,
Roth, & Dacin, 2008; Kostova & Zaheer, 1999). Failure to embed in local institutions led to failure of
many multinational retail giants in Chile (Bianchi & Ostale, 2005). As per neo-classical economics,
institution of a country has three dimensions, namely regulatory, political and societal (North, 1990).
Regulatory Institutions: Government regulations affect the growth of firms (Capelleras, Mole,
Greene, & Storey, 2008). Internationalization paths are considered as outcome of assessment of transaction
cost and risks (Buckley & Casson, 1976). Effectiveness of laws pertaining to intellectual property rights
influence the international location choice of American firms (Globerman & Shapiro, 2003), and German
and British new-tech firms (Coeurderoy & Murray, 2008). The relationship was moderated by similarity in
home and host country legal regime (Coeurderoy & Murray, 2008).
Political institutions: Policy environment of a country includes its set of laws, regulations,
administrative procedures, and policies formally sanctioned by the government that impact a firm’s
profitability by altering its costs and revenue (Delios & Henisz, 2003b). Uncertainty in policy environment,
degree of corruption, and opaqueness in government and legal processes make a country politically
hazardous (Henisz, 2000) and can deter FDI (Delios & Henisz, 2003a; Dunning, 2009; Globerman &
Shapiro, 2003; Henisz & Delios, 2001; MacCarthy & Athirawong, 2003; Woodward & Rolfe, 1993).
Political risk has been shown to moderate the relationship between market size and FDI location (Flores &
Aguilera, 2007).
Among different constituents of political institutions, corruption has received special attention in
location literature. Corruption can be pervasive or arbitrary (Rodriguez, Unlenbruck, & Eden, 2005).
Corruption in a host country affects the sources of inward FDI whereby attracting higher FDI from
countries where corruption is high (Cuervo-Cazzura, 2006). It poses a negative trade-off between market-
attractiveness and FDI (Grosse & Trevino, 2005; Wei, 2000), but does not deter FDI when corruption is
47
arbitrary rather than pervasive (Cuervo-Cazurra, 2008). Multinational firms take recourse under entry
modes such as short-term contracting or joint ventures to adapt to corruption in host countries (Uhlenbruck,
Rodriguez, Doh, & Eden, 2006). Moreover, corruption affects resource-seeking but not market-seeking FDI
(Brouthers, Gao, & Mcnicol, 2008). A country’s level of political constraints and economic development
negatively affect the bribery activity of local firms (Husted, 1999; Martin, Cullen, Johnson, & Parboteeah,
2007). As emerging markets have fewer political constraints and lower economic development than the
developed countries (Henisz, 2000), firms from emerging markets are likely to engage in bribery and are
likely to remain undeterred by the level of corruption in a host country. Corruption is said to be rampant in
Asian countries (Luo, 2002). Thus, level of corruption is unlikely to be a factor affecting the location
choice of EMFs.
Societal institutions: Societal trust has been shown to moderate the relationship between market
size and FDI location (Flores & Aguilera, 2007). Xu and Shenkar (2002) used the sociological dimensions
of institutions, namely cognitive, normative and regulative to propose that (i) multinationals with routine-
based advantage are likely to enter normatively proximate market, (ii) multinationals following global
strategy are likely to enter normatively and cognitively closer markets, and (iii) multinationals with multi-
domestic strategy are likely to enter distant normative and cognitive markets.
(iii) Experiential learning and managers’ background: Experiential learning has been highlighted
and oft-researched in International Business literature (Hutzschenreuter, Pedersen, & Volberda, 2007).
Prior investment and duration of stay in a host country increases the probabilities of further investment in
the same host country (Erramilli, 1992; Mudambi, 1998) and survival rate (Shaver, Mitchell, & Yeung,
1997), even in host countries with high frequency and amplitude of changes in resource environment
(Dowell & Killaly, 2008). Relevant types of international experiences turn firms less sensitive to the
deterring effects of uncertain policy environments on investment (Delios & Henisz, 2003; Henisz &
Macher, 2004) but still a firm needs to balance exploitation of its experience in a host country or its region
and exploration of new geographies while making a host country selection (Barkema & Drogendijk, 2007).
Experience engenders time-based and transfer-based learnings which ease the process of cross-border tacit
48
knowledge transfer, increasing a firm’s propensity to set up manufacturing plants in a foreign location
(Martin & Salomon, 2003).
An under-researched determinant of the international paths of firms is about managers’
background (Coeurderoy & Murray, 2008; Hutzschenreuter et al., 2007). Top management characteristics
such as international experience, higher elite education, lower average age, and higher average tenure are
positively related to firm’s international diversification (Sambharya, 1996; Tihanyi, Ellstrand, Daily, &
Dalton, 2000). Managers with more international experience seem to provide consistent models of country-
selection and appear to be less risk-averse in making decisions (Buckley et. al, 2007). Top management
team’s educational and tenure heterogeneity, and international experience are shown to be positively
related to a firm’s global strategic posture which has three dimensions, namely foreign sales, foreign
production and geographic diversity (Carpenter & Fredrickson, 2001). Governance issues such as family,
non-family insiders such as directors and CEOs and institutional shareholding pattern are found to impact
the location choices of Taiwanese firms into China (Filatotchev et. al, 2007).
(iv) Agglomeration: As knowledge-intensive, innovative and entrepreneurial activities become
geographically concentrated in clusters due to cross-border differences in institutional and human
environments (Breschi & Malerba, 2001; Cantwell, 1995; Cantwell, 2009; Dunning, 2009); multinational
firms agglomerate in these clusters for network externalities (Almeida, 1996; Almeida & Phene, 2004;
Baum, Li, & Usher, 2000; Frost, 2001; Krugman, 1991; Jaffe, Hendersen & Trajtenberg, 1993; Marshall,
1920; McCann & Folta, 2008; McCann & Mudambi, 2004; Mudambi, 2008; Nachum, 2000; Phene &
Almeida, 2008; Porter, 2000; Saxenian, 1996; Smith & Florida, 1994). Firms also agglomerate when prior
stock of investment by multinationals in a host country gives signals to other firms to locate there
(Dunning, 1998).
Another reason for agglomeration to occur is when firms imitate each other in order to gain
legitimacy or reduce uncertainty associated with internationalization (DiMaggio & Powell, 1983; Gimeno
et. al, 2005; Levitt & March, 1988; Suchman, 1995) or when firms are under pressures of oligopolistic
rivalry or face multimarket contact (Ghemawat & Thomas, 2008; Graham, 1978; Haveman & Nonnemaker,
2000; Knickerbocker, 1973). International firms are attracted to a host country for any of these reasons.
49
Firms are shown to avoid agglomeration too, depending upon their status as a leader or laggard (Shaver &
Flyer, 2000), and the extent of their R&D or product differentiation (Chung & Alcacer, 2002; Nachum &
Wymbs, 2005). A further detail is given below.
Network Externalities: Domestic and international firms co-locate to increase their learning
(Almeida, 1996; Baum et al., 2000; Jaffe et al., 1993) or competence by undertaking specific activities at a
particular location by taking advantage of spatially-embedded resources (Cantwell, 2009; Cantwell &
Mudambi, 2005; Dunning, 2009; Nachum & Zaheer, 2005; Porter, 2000). Strong network externalities are
found within firms than across firms, from firms in the same industry than firms from different industries,
and from firms of same nationality than firms from different nationalities (Henisz & Delios, 2001).
Relationship of network externalities and location choice is shown to be both positive and negative
(Aharonson, Baum, & Feldman, 2007; Alcacer, 2006; Baum & Haveman, 1997; Baum & Mezias, 1992;
Chung & Kalnins, 2001; Shaver & Flyer, 2000). First time entrants into the USA markets during 1985-94
located to maximize their net spillovers as a function of location’s knowledge activity, their own
capabilities and competitor’s anticipated actions (Alcacer & Chung, 2007). A recent study demonstrated
curvilinear relationship between network externalities and the likelihood of co-location (Chang & Park,
2005). The sales and production subsidiaries were illustrated to be more geographically dispersed while
R&D subsidiaries were more concentrated in worldwide cellular handset industry in year 2000 (Alcacer,
2006).
Imitation: Guillen (2002) showed that emerging market multinationals which are in early stage of
internationalization imitated other firms in making the location decisions.
Industrial Oligopolistic Rivalry or Multimarket Contact: Competition determines the co-location
of firms in a host country (Flowers, 1976; Haveman & Nonnemaker, 2000; Hennart & Park, 1994; Li &
Guisinger, 1992; Ito & Rose, 2002; Knickerbocker, 1973; Sethi et al., 2003;Yu & Ito, 1988). Knickerboker
(1973) theorized that FDI is the result of an oligopolistic reaction to other competitors, producing a
bandwagon effect. The host country location choice of global tire firms and Japanese largest automobile
firms is guided by the presence of their domestic rivals in the host country (Ito & Rose, 2002; Rose & Ito,
2008). The authors illustrated that these firms competed in some key markets but avoided unnecessary
50
competition in other markets (Rose & Ito, 2008). Similarly, international location choice of six largest
multinationals in cement industry is explained with help of rivalry resulting from multimarket contact
(Ghemawat & Thomas, 2008).
(v) Macroeconomic factors: It includes market size and market growth, barriers to trade, costs of
labor and other resources, transportation and information costs and availability of infrastructure such as
energy, roads and communication; government’s initiatives, and trade and tax regulations. Exchange rates
also affect location decisions (Dixit, 1989). It is highly likely that different internationalization strategies
may make some macroeconomic factors more salient in location choice than others.
(vi) Customer or Partner following: Firms, especially in services sector and business networks
follow abroad their customers or business partners (Chen & Chen, 1998; Hennart & Park, 1994; Li &
Guisinger, 1992; Martin, Swaminathan, & Mitchell, 1998; Miller & Parkhe, 1998; Nigh, 1985; Petrou,
2007; Yamori, 1998).
(vii) Availability of natural resources: Dunning (2009) expects a continued renaissance in natural-
resource-seeking FDI especially by emerging market and emerging transition market firms.
It can be observed further from the literature review that only a few empirical studies have
examined the location choices of EMFs. Almost all of these studies have used categorical variables to
divide the host countries into two basic categories of developed or developing countries. For example,
Erramilli et al. (1997) proposed that EMFs may find themselves into two kinds of markets: those that are
more advanced or less advanced than their own market. Aulakh et al. (2000) studied relationship between
the location choice of and strategies for host countries by Latin American firms and categorized the host
countries as developed or other developing countries. Makino et al. (2002) suggest that Taiwanese firms
can go to either developed countries or less developed countries. Makino et al. (2004) suggest that a
developed country firm can go either to another developed country or to a less developed country which
includes emerging markets, and least developed countries. A special issue in Journal of Management
Studies (2005) edited by Wright et al. suggests that EMFs can go to other emerging markets and developed
countries.
51
EMFs are unlikely to consider all seven categories of host country location determinants proposed
above. For example, experiential learning and managerial background may not be as important a
determinant for EMFs at this point in time, as these firms are in their initial stages of internationalization.
Moreover, availability of natural resources or customer or partner following is unlikely to affect the
location decisions of Indian software firms. Consequently, the dissertation controls for the remaining four
categories, namely distance, institutions, agglomeration, and macroeconomic factors.
To understand the location choice of Indian software firms in the dissertation, a two-way
categorization of various countries (either developed or emerging market countries) is used by proposing
that Indian software firms can internationalize by seeking to enter either developed countries or other
emerging markets. NICs are clubbed with developed countries in the dissertation. It is highly unlikely that
Indian software firms would be locating in the least developed countries at this time because of market and
resource configurations in these countries. It is believed that the two-way categorization would help
understand the research question better as the factors that impede or facilitate internationalization such as
economic, institutional, societal, and resource environments are different in these two categories of
countries.
52
CHAPTER VI
MODEL AND HYPOTHESES
It can be observed from the literature reviews in chapters 2-5 that EMFs possess different types of
resources and internationalization strategies. Further, it can be observed that there are limited studies that
researched the host country location determinants of EMFs. Specially, the question as to what location
determinants EMFs consider for choosing a host country location has not been addressed well by scholars.
Also, how EMFs’ resource types and internationalization strategies affect the location choice have received
limited scholarly attention.
The conceptual model presented in this dissertation examines the impact of the above-mentioned
variables, namely resources and internationalization strategies on the location choices of EMFs. The
conceptual model proposed in the dissertation argues that resource types of EMFs, namely relational
resources (which are non-market resources, ethnic relational resources, business relational resources),
market-knowledge resources, process resources, and natural-asset based resources, determine the
internationalization strategies, namely market-seeking, efficiency-seeking, asset-seeking, resource-seeking,
or opportunity-seeking, to be adopted by the EMFs in their overseas journey. For example, EMFs with
business relational resources may go international with a market-seeking or an asset-seeking strategy,
whereas EMFs with natural asset based resources may go international with market-seeking or resource-
seeking strategy. On the other hand, EMFs with non-market resources may go overseas mainly with
market-seeking strategy.
Further, the model suggests that the internationalization strategies of the EMFs will directly affect
the host country location choice, after controlling for various location determinants. The model further
proposes that the direct relationship between EMF’s internationalization strategy and the location choice
will be moderated by the resource type possessed by the EMF. Thus, it is not necessary that all EMFs with
market-seeking strategy will choose the same location. The location choice is decided by the combination
of resource type and internationalization strategy of the EMF. For example, a market-seeking EMF with
non-market resources and another market-seeking EMF with business relational resources will make
different location decisions. The former may locate in an emerging market where it can advantageously
53
deploy and exploit its non-market resources whereas the latter may locate in a country where its major
customer opens up operations. That country could be developed nation or another emerging market.
Similarly, a market-seeking EMF with ethnic relational resource will mainly locate in a country where the
size of the ethnic population is among the largest because it may target to sell its product to the ethnic
population in the host country, and this host country could be a developed nation or another emerging
market or a least developed nation.
Model
Resources
Relational Resource (Non-
market, Ethnic, Business)
Process-related Resource
Market Knowledge Resource
Natural-asset based Resource
Internationalization
Strategies
Market-seeking
Asset-seeking
Resource-seeking
Efficiency-seeking
Opportunity-seeking
Location Choice
DC (developed country)
EM (emerging markets)
LDC (least developed country)
Location Determinants (Control Variables)
Distance (Psychic, Cultural,
Geographic, Economic)
Institutions (regulatory, political,
societal)
Experiential learning and
managers’ background
Agglomeration (network
externalities, imitation,
oligopolistic reaction)
Macroeconomic factors (labor
cost, infrastructure, tax, exchange
rate)
Customer following
Availability of natural assets
54
Some scholars such as Cuervo-Cazurra (2007), Makino et al. (2002), Sethi et al. (2003), and Xu
and Shenkar (2002) have investigated relationships among some variables that form part of the model
described in the preceding paragraphs. Makino et al. (2002) study the relationships among home country
factors, firm’s internationalization strategies and capabilities on the host country location choices of firms
from Taiwan. These authors, however, do not discuss the role played by the location determinants.
Moreover, the host countries are divided into developed or the least developed countries in the said study.
Cuervo-Cazurra (2007) employs the eclectic paradigm of international production to explain the
relationship between the firm capabilities derived from home or host country assets and the
internationalization strategies chosen by multilatinas investing in other Latin American countries. Cuervo-
Cazurra (2007), however, studies a sub-segment of the comprehensive model proposed in this dissertation
and his coverage of FDI location is restricted to the Latin American countries. Sethi et al. (2003) study, as
one of the relationships in their model, the effect of efficiency-seeking and market-seeking
internationalization strategies on the location determinants considered by the US multinationals. The focus
of their study, however, is to examine the changing trends in flow and determinants of US FDI as a result
of macro-economic and firm strategy considerations. The emphasis of Xu and Shenkar (2002) is to explore
conceptually the effect of normative, regulatory and cognitive dimensions of institutional distance as a
location determinant on various aspects of FDI by multinationals. As part of their model, these authors
propose a relationship between institutional distance as a location determinant and resource-types of
multinational firms.
Based on the above discussion, it can be concluded that these studies do not examine the
simultaneous effect of firm resources and internationalization strategies on the host country location choice,
after controlling for various location determinants. Dunning and Lundan (2008), in their eclectic paradigm
of international production propose that the firms’ internationalization strategies, resources and the location
determinants chosen by firms affect the international production decision of firms; they fall short of
proposing a model that delineates specific relationships among these variables. Furthermore, the eclectic
paradigm of international production does not include the host country location chosen by the firm as one
of the variables in the framework. For example, it does not discuss whether firms will locate in developed
55
countries, NICs, or emerging markets to follow any specific internationalization strategy. The eclectic
paradigm of international production as a framework, however, provides a basis to study the model
proposed in the dissertation as this framework suggests that there exists a relationship among resources,
internationalization strategies and host country location determinants of a firm and these variables also
form part of the model discussed in this dissertation. Hence, the eclectic paradigm of international
production can provide guidance to delineate relationships among the variables.
Before specific hypotheses are proposed, it is important to understand if EMFs have a resource
portfolio consisting of more than one resource out of the relational resources (which are further of three
types namely non-market, ethnic, and business relational resource), market-knowledge resources, process-
related resources, and natural-asset based resources; and if these organizations use more than one resource
in order to internationalize as it has become tough to internationalize based on a single advantage (Aulakh
et al., 2000). For example, Indian software, besides process-related resource, have also been said to use
ethnic relational resource (Kapur & Ramamurti, 2001) and business relational resource (Elango &
Pattanaik, 2007) to internationalize. To understand what resources have been utilized by the sampled Indian
software firms in their internationalization effort, a short questionnaire is administered to top and senior
managers of the sampled Indian software companies. In each company, two managers are asked to rate on a
scale of 1-5 the importance of various resources namely the three types of relational resources, market-
knowledge resources, process-related resources, and natural-asset based resources that have helped their
organizations internationalize. The survey results demonstrate that majority of these firms utilize one
resource predominantly to locate in a specific location. It can be further noted from the survey result that 52
of the 64 contacted managers, forming 81.25% of the respondents rate process-related resources as the most
important resource that helped their company internationalize. Thus, it is proposed in this dissertation that
the key resource of Indian software firms that motivated them to internationalize and start looking outward
is process-related resource. This assumption helps make the model simpler and parsimonious.
Internationalization Strategies and Location Choices of Indian Software Firms.
Resources of a firm provide sustainable competitive advantage to the firm in the marketplace
(Barney, 1991) and guide its strategy (Malik, 2008; Noda & Bower, 1996; Papadakis, Lioukas, &
56
Chambers, 1998; Prahalad & Hamel, 1990; Tseng et al., 2007). Firm-specific resources may be
advantageous and fungible to varying degrees when transferred to another host country due to institutional,
economic or social environmental dissimilarities between home and host countries (Anand & Delios, 1997;
Buckley & Casson, 1996; Cuervo-Cazurra, 2007; Erramilli et al., 1997; Miller & Shamsie, 1996; Rugman
& Verbeke, 1992). Thus, firms are likely to choose different internationalization strategies in different
international markets to match their resources and overcome the disadvantages in a host country.
Consequently, there may be heterogeneity in firms’ internationalization strategies depending upon their
resource configurations.
Indian software firms may locate in international markets with a market-seeking strategy or
augment their resource base with a labor-seeking strategy. Process-related resources possessed by Indian
software firms have enabled these firms to offer low cost alternatives to the standardized and mature
products. Indian software firms have excelled in offshore codification of the standardized processes of
multinational firms. The value-proposition created on strength of process-related resources has made these
firms internationally competitive. The process-related resources may be fungible in different institutional
environments since the customers mainly derive value from the low-cost offerings of these EMFs.
It has been observed that high wages in developed countries made multinational firms in these
countries look for low-cost alternatives when they are developing software to automate their processes
(Doh, 2005). Indian software companies have the potential to fill this gap and offer low cost solutions to
the large multinational firms in developed nations. Firms in emerging markets may not face as much cost
pressure as developed nation firms do, though the former set of firms may be motivated to harness the
advantage of information technology in order to increase their competitive advantage. However, the market
in emerging countries is not likely to be as big or profitable as in developed countries. Thus, the process-
related resources of Indian software firms are likely to be advantageous in their international journey by
helping them overcome liabilities of foreignness in developed nations, and by increasing their international
sales and profitability, which is the prime objective of a market-seeking strategy (Dunning, 1983).
Indian software firms are likely to face competition from firms from other countries that have
availability of skilled but cheap manpower. At the same time, the growth of Indian software firms has
57
resulted in recruitment of talented Indian engineers in large numbers which has scaled up the wages in
India (Doh, 2005; Scheiber, 2004). The salaries in India have been moving up in double digits for last
several years (Economist, 2006) and are forecast to show a similar trend at least until 2011 (Minder, 2008).
Consequently, Indian software firms may face shortages of skilled labor in their home country and the cost-
arbitrage opportunity available to these firms may shrink with the upward movement of the labor cost.
Hence, to sustain their international competitiveness, Indian software firms need to augment their resource
base with a labor-seeking strategy (Dunning, 2006; Sethi et al, 2003). For example, many Japanese firms
moved out of Taiwan to locate in Malaysia when the labor cost went up in Taiwan (Song, 2002).
Developed countries offer abundant skilled labor but at a high cost. Some emerging markets offer
affordable skilled labor such that locating with a labor-seeking strategy in those nations has the potential to
serve or maintain the cost differential, one of the drivers of offshoring. Indian software firms are unlikely to
face a big liability of foreignness in other emerging markets as the economic conditions there are similar to
India. Moreover, the past international experience of Indian software companies, though in developed
nations, may have built in these firms some managerial skills required to coordinate dispersed teams and
increased their knowledge stock about the unique business environments in other countries. International
experience also prepares the managers to overcome the complexities inherent in an international expansion
(Zahra, 2003). Consequently, several Indian software firms have set up operations in other countries such
as Mexico, Malaysia, and China to tap the human resources there, which is the prime motive behind a
labor-seeking strategy (Dunning, 1983). Hence, the following is proposed:
Hypothesis 1: Among Indian software firms, those pursuing a market-seeking strategy are more
likely to locate in a developed country than those pursuing a labor-seeking strategy, which in turn,
are more likely to locate in an emerging country.
To study the impact of resources on the location decisions of Indian software companies, it is
important to comprehend what variables constitute the construct called process-related resources. To
understand and establish the content validity of the construct of process-related resource in the context of
Indian software firms, the industry experts and the top managers of some of the largest Indian software
companies were contacted. Based on discussions with them, the process-related resources that are
58
responsible for the remarkable success of Indian software companies may have the following three
dimensions:
Knowledge management by the software vendor firm,
Internal processes of the software vendor firm to deliver quality goods to the client,
Low cost but qualified human talent with the software vendor firm.
Each of these resources and its impact on location decision is detailed below.
Electronic Knowledge Sharing Database as a Resource and Location Choice
Knowledge sharing in an organization benefits it in many ways such as by enhancing customer
satisfaction or reducing costs (Nonaka, 1994). An absence of knowledge sharing in an organization may
result in missed opportunities and inefficiencies (Milliken, Morrison, & Hewlin, 2003). A study by
Babcock (2004) reports that Fortune 500 companies may loose at least USD 31.5 billion a year if
employees fail to share knowledge.
An effective approach adopted by multinational firms to capture knowledge sharing may be to
establish an electronic knowledge sharing database. A knowledge sharing database provides organization-
wide information at a click and makes it easier for employees to access and contribute to the information.
In a software company, a knowledge sharing database may help a project team by informing it of client-
specific knowledge if the firm has worked with the same client in the past, or project-specific knowledge if
the firm has worked on a similar project in the past, or any other technology-related information vital to the
successful completion of the project. Ready availability of client or project specific knowledge may lower
the effort, elapsed time, and rework – the hallmark of a successful offshoring project (Gopal,
Mukhopadhyay, & Krishnan, 2002). A knowledge management database by codifying the knowledge may
contribute in an effective manner to the success of a software project where project team members often
work with each other in distributed location settings and different time zones which may have adverse
impact on knowledge sharing among employees. Thus, a knowledge sharing database may assist the project
team in completing the project within envisaged time and cost budgets, and may enhance the client
satisfaction with the vendor software company. No wonder, Dyer & Singh (1998) and Tallman, Jenkins,
59
Henry & Pinch (2004) report increase in the competitive advantage of the firms that effectively implement
knowledge sharing practices among their employees.
Despite these obvious benefits, not all companies can successfully implement and document
knowledge sharing among their employees (Lepak & Snell, 1999) because firms may lack organizational
mechanisms (Majchrzak, Rice, King, Malhotra, & Ba, 2000), technological resources that help employees
share knowledge (Fulk, Flanagin, Monge, & Bar, 2004) or an organization culture conducive to motivate
employees to share knowledge (Bock, Zmud, Kim & Lee, 2005). Thus, an electronic knowledge sharing
database may act as a competitive enhancing resource for a firm.
As the expertise residing in an organization can be harnessed and put to use effectively with the
aid of an electronic knowledge sharing database, a vendor Indian software company, by effectively tapping
the expertise within the organization, can ride the experience curve to lower its cost of project and deliver
the software product within the time schedules originally decided with a client. Such vendor companies can
lay credible claims to client companies distantly located in developed markets and may sound more
convincing to client companies considering automating their processes than a vendor software company
that does not possess an electronic knowledge sharing database. Gaining client credibility is imperative
because outsourcing decision by a client company is a complex and daunting task (Aydin & Bakker, 2008)
and may deter the latter to engage in offshoring projects. Moreover, a vendor software company that
possesses a knowledge sharing database is more likely to offer know-how to the client company facing
knowledge drain occurring at its premises because of downsizing or outsourcing business activity than a
vendor company that does not possess knowledge sharing database (Aydin & Bakker, 2008).
Though developed countries offer bigger markets for software firms than emerging markets, the
former are highly competitive. Developed countries see many big multinational vendor software firms such
as IBM etc. competing fiercely for large and prestigious projects initiated by large multinational client
companies. Many of the client companies have several years of experience in offshoring and have the
expertise to assess the capabilities of vendor companies. Hence, it is possible that the Indian software firms
that possess a knowledge sharing database will have an increased tendency to locate in developed countries
with a market-seeking strategy than firms that do not possess a knowledge sharing database. However, as
60
the Indian software firms that do not possess a knowledge sharing database have an already high likelihood
of locating in a developed country with a market-seeking strategy, the increased propensity of firms with a
knowledge sharing database to locate in a developed country with a market-seeking purpose is unlikely to
be much higher than for firms without such a database. Hence, the effect may not be observable in the case
of market-seeking strategies of Indian software firms.
On the contrary, the interaction effect will be pronounced in case of a labor-seeking behavior of
the two set of firms. Since, by virtue of having the potential to lower their costs by effectively harnessing
the knowledge distributed in various pockets of the organization, an Indian software firm with a knowledge
sharing database may be able to set up a global delivery center in a developed country and still be cost
competitive. Locating with a labor-seeking strategy in close vicinity of its client companies in a developed
nation may enable the vendor software company to serve its client companies more effectively than a
vendor company that does not locate proximally to the client company. Thus, a knowledge sharing databse
is likely to enable an Indian software firm to locate in a developed country with a labor-seeking strategy.
Hence, the following is argued.
Hypothesis 2: Among Indian software firms, the relationship between their strategy-type and
location choice will be moderated by the possession of an electronic knowledge sharing database,
such that a firm that possesses the database will show more likelihood to locate in a developed
country than one without the database when its strategy-type is labor-seeking, but not when its
strategy-type is market-seeking.
Low Cost of Software Production as a Resource and Location Choice
The high cost of production in developed countries due to high wages has boosted the offshoring
of IT work to emerging countries such as India (D’Costa, 2004). However, over a period of time, the wages
in India have seen a steady increase in the software industry (Doh, 2005; Scheiber, 2004). This may render
many Indian software companies uncompetitive to locate in developed markets with a market-seeking
strategy. Unaffordability of an expensive IT solution and unclear associated cost-benefit advantages may
deter client companies to engage in IT offshoring projects. It is interesting to note that the firms that have
managed their labor bills efficiently also happen to be some of the most reputed Indian software companies.
61
Their brand-names, good organization culture, and reputation to offer professional growth to their
employees may have helped these companies attract the best talent but at competitive rates. Hence, the
Indian software companies that have been able to keep their labor cost low may find themselves in a
competitive advantageous position to serve developed markets than the firms that have not been able to do
so. Hence, the Indian software firms that have low wage bills will have an increased tendency to locate in
developed countries with a market-seeking strategy than the firms that have high wage bills. However, as
the Indian software firms that have high wage bills are likely to have an already high likelihood of locating
in a developed country with a market-seeking strategy, the increased propensity of firms with low wage
bills to locate in a developed country with a market-seeking purpose is unlikely to be much higher than for
firms with high wage bills. Hence, the effect may not be observable in case of market-seeking strategies of
Indian software firms.
On the contrary, the interaction effect is likely to be observable in case of a labor-seeking behavior
of the two set of firms. Indian software companies that have kept their wage bills low may be in a more
comfortable position to locate in developed countries with a labor-seeking strategy than the companies
whose wage bills are high. Opening global delivery centers in developed countries can improve
communication, coordination, and conflict resolution between client and vendor companies resulting in an
increased psychological contract between the two (Miranda & Kavan, 2005), reduced project uncertainties,
and improved performance (Gopal et. al, 2002). Locating in a developed nation with a labor-seeking
strategy may enable client and vendor companies to build relational capital and trust that positively affect
the client’s propensity to outsource (Holcomb & Hitt, 2007; Miranda & Kavan, 2005). However, locating
in a developed market with labor-seeking strategy can push the wage bills even higher for those vendor
companies that already have higher wage bills because the labor cost is high in developed nations. A high
wage bill may turn these vendor companies uncompetitive in the offshoring business – a trend that was
observed in Ireland (Doh, 2005). On the contrary, firms with low wage bills may be able to soft-land
against the impact of high wages when they locate in a developed country with labor-seeking strategy.
Therefore, Indian software companies with high wage bills are not as likely to locate in developed markets
with labor-seeking strategies as an Indian software firm with low wage bill. Hence the following is argued.
62
Hypothesis 3: Among Indian software firms, the relationship between their strategy-type and
location choice will be moderated by wage bill, such that a firm with low wage bill will show
more likelihood to locate in a developed country than one with high wage bill when its strategy-
type is labor-seeking, but not when its strategy-type is market-seeking.
Process Improvement Implementation as a Resource and Location Choice
The capability maturity model integration (CMMI) in software engineering is a process
improvement approach which supports project teams, departments or entire organizations in improving
their internal processes by providing them guidance and reference points. An organization is appraised at
various CMMI maturity levels ranging from 1 to 5, with level 5 being the highest achievable level
(http://www.sei.cmu.edu/cmmi/index.cfm. accessed on December 29, 2009). The CMMI appraisal is being
adopted worldwide and is typically carried out by an independent consulting or auditing firm trained by
Software Engineering Institute. CMMI may improve a firm’s coordination capabilities required to manage
a set of geographically dispersed activities. CMMI infuses a firm with disciplined processes which pay off
by increasing a project success rate by impacting the rework, elapsed time and effort required to complete
the project (Gopal et. al, 2002).
Software offshoring firms seek the CMMI appraisal voluntarily as CMMI employs systems
engineering principles in software development resulting in improved performance for software firms.
Thus, CMMI appraisal has become the most well-known certification standard in the context of the IT
offshoring industry worldwide (Herbsleb, Zubrow, Goldenson, Hayes, & Paulk, 1997). The appraisal at
CMMI level 5 of a vendor software firm may send signals to the client of the former’s improved internal
mechanisms that may enhance the client’s trust in the capabilities of the Indian company (Gopal & Gao,
2009; Terlaak & King, 2006). Such an appraisal may result in legitimacy and efficiency gains to a vendor
software firm (Gopal & Gao, 2009). Improved disciplined internal processes of a vendor company allay
one of the top concerns of clients regarding the vendor’s ability to provide efficient and high-quality
services (Gopal & Gao, 2009). The service quality and operational efficiency of vendor software firms have
been listed as the top concerns of U.S. based client companies considering offshoring decisions (Couto,
Mani, Lewin, & Peeters, 2006).
63
Hence, CMMI appraisal may play an important role in improving the internal processes in Indian
software companies that have distributed project teams placed in multiple project sites and working on
different segments of a big software project. It is highly likely that the Indian software firms that have been
appraised at CMMI level 5 can effectively signal to the firms of their enhanced capabilities and can engage
in increased exports. On the other hand, firms that have not been appraised at CMMI level 5 may not be
able to convince client companies that effectively. Indian software firms with CMMI level 5 appraisals, by
virtue of possessing disciplined processes, may be in a different league of operations wherein they find
themselves competing with multinational IT services companies such as IBM. This may further increase
the propensity of firms with CMMI level 5 appraisal to locate in developed countries for market-seeking
purpose. Hence, the Indian software firms that CMMI level 5 appraisal will have an increased tendency to
locate in developed countries with market-seeking strategies than the firms that do not have the appraisal.
However, as the Indian software firms without the CMMI level 5 appraisal are likely to have already high
propensities to locate in a developed country with market-seeking strategies, the increased propensities of
firms with CMMI level 5 appraisal to locate in a developed country with market-seeking purposes is
unlikely to be much higher than the firms without the database. Hence, the effect may not be observable in
case of market-seeking strategies of Indian software firms.
On the contrary, the interaction effect is likely to be observable in the case of labor-seeking
behaviors of the two set of firms. The CMMI appraisal at the highest level of maturity may give Indian
software firms cost competitive advantages to run global delivery centers efficiently even in a developed
nation so as to increase the service levels to their clients. Hence the following is proposed.
Hypothesis 4: Among Indian software firms, the relationship between their strategy-type and
location choice will be moderated by CMMI level-5 appraisal, such that a firm with appraisal will
show more likelihood to locate in a developed country than one without appraisal when its
strategy-type is labor-seeking, but not when its strategy-type is market-seeking.
To test these hypotheses, data is hand-collected from Indian software companies on their
international location decisions between April 2000 and March, 2009. The next chapter details the data
collection procedure.
64
CHAPTER VII
VARIABLES AND DATA COLLECTION
The extant literature was reviewed in order to understand how other scholars have operationalized
the likely dependent, independent and control variables to be used in this dissertation.
1) Dependent variable
The dependent variable in this dissertation is categorical namely developed country (coded 1) and
emerging markets (coded 0). Data is collected from the annual reports of the surveyed firms and also from
internal web-portal, also called intranet of the company. The data, so collected, is scrutinized and validated
by the interviewed managers of each firm. This dissertation combines the IMF and Hoskisson et al. (2000)
categorization of countries to divide the countries into the two categories.
2) Independent Variables
Data were collected on four independent variables namely internationalization strategy, whether a
firm possessed electronic knowledge sharing database, wage bill per employee, and whether a firm was
appraised at CMMI level 5. Each of the independent variables is explained below.
(i) Internationalization Strategy: A literature review was conducted to understand how past studies
operationalized the international strategy variable. The summary of the findings of the literature review is
given below in Table 3.
Table 3: Variables related to Internationalization Strategies
Authors
and Year
Independent Variables Control Variables
Makino et
al. 2002
Labor-seeking, Asset-seeking, Market-seeking
(dummy = 1 if motivation present, 0 otherwise)
Data Source: Survey
Subsidiary Age (Year of foundation)
Entry Mode (dummy = 1 for joint-
venture, 0 for wholly-owned
subsidiary)
Firm size (number of employees)
Foreign sales (overseas sales / total
sales)
65
Galan et
al. 2007
Cost factors;
Market factors;
Infrastructure and technological factors;
Political and legal factors;
Social and Cultural factors.
Data Source: Survey.
Firm size (sales);
Business sector (services /
manufacturing).
Other studies that have used market-seeking, asset-seeking, and resource-seeking as independent
variables are Brouthers et al. (2008), Cuervo-Cazurra (2007), Makino et al. (2007), Sethi et al. (2003), and
Song (2002).
Song (2002) uses local sales ratio to operationalize market-seeking versus export-seeking
strategies. High local sales ratio indicates market-seeking, whereas low local sales ratio is part of export-
seeking strategy. Sethi et al. (2003) use principal component analysis to label market-attractiveness of a
country, which includes Gross National Product (GNP) and population of the country. As per Curevo-
Cazurra (2007), firm is market-seeking when its initial operation in a host country is sales, services, repair,
or distribution. Makino et al. (2007) use internationalization strategies as independent variables to study the
termination of international joint ventures. They operationalize these strategies with help of survey-
questions. Labor-seeking is access to low-cost inputs; market-seeking is following customers and market-
expansion; and strategic asset-seeking is research and development (R&D). Brouthers et al. (2008) use
industries of the firms to divide them into either market-seeking or resource-seeking types.
Based on the above, it may be observed that there are few studies that provide scales to measure
the constructs of market-seeking, asset-seeking, and resource-seeking strategies. Although Galan et al.
(2007) provides scales to measure the market-factors and cost-factors for internationalization, these scales
use location determinants to develop the items.
This dissertation uses categorical variable namely purpose to denote market-seeking and labor-
seeking strategy. When an Indian software firm located in a host country with a global delivery center, it is
used as a proxy for labor-seeking strategy and is coded 0. On the other hand, if a firm locates in a host
66
country with a sales office, it is coded as 1. There are instances when a company locates in a host country
with more than one strategy. For example, a firm may locate in Brazil with both market-seeking and labor-
seeking strategies. There is, however, never an instance when the firm opens both a sales and global
delivery center at the same time in the same location. Hence, each time a firm enters a country where it is
already located, the new entry is counted as a new location and its strategy is coded accordingly.
Besides internationalization strategy, data were also collected on the following resources of Indian
softweare firms. These resources constitute other independent variables, besides internationalization
strategy in the dissertation.
(ii) Knowledge-management by the Indian software vendor firm: The knowledge management by the
software company enhances the knowledge about the client and the projects. The existing knowledge stock
helps bid in a cost-effective manner for similar projects announced by other companies and new projects
announced by the existing clients. The variable namely “whether or not a company possesses knowledge-
management software in each of the 9 years of the study period” has been used as a proxy for Knowledge-
management by the Indian software vendor firm. It is a dichotomous variable, coded 0 in the year when the
firm did not have the electronic database and coded 1 in the year in which the firm had such a database.
(iii) Internal processes of the software vendor firm to deliver quality goods to the client: The steps
undertaken by Indian software firms to streamline their internal processes may have helped these
companies become successful. The better the internal processes, the better is likely to be the quality of the
products delivered to the client companies in a cost-effective manner. A variable namely “whether or not a
company possesses CMMI-level 5 (Capability Maturity Mode Integration level 5) has been used as proxy
to measure the level of internal processes of the software vendor firm to deliver quality goods to the client.
It is a dichotomous variable, coded 0 in the year when the firm did not have the appraisal and coded 1 in the
year in which the firm had the appraisal.
As per the industry experts, the state-of-the-art hardware and software platforms used by Indian
software firms may help Indian software companies deliver cost-efficient services and products to their
clients. However, a variable that may act as a proxy for the status of technological sophistication and
67
recency of the hardware and software platforms used by Indian software firms could not be collected in the
dataset that.
(iv) Low cost but qualified human talent with the software vendor firm: The cost and quality of the human
talent possessed by the software firm is one of the dimensions of the construct called process-related
resource. The low cost skilled manpower has been long recognized to result into the international growth of
Indian software firms (Pradhan, 2007). A variable namely “wage bill per employee” has been used as a
proxy for Low cost but qualified human talent with the software vendor firm. This is calculated by dividing
the total wage bill of the company in a year by its total number of employees in that year.
3) Control variables
(i) Variables related to Macroeconomic Determinants: In Table 4 below, some studies are listed that have
studied the impact of macroeconomic variables on location choice. The macroeconomic variables are used
as control variables in the dissertation.
Table 4: Macroeconomic Determinants.
Authors and
Year
Independent Variables
Li &
Guisinger,
1992
Market size (GDP)
Cultural distance: K&S 1988 method using scores from Hofstede’s 1980
Openness index
Oligopolistic reaction: number of other service firms / total number of service
firms
Growth in firm size: annual growth rate in sales
Home country business presence in a host country: book value of FDI in host
country
Data Source: archival
Woodward
& Rolfe
1993
Level of infrastructure development: Per capita GNP
Wage rate: Log of Adjusted 1985 hourly wage, including fringe benefits, for
unskilled operators in each country in 1985.
68
Political stability:1 = most stable; 0= least stable
Income tax incentives: 1 = more than 10 years; 0 = less than or equal to 10
years, 1984).
Free trade Zones: Log of area in acres
Profit repatriation restrictions: 1 = most restrictive; 0 = less restrictive, 1984
Exchange rate devaluation: Log of annual growth rate 1974-83
Land area: Log of area in sq kilometers
Manufacturing concentration: proportion of total labor force in manufacturing
in 1984
Inflation rate: Log of annual growth rate 1974-83
Transport costs: Log of transportation cost as proportion of trade
Unionization rate: Log of proportion of total labor force.
Data Source: archival
Loree &
Guisinger
1995
Performance requirements: survey-questions
Investment incentives: survey-questions
Political stability: index of political, financial and economic risk
Cultural distance: Kogut & Singh 1988 index
Income level in host country: Log of GDP per capita
Infrastructure: 22 variables factor analyzed to obtain 2 factors namely
communication and transportation
Wages: Log of average salary in host country
Developed country dummy: 1, 0 otherwise
Data Source: Survey and archival
Grosse &
Trevino,
1996
Bilateral trade: export and import to and from the USA Market-size of home
country: GDP
Per-capita income: GNP / population
69
Political risk: Source: Political risk services “Investment risk” ratings
Geographic distance: miles from home country capital to the host city.
Cultural distance: Gross & Goldberg index with scores from Hofstede.
Relative cost of borrowing: home country prime rate / prime rate in the USA
Exchange rate: level of bilateral exchange rate at the year end.
Relative rate of return: average stock market share price index
Data Source: Archival
Zhao & Zhu
2000
Market potential: GNP per capita; CFTR (ratio of the product sales income to
the average balance of circulating fund).
Export: export value as % of GDP
Efficiency: profit ratio: after-tax profit to total sales
Cost advantage: labor cost (average wage of a city), rental cost
Productivity level: net value added per employee
Technological level: ratio of technology development expenditure to GDP.
Infrastructure: index with 3 elements: transportation route (road space) per sq
kilometer; number of post office per capita; number of telephone lines per
capita.
Data Source: Archival
The dissertation uses the following variables to control for various macroeconomic variables.
Logofgdp: This variable is the natural log of GDP of the host country in current US Dollar for the
year in which an Indian software firm chose the host country as its location. The source of the data is world
development indicators by the World Bank. There were no missing data for this variable.
Logoftax: This variable is the natural log of highest marginal corporate tax rate in percentage in
the host country for the year in which an Indian software firm chose the host country as its location. The
source of data is world development indicators by the World Bank. The missing data were treated with help
of the KPMG’s Corporate and Indirect Tax Rate Survey 2008 available at
70
http://www.kpmg.com/SiteCollectionDocuments/Corporate-and-Indirect-Tax-Rate-Survey-2008v2.pdf.
This website was accessed on August 5, 2009.
Logofinflation: This variable is the natural log of percentage of annual inflation in consumer
prices in the host country for the year in which an Indian software firm chose the host country as its
location. The source of data is world development indicators by the World Bank. The data was transformed
by adding integer 6 to each number so as to convert all data pertaining to this variable into a positive
integer. The natural log was calculated on the transformed data. There were no missing data for this
variable.
ICT Development Index: This variable is an index of the development of internet and
telecommunication infrastructure of the host country in the year in which an Indian software firm chose to
locate there. The data were collected form the publication by the International Telecommunication Union
entitled “Measuring the Information Society, 2009”.
Logofpopulation: This variable is the natural log of the total population of the host country for the
year in which an Indian software firm chose to locate in the host country. The source of data is world
development indicators by the World Bank.
Logoffdistk: This is the natural log of the FDI stock in the host country in US Dollar at current
prices for the year in which an Indian software firm chose to locate in the host country. The source of the
data was the UNCTAD. The data were lagged by 1 year. The missing value were treated by multiplying the
base value in the year preceding the missing year by the simple average of the increase in FDI stock in
preceding three years for the particular host country.
ExchangeRate: This variable indicates the fluctuation in the exchange rate of the local currency of
the host country against the US Dollar compared to its exchange rate in the preceding year. A number less
than 1 indicates that the local currency depreciated against US Dollar in the year of interest and a number
greater than 1 indicates that the local currency appreciated against US Dollar in the year of interest. A
number equal to 1 indicates that the local currency did not change against the US Dollar. The extent of
decrease or increase from one indicates the extent of fluctuation of the local currency against the US
71
Dollar. The sources of data are International Monetary Fund (IMF) and the US Department of Commerce.
The data were lagged by a year.
Logoflabor: The variable is the natural log of the ratio of the labor rates between the capital city of
the host country and India (the labor rates in city of Mumbai). The data source is Prices & Earnings report
compiled by the Union Bank of Switzerland (UBS). The data are lagged by 1 year. The Prices & Earnings
reports are available for years 2003, 2005, 2006, 2008 and 2009. The missing data were treated by imputing
the same ratio number as was available for the closest year to the missing year. The labor rates may change
from year to year. However, the method of imputing the number available for the closest year to treat a
missing value was deemed appropriate.
HDI: This variable indicates the index of human development as published in the various volumes
of the Human Development Report compiled for the United Nations Development Program (UNDP). The
index assigns different countries a score out of 1. The higher the score assigned to a nation, the better the
human development in that nation is. The missing data were treated by imputing the same index number as
was available for the closest year to the missing year for the host country. As the human development index
of a host country is not envisaged to undergo significant changes in a shorter time frame, the method of
imputing the number available for the closest year to treat a missing value was deemed appropriate. The
data were lagged by 1 year.
(ii) Variables related to Institutions: Table 5 given below details some studies that have studied the impact
of institutions on location decisions of firms. The proxy variable for institution is used as a control variable
in the dissertation.
Table 5: Institutions
Authors
and Year
Independent Variables Control Variables
Globerman
& Shapiro
2003
Governance infrastructure index (GII):
from Kaufmann, Kraay, and Zoido-
Lobaton (1999)
Legal systems (common or civil): by
Host country size: Real GDP
Human Development Index
If the country uses a fixed
exchange rate: Dummy =1
72
La Porta et al. 1999, and from
university of Ottawa Faculty of Law
(used as substitutes of GII).
Index of Economic Freedom
Data source: Archival
Exchange rate: ratio of end period
trade-weighted exchange rate to
average for preceding 2 years in
terms of USD.
NAFTA effects: dummy (=1) for
Canada and Mexico
Omitted variables: Wage rate, tax
rates, openness of the country
(measured as (imports + exports) /
GDP, Cultural distance: dummy
variable (=1) for English.
Flores &
Aguilera,
2007
Market affluence: GDP in billions
Market magnitude: Population
Infrastructure: Total number of phone
lines
Wage level: average wages received
by manufacturing workers who work
40 hrs per week
Political institutions: 1 when a country
is democracy, 0 otherwise. Legal
Systems: 1 when same legal system, 0
otherwise
Cultural distance: modified K&S
(1988) index using Hofstede (1983),
controlling for language, geographic
distance, level of development, market
size and company size
Firm size: total no. of employees
Firm Performance: return to
investors in last 10 years;
Firm industrial sector: 2-digit SIC
coding scheme
Official language: dummy (1 if
English, 0 otherwise)
Geographic distance: between
capital cities of home and host
countries
Regional location: (19 UN
regional categories), dummy 1 if
country is in that region, 0
otherwise);
Economic development: dummy
(1 if economically developed, 0
73
Level of trust in a society
Data Source: Archival
otherwise).
Cuervo-
Cazurra,
2008
Pervasive corruption
Arbitrary corruption (using a method
by Uhlenbruck et al. 2006).
Data Source: Archival
Country size: GDP,
Population
Host country inflation
Geographic distance: natural log
of greater circle distance between
the home and host country centers
in miles Landlocked or Common
border: Common Language
Common Colonial histories
FDI limitations
Brouthers
et al., 2008
Market attractiveness for market-
seeking FDI: GDP per capita
Market attractiveness for labor-seeking
FDI: average monthly wage in
manufacturing
Market attractiveness for raw-material
seeking FDI: level of domestic energy
production
Corruption: rating from 0 – 10.
Data Source: Archival
International orientation of a
country: Trade / GDP
Ratio of Government consumption
to GDP
Based on the above, it may be noted that Kaufmann et. al’s scale to measure the institutional
development of a country suits this dissertation. The scale developed by Kaufmann et. al (1999) includes (i)
voice, political freedom and civil liberties; (2) political instability, terrorism and violence; (3) the rule of
law, crime, contract enforcement and property law; (4) level of graft and corruption in public and private
74
institutions; (5) extent of regulation and market openness; (6) measure of government effectiveness and
efficiency. The scale provides a comprehensive coverage of the institution-related location determinants
which may impact the location decision of a firm. Other recent study that has used Kaufmann et. al’s
(1999) scale to measure the governance structure of a country is by Bhardwaj, Dietz, and Beamish (2007).
There are other scales developed by scholars such as Henisz (2000) and La Porta, Lopez-de-
Silanes, Shleifer, and Vishny (1998). But the foci of these scales are narrower. For example, Henisz (2000)
focus is the level of political hazard in a country; whereas La Porta et. al (1998) finds the type of legal
system in a country. Delios and Beamish (2001) measure institutional environment of a country with help
of (i) extent of its political and economic risks, (ii) extent of its restrictions on foreign ownership, and (iii)
extent of intellectual property protection in the country. Steensma, Tihanyi, Lyles, and Dhanaraj (2005) use
regulatory quality, liberalization index, corruption index, government stability, government transparency
ranking, and private sector as percent of GDP to measure the institutional development of transition
economies.
Other studies such as Globerman and Shapiro (2003) have used the index of economic freedom
compiled by the Heritage Foundation. The index is updated for year 2009. Hence, it is used as a proxy to
denote institutions in various host countries. The variable is denoted by “institution”. This variable is an
index of the level of institutional development of a host country for the year in which an Indian software
firm chose to locate there. The missing data were treated by imputing the same index number as was
available for the closest year to the missing year for the host country. As the institutions are not envisaged
to undergo significant changes in a shorter time frame, the method of imputing the number available for the
closest year to treat a missing value was deemed appropriate.
(iii) Variables related to Cultural Distance: Table 6 below gives a summary of how other scholarly studies
in past have operationalized cultural distance.
Table 6: Cultural Distance
Authors and
Year
Independent
Variables
Control Variables
Benito & Cultural Distance: - Export share of the parent firm
75
Gripsrud 1992 Kogut & Singh
(1988) index
- Sales of the parent firm
- Mode of entry (greenfield / acquisition)
- Ownership % of FDI
Besides the above paper, many other scholars in the location literature have used cultural distance
as a control variable and it may be observed from these studies that the Kogut and Singh (1988) measure of
cultural distance is widely used to measure the cultural distance between the home and host countries (For
example, Berkema et al., 1996; Ellis, 2008; Tsang & Yip, 2007). Some recent studies have included other
variables along with the cultural distance index to overcome the criticism of using the Hofstede’s scores in
measuring the cultural distance (Shenkar, 2001). These variables are language and time zone.
Recently, Dow and Karunaratna (2006) provide a comprehensive measure of perceived psychic
distance between home and host countries. Their measurement of psychic distance includes differences in
culture (Kogut & Singh, 1988), language, political system, religion, industrial development (GDP per
capita, passenger car per 1000, energy consumption per capita, population living in urban cities,
manufacturing as % of GDP, telephone as % of GDP), education, time zones.
As this dissertation measures the levels of institutional and infrastructure development of a host
country with help of other indices, use of psychic distance as measured by Dow and Karunaratna (2006)
might result in multicollinearity in the dataset as Dow and Karunaratna (2006) include industrial
development and political systems as part of their index. To avoid multicollinearity, this dissertation
includes cultural distance as measured by the Kogut and Singh (1988) index using Hofstede’s score (2001).
At the same time, this dissertation controls for variables such as language and time zone in line with
suggestions made by Shenkar (2001). The following measures are used as proxy for cultural distance
between India and host countries.
Time Diff: This variable is the absolute difference in time in hours between the capital city of the
host country and the city of the headquarters of the Indian software firm that chose to locate in the host
country of interest. The source of data is the World Clock.
76
Isenglishspoken: This is a dichotomous variable and takes the value equal to 2, if English is
spoken in a country, otherwise it assumes a value of 1. The source of the data is the CIA Fact Book.
CulturalDistance: This variable indicates the cultural distance between the host country and India.
The Kogut & Singh (1988) measure of cultural distance was used in the dissertation.
(iv) Variables related to Geographic and Economic distance: In extant literature, geographic distance is
measured by the physical distance between capital cities or some focal cities of the home and host countries
(Erramilli et al., 1999). This dissertation operationalizes geographic distance in the same manner. The
variable is called LogofGeoDistance and is the natural log of the geographic distance in nautical miles
between the capital city of the host country and the city of the headquarters of the Indian software firm that
chose to locate in the host country of interest. The source of data is the World Atlas.
Economic distance is used by Tsang and Yip (2007) as a location determinant affecting the
location decisions of firms. They operationalize economic distance as difference in natural logs of real
GDP per capita of Singapore and the host country. As GDP per capita is included as one of the variables in
this dissertation to reflect the market size of a host country, economic distance is not included separately as
a variable in this dissertation.
(v) Variables related to International Experience and Top Management Characteristics (TMT): Table 7
below summarizes how past studies have operationalized the variables related to international experience
and top management characteristics.
Table 7: International Experience and TMT
Authors and
Year
Independent Variables Control Variables
Sambharya,
1996
International experience in years; proportion
of top managers with international
experience;
homogeneity of international experience.
Firm size (Sales)
TMT size
Tihanyi et
al., 2000
Age (in years);
Team tenure;
Firm profitability (Prior ROA in 3
years);
77
Elite education (categorical variable);
International experience (through
assignment or education);
Educational background;
Functional background (seven categories).
Firm size (number of employees);
TMT team size.
The data pertaining to top management international experience and tenure could only be collected
for the year 2009 in the dissertation. As other time variant variables in the dissertation have been collected
for each time period, it was considered better to not control for TMT experience and its tenure in the
dissertation. In the model, there are other control variables such as international experience and firm age of
the firm that may overcome, to some extent, the lack of the TMT variables in the model.
(vi) Variables related to Agglomeration: Rose and Ito (2008) operationalize oligopolistic reaction as
number of rivals, aside from the focal firm, with a subsidiary in the host country prior to the focal firm’s
entry. A measure of imitation as used by Filatotchev et al. (2007) is the cumulative Indian FDI in the host
country. However, the lack of reliable data denoting the number of already operating Indian firms or year-
wise cumulative Indian FDI in each of the host country precludes this dissertation from controlling for
agglomeration factor. Agglomeration becomes an important variable if firms internationalize with asset-
seeking strategy. Since, Indian software companies are not internationalizing with asset-seeking strategy,
inability to control for agglomeration is not likely to have impact on the validity of the statistical results
obtained in the dissertation.
A list of dependent, independent, firm-specific and host country-specific control variables used in
this dissertation and their sources are given in the following tables.
Table 8: Data-source for Dependent and Independent Variables
Variable Name Source of data
Location (Dependent Variable) Firm annual reports and intranet
Purpose (Independent variable) Firm annual reports and intranet
CMMI (Independent variable) Firm annual reports, firm archives such as presentations,
78
newspaper items, firm websites, intranet, stock exchange
archives
KMSW Firm annual reports, firm archives such as presentations,
newspaper items, firm websites, intranet, stock exchange
archives
Wage Bill per employee Annual report
Table 9: Data-source for Firm-specific Control Variables
Variable Name Source of data
Reentry Self-compiled based on the data on location.
EntryMode Firm annual reports and intranet
LogofSales Firm annual reports
Profit Firm annual reports
ROS Firm annual reports
Cash Firm annual reports
Logtotalempl Firm annual reports, firm archives such as presentations, newspaper
items, firm websites, intranet, stock exchange archives
FirmAge Firm annual report
BusinessGroup Annual report
Indiagdc Firm annual reports, firm archives such as presentations, newspaper
items, firm websites, intranet, stock exchange archives
Internationalexp Annual report
Table 10: Data-source for Host Country-specific Control Variables
Variable Name Source of data
Logofgdp World Bank
79
Logoftax World Bank, KPMG’s Corporate and Indirect Tax Rate Survey 2008
Logofinflation World Bank
ICT International Telecommunication Union
Logofpopulation World Bank
Institution Heritage Foundation
LogofGeoDistance World Atlas
TimeDiff World Clock
HDI UNDP
Isenglishspoken CIA Fact Book
Logoffdistk UNCTAD
ExchangeRate IMF and US Department of Commerce
Logoflabor Prices & Earnings report by the Union Bank of Switzerland
CulturalDistance Hofstede’s scores
Data Collection
In the summer of 2009, the top 32 publicly-listed Indian software companies were contacted to
hand-collect data to study their international location choices made by them between April 2000 and March
2009. The data were compiled from several publicly available sources such as company websites, news
items, initial public offer documents of the firms, annual reports, presentations to media and investors, and
also from firms’ internal archival documents and their intranets. Data-sources such as the ones compiled by
Ministry of Finance, Government of India are inadequate to analyze the research questions outlined in the
dissertation as these data-sources do not provide much information on foreign subsidiaries of Indian firms
(Pradhan, 2007), and it necessitated the compilation of the dataset.
The first year of data collection is taken as 2000 as significant policy changes by the government
of India resulted in rapid outward FDI since then (UNCTAD, 2004). The variables pertaining to the firms
were collected at two levels. Some variables could be collected at the level of each location decision made
by the firm within a year, so these variables change with each location decision made by the firm. The
80
remaining variables could be collected at the level of each year of the study. Consequently, these variables
change only year to year and remain invariant for all the location decisions made within a year by a firm.
The ideal scenario would have been the collection of all time-variant data at the level of location decision.
However, paucity of the firm records at the level of location decision for all variables restricted the
collection of data at two levels namely location decision and year level. The list of variables and the level at
which these variables were collected are tabulated below. Table 11 below also gives a brief description of
the variables and lists if a variable is continuous or dichotomous.
Table 11: Level of Firm-specific Data Collection
Variable Name Level at which the
data collected
Variable Description Continuous or
Dichotomous
Location (Dependent
variable)
Location decision Name of the country where
the firm located
Dichotomous: 1 for
developed country and
0 for emerging market
Purpose (Independent
variable)
Location decision Strategy of the firm Dichotomous: 1 for
market-seeking and 0
for labor-seeking
CMMI (Independent
variable)
Yearly level If the firm possessed CMMI
level 5 appraisal at the end
of each year
Dichotomous: 1 if yes,
0 if no.
KMSW (Independent
variable)
Yearly level If the firm installed a
knowledge management
software at the end of each
year
Dichotomous: 1 if yes,
0 if no.
WageBillperemployee
(Independent variable)
Yearly level Total wage bill divided by
total employees
Continuous
Reentry Location decision Whether the firm already
has presence in the country
Dichotomous: 1 if
already present, 0 if not
81
EntryMode Location decision The entry mode for the
location decision
Dichotomous: 1 if
wholly-owned
subsidiary, 0 if not
LogofSales Yearly level Natural log of the revenue
of the firm in each year
Continuous
Profit Yearly level Profit after tax of the firm in
each year
Continuous
ROS Yearly level Net Profit Margin of the
firm in each year
Continuous
Cash Yearly level Net Cash at the end of each
year
Continuous
Logtotalempl Yearly level Natural log of total
employees at the end of
each year
Continuous
FirmAge Yearly level The number of years
elapsed since the year of
firm incorporation
Continuous
BusinessGroup Firm level If the firm is affiliated with
a business group
Dichotomous: 1 if yes,
0 if no.
Indiagdc Yearly level Number of Indian cities in
which the firm had a Global
Delivery Center at the end
of each year
Continuous
Internationalexp Yearly level Years elapsed since the first
year of internationalization
Continuous
82
The top 32 publicly listed Indian software firms from which the data is collected have made 650
international location choices in 67 different countries since April 2000 till March 2009. There are 110
publicly listed Indian software companies as on June, 2009, making the sample 29% of the publicly listed
IT and software organizations in India. NASSCOM, the industry association of Indian software companies,
categorizes these companies as Tier-I and Tier-II software companies. The top 10 Indian software
organizations comprise tier-I and the remaining companies form tier-II of the NASSCOM classifications.
The data has all tier-I companies and 22 tier-II companies. According to Gopal & Gao (2009), Dataquest
magazine in 2005 reports that Indian software exports are highly concentrated with the top 5 firms
accounting for 30% while the top 20 firms account for 53% of the total software exports from India.
Another study by Pradhan (2007) mentions that there were 165 Indian information and software technology
multinational companies in year 2006 and the top 60 firms among these 165 firms accounted for 72.6% of
the total overseas investments by these companies. Thus, a high concentration of exports and international
location decisions in the Indian software industry turns the small number of 32 firms to represent
adequately the location decisions made by Indian software companies, though it forms 19.4% of the 165
information and software technology companies that are reported to have established operations overseas.
It has been recognized that data collection poses problems in the context of emerging economies
(Hoskisson et. al, 2000). Given the scenario, it is likely that the sample may be biased towards the more
active software firms.
The next step in the data collection process was to understand that it is the process-related
resources that have assisted Indian software companies go international more than any other resources. To
accomplish this, two employees working at the rank of General Manager or above were contacted in each
of the 32 software firms at two different time periods and were separately administered a short semi-
structured interview. Many of the contacted managers were part of the top management team of the
companies such as head of human-resources, finance, or marketing departments, or executive directors,
CEOs or promoters of the companies. These employees were given a list of the following six resources.
Non-market resources
Ethnic relational resources
83
Business relational resources
Process-related resources
Market knowledge resources
Natural asset based resources
The meaning of each of the resources was explained to them. They were, then, asked to rate each
of the resources on a scale of 1-5 (where 1 meant not at all helped and 5 meant helped the most in
internationalization) if that particular resource helped their firm achieve its present extent of international
geographical expansion. The data analysis suggests that it is the process-related resources that have helped
Indian software companies go international more than any other resource. Out of the 64 interviewed
managers, 81.25% (52 in number) of them rate process-related resources as the most important resource
that boosted their firms’ internationalization effort.
Host country related variables
There are 67 countries in which the 32 sampled Indian software firms have located between April
2000 and March 2009 for market-seeking and labor-seeking purposes. These countries are a mix of
developed countries and emerging markets. The control data for each of the host country was collected at
the level of the location decision as given in the following Table 12.
Table 12: Level of Host Country-specific Data Collection
Variable Name Level at which the
data collected
Variable Description Continuous or
Dichotomous
Logofgdp Location decision Natural log of GDP of the
country in the year in which a
firm located there
Continuous
Logoftax Location decision Natural log of tax rate of the
country in the year in which a
firm located there
Continuous
Logofinflation Location decision Natural log of annual inflation
rate of the country in the year in
Continuous
84
which a firm located there
ICT Location decision An index of development of
Information and
telecommunication
infrastructure in the country in
the year in which a firm located
there
Continuous
Logofpopulation Location decision Natural log of population of the
country in the year in which a
firm located there
Continuous
Institution Location decision Institutional score out of 100 of
the country in the year in which
a firm located there
Continuous
LogofGeoDistance Location decision Natural log of geographic
distance of the country in which
a firm located from India
Continuous
TimeDiff Location decision Time difference in hours
between the country of location
and India
Continuous
HDI Location decision Human Development Index
score (out of 1) of the country
in the year in which a firm
located there
Continuous
Isenglishspoken Location decision If English is spoken or not in
the country in which a firm
located
Dichotomous: 1 for
yes, 0 for no.
Logoffdistk Location decision Natural log of FDI stock of the Continuous
85
country in the year in which a
firm located there
ExchangeRate Location decision Ratio of exchange rates of local
currency of the host country to
US dollar in the year in which a
firm located in the host country
Continuous
Logoflabor Location decision Natural log of ratio of labor
rates of the host country to
those in India in the year in
which a firm located there
Continuous
CulturalDistance Location decision Cultural Distance between the
host country and India
Continuous
In the following chapter, the statistical analysis is applied on these variables with help of
hierarchical linear modeling or random coefficient growth modeling using software HLM 6.2.
86
CHAPTER VIII
METHODOLOGY, RESULT, AND DISCUSSION
Hierarchical Linear Modeling (HLM) is an appropriate methodology to analyze a hierarchical
data, longitudinal data, and test hypotheses about how variables measured at one level affect relations
occurring at another level (Raudenbush & Bryk, 2002). The data collected for the dissertation has a
hierarchical structure as the location decisions made by the 32 sampled firms are nested within each firm
and each of these firms make several international location decisions over the period of the data collection.
Thus, a location decision made by a firm may not be independent of the other decisions made by the same
firm in previous years as multiple observations from the same source tend to be correlated (Bliese &
Ployhart, 2002). Further, the resulting design is unbalanced as the number of location decisions made by a
firm may differ from those made by other firms. For a nested dataset such as this with dichotomous
dependent variable, logistic regression is not an appropriate methodology as the assumption of independent
observations is violated in this data and hence estimates are likely to be biased.
Application of non-multilevel analyses such as logistic regression in the scholarly area of entry
mode research which is a multilevel research question has been recently criticized by Arregle et al (2006)
and they suggested the use of HLM to address entry mode related research questions. Ignoring the
multilevel nature of the dataset creates conceptual and statistical limitations such as a risk for validity and
robustness of the results (Arregle et. al, 2006). Similarly, in the scholarly field of FDI location decisions
which poses a multilevel research question, the use of HLM is appropriate but has hardly been used.
Though the data is longitudinal by nature, time series analysis is not an appropriate
methodological tool as it requires observing each firm for a larger period of time such as 50 or more times
(Tabachnik & Fidell, 2007).
As the data in the dissertation has 3 levels namely location decision level, year level and the firm
level and since the outcome variable is dichotomous namely developed nation or emerging market; a 3-
level Bernoulli HLM is an appropriate methodological tool to analyze this dataset. Bernoulli HLM has been
suggested for entry mode research where the outcome variable is dichotomous (Arregle et. al, 2006).
87
For building models at each of the three levels, the approach used by Arregle et. al, 2006) is
followed. The model building begins with an unconditional model and in the second step, all level-1
control variables are entered (though Arregle et. al, 2006 enter the variables one at a time) and only those
control variables are retained that are found to be statistically significant with p < 0.05 in line with Arregle
et. al (2006). In the third step, dichotomous independent variable named purpose that denotes strategy
(whether market-seeking or labor-seeking) is entered for each of the location decisions. This completes the
level 1 model. Once the level 1 mode is in place, the variables at the level 2 should be introduced (Bliese &
Ployhart, 2002). Hence, the fourth step is to build the level 2 model.
For building the level 2 model; the control variables namely time, log of sales, cash, and return on
sales are entered together. As in the step 2, only those variables that are found to be statistically significant
are retained for further analysis. In the fifth step, the independent variables pertaining to firm resources are
entered. This completes the level 2 model. In the sixth step, the variable called BusinessGroup is entered at
the level of the firm. This step introduces the third and the last level in the model. This step completes the
introduction of all linear effects in the 3-level model.
In the seventh and the last step, the interaction terms of firm resources with firm strategy are
entered one by one. The interaction terms are introduced not as a multiplicative term but by varying the
slope of the purpose variable by the firm resources. The statistical significance of the interaction term is
checked.
The step-by-step approach from univariate to bivariate to trivariate adopted in the dissertation is
recommended for the multilevel model as a saturated level-1 model is helpful only when the sample size is
very large (Raudenbush & Bryk, 2002). At each step, decision whether or not to retain a variable in the
model for further analysis is undertaken (Arregle et. al, 2006).
At each step, two decisions need to be taken before running the analysis. First, if a variable needs
to be entered as uncentered, group-centered, or grand-centered at the level 1 and level 2. Second, if the
intercept or slope pertaining to the variable needs to be specified as fixed, random, or non-randomly
varying. Enders & Tofighi (2007) suggest group-centering the variables when the primary interest is to (i)
check the relationship between a level-1 predictor and the outcome variables and (ii) when the cross-level
88
interactions are of interest. This dissertation primarily looks at the relationship between the firm strategy
(denoted by the variable called purpose which is a level 1 variable) and the firms’ international location
decision. The dissertation also seeks to understand if the interaction of firms’ resources which are level 2
variables with purpose (a level 1 variable) is statistically significant. So, in line with the recommendations
of Enders & Tofighi (2007), all variables are group-centered.
On the question of random, fixed, or non-randomly varying specification of the intercept and
slopes, variables are entered as randomly varying initially, and the estimated reliability and pace of model
convergence, i.e. whether quick or slow to converge, are checked. Raudenbush & Bryk (2002) suggest
replacing random variation with fixed variation for reliability estimates of less than 0.1.
The 3-level Bernoulli HLM gives the option to use Laplace iterations. Laplace estimation gives
the deviance ratio to test the overall model fit at each subsequent model (Raudenbush & Bryk, 2002).
As explained elsewhere, the 32 Indian software firms in the dataset have made 650 international
location decisions between April 2000 and March 2009. These 650 location decisions make level 1 of the
dataset. The 9 years during which the data pertaining to the firm resources has been collected for the 32
firms form level 2 of the dataset. There are 201 observations at level 2 and it shows that not every firm
makes international location decision in each of the year of the study and many firms make multiple
location decisions in a year. The variable business-group is time-invariant for the sampled 32 firms and it
forms level 3 (firm-level) of the dataset. Thus, the dataset is unbalanced as number of location decisions per
firm varies.
Correlation Matrix and Variance-Inflation Factor (VIF) Results
Before the actual building of HLM Bernoulli model begins; correlations, and VIFs of the variables
are checked. The correlation matrix and collinearity statistics are given in Table 13 and Table 14 below.
Table 13: Correlation Table
1 2 3 4 5 6
1
Developed 1
2
Rentry .280** 1
3
Purpose .311** -.098* 1
89
4
EntryMode .029 .044 .117** 1
5
LogofSales -.314** .120** -.261** .061 1
6
Profit -.280** .134** -.287** .145** .776** 1
7
ROS -.068 .106** -.058 .094* .322** .277**
8
Cash -.217** .068 -.197** .094* .539** .737**
9
logtotalempl -.303** .117** -.275** .029 .975** .776**
10
logofswemp -.308** .107** -.277** .029 .969** .782**
11
CMMI -.174** .095* -.217** -.094* .457** .378**
12
KMSW -.224** .024 -.233** -.164** .501** .361**
13
BusinessGroup -.063 .159** -.042 .211** .228** .258**
14
IndiaGDC -.294** .121** -.253** .223** .805** .810**
15
WageBillperemployee -.132** .058 -.085* -.092* .237** .129**
16
Internationalizationexp -.238** .142** -.190** .100* .764** .639**
17
logofGDP .332** .464** .056 .015 -.226** -.123**
18
logoftax .251** .183** .168** .031 -.176** -.113**
19
FirmAge -.254** .165** -.183** .243** .787** .696**
20
logofInflation -.326** -.066 -.051 -.047 .180** .139**
21
ICT .708** .293** .208** .044 -.116** -.109**
22
logofPopulation -.162** .314** -.110** -.002 -.119** -.021
23
Institution .690** .310** .188** .046 -.272** -.218**
24
logofGeoDistance .235** .329** -.080* .066 -.061 .008
25
TimeDiff .259** .403** .003 .076 -.130** -.046
26
IsEnglishSpoken .138** .157** .116** .011 -.142** -.114**
27
logofFDItock .416** .464** .050 .022 -.216** -.126**
28
ExchangeRate .199** .055 .053 .012 -.018 -.047
29 logoflabor .804** .282** .312** .033 -.365** -.332**
30 CulturalDistance .364** .111** .107** .083* -.062 -.076
31 HDI .779** .257** .214** .059 -.208** -.201**
90
Table 13: Correlation Table (Cont’d)
7 8 9 10 11 12
7 ROS 1
8 Cash .210** 1
9 logtotalempl .309** .537** 1
10 logofswemp .297** .543** .996** 1
11 CMMI .207** .292** .492** .501** 1
12 KMSW .075 .255** .506** .526** .465** 1
13 BusinessGroup .036 -.059 .234** .223** .120** -.027
14 IndiaGDC .293** .440** .817** .830** .414** .508**
15 WageBillperemployee .027 .250** .193** .200** .186** .233**
16 Internationalizationexp .175** .356** .767** .755** .160** .335**
17 logofGDP -.124** -.056 -.214** -.212** -.074 -.090*
18 logoftax -.063 -.065 -.176** -.170** -.054 -.073
19 FirmAge .325** .335** .774** .765** .245** .247**
20 logofInflation .014 .127** .176** .177** .162** .136**
21 ICT -.081* -.072 -.098* -.102** .031 -.078*
22 logofPopulation -.086* .026 -.117** -.112** -.041 -.005
23 Institution -.077* -.176** -.260** -.265** -.150** -.152**
24 logofGeoDistance -.043 .000 -.051 -.049 -.069 -.007
25 TimeDiff -.069 -.018 -.123** -.122** -.079* -.026
26 IsEnglishSpoken -.082* -.057 -.144** -.148** -.028 -.028
27 logofFDItock -.104** -.057 -.204** -.204** -.087* -.105**
28 ExchangeRate .071 -.042 -.014 -.009 .012 .037
29 logoflabor -.068 -.268** -.356** -.357** -.296** -.226**
30 CulturalDistance .037 -.050 -.062 -.060 -.086* -.094*
31 HDI -.034 -.141** -.196** -.199** -.141** -.161**
Table 13: Correlation Table (Cont’d)
13 14 15 16 17 18
13 BusinessGroup 1
14 IndiaGDC .385** 1
15 WageBillperemployee -.240** .039 1
91
16 Internationalizationexp .154** .695** .133** 1
17 logofGDP -.037 -.186** .008 -.130** 1
18 logoftax -.029 -.124** .005 -.138** .411** 1
19 FirmAge .380** .776** .048 .835** -.166** -.128**
20 logofInflation .016 .168** .116** .104** -.167** .057
21 ICT -.049 -.145** -.041 -.090* .360** .065
22 logofPopulation -.010 -.071 .037 -.054 .824** .337**
23 Institution -.047 -.232** -.102** -.186** .242** -.096*
24 logofGeoDistance -.019 -.015 .041 .022 .600** .207**
25 TimeDiff -.019 -.065 .014 -.047 .609** .311**
26 IsEnglishSpoken -.049 -.118** -.016 -.140** -.010 .048
27 logofFDItock -.067 -.187** .004 -.120** .828** .158**
28 ExchangeRate -.015 .012 -.008 -.014 .061 .012
29 logoflabor -.055 -.309** -.117** -.241** .463** .183**
30 CulturalDistance -.019 -.060 -.016 -.022 .194** .017
31 HDI -.051 -.210** -.080* -.130** .373** .112**
Table 13: Correlation Table (Cont’d)
19 20 21 22 23 24
19 FirmAge 1
20 logofInflation .113** 1
21 ICT -.132** -.298** 1
22 logofPopulation -.074 .027 -.128** 1
23 Institution -.202** -.393** .655** -.160** 1
24 logofGeoDistance -.011 -.054 .315** .452** .317** 1
25 TimeDiff -.066 -.040 .288** .465** .408** .826**
26 IsEnglishSpoken -.125** .121** .060 -.057 .438** .051
27 logofFDItock -.170** -.231** .497** .609** .477** .591**
28 ExchangeRate -.016 -.239** .125** -.053 .135** .057
29 logoflabor -.260** -.384** .613** .020 .597** .398**
30 CulturalDistance -.034 -.245** .527** -.064 .207** .366**
31 HDI -.162** -.462** .820** -.155** .661** .352**
92
Table 13: Correlation Table (Cont’d)
25 26 27 28 29 30 31
25 TimeDiff 1
26 IsEnglishSpoken .368** 1
27 logofFDItock .583** .087* 1
28 ExchangeRate -.035 -.056 .132** 1
29 logoflabor .360** .034 .495** .211** 1
30 CulturalDistance .138** -.322** .183** .149** .510** 1
31 HDI .310** -.011 .467** .234** .774** .531** 1
** p < 0.01, * p < 0.05 (2-tailed test)
Table 14: Collinearity Statistics
Collinearity Statistics
Tolerance VIF
(Constant)
Rentry .581 1.722
Purpose .714 1.401
EntryMode .706 1.417
LogofSales .040 25.272
Profit .133 7.533
ROS .717 1.394
Cash .297 3.366
logtotalempl .005 189.353
logofswemp .006 162.049
CMMI .496 2.016
93
FirmAge .149 6.717
KMSW .468 2.138
BusinessGroup .564 1.774
IndiaGDC .125 8.017
WageBillperemployee .738 1.354
Internationalizationexp .183 5.455
logofgdp .038 26.435
logoftax .646 1.549
logofInflation .659 1.518
ICT .184 5.430
logofpopulation .045 22.096
logofGeoDistance .187 5.351
TimeDiff .166 6.006
HDI .119 8.405
IsEnglishSpoken .495 2.020
logoffdistk .187 5.347
ExchangeRate .838 1.194
logoflaborrate .232 4.301
CulturalDistance .404 2.477
The correlation matrix shows that many variables are highly correlated (r > 0.70) with each other
and with the dependent variable. The VIFs for these variables are also high (VIF > 4.0), which may result
in multicollinearity and may bias the statistical results. The variables that are highly correlated with the
94
dependent variable are not retained in the model for further analysis. The variables that are highly
correlated with the dependent variable are ICT (r = 0.71, p < 0.01), Institution (r = 0.70, p < 0.01), HDI (r =
0.78, p < 0.01), logoflabor (r = 0.81, p < 0.01). The VIFs for these variables were also greater than 4 and
tolerance less than 0.2 as can be seen in the table 14 above, indicating that multicollinearity among the
variables may be present. This is likely to be the case, as developed nations receive higher scores for their
institutions, ICT and HDI than emerging markets. The labor cost is also higher in developed ountries than
emerging markets. So, these control variables are dropped from further analysis.
The variables on the left side of the equation that are highly correlated with each other are as
follows.
logofsales with profit (r = 0.78, p < 0.01), logtotalempl (r = 0.98, p < 0.01), logofswemp (r = 0.97,
p < 0.01), firm age (r = 0.79, p < 0.01) indiagdc (r = 0.81, p < 0.01), and internationalexp (r =
0.76, p < 0.01);
logofgdp with logofpopulation (r = 0.82, p < 0.01), and logoffdistk (r = 0.83, p < 0.01);
loggeodistance with timediff (r = 0.83, p < 0.01).
One variable from each set of highly correlated independent variables are included in further
analysis and model building. Hence, logofsales, logofgdp, timediff are retained while other variables are
dropped from the HLM model. The VIFs for the dropped variables were also greater than 4 as can be seen
in table 14.
The list of independent and control variables entered in the HLM Bernoulli model is as follows.
The dependent variable is dichotomous (1 = developed country, 0 = emerging market).
Table 15: List of Variables in HLM
Level of the model Control Variable Independent Variable
Level 1 Reentry
Entrymode
Logofgdp
Logoftax
Logofinflation
95
Timediff
isenglishspoken,
Exchangeratevolatility
Culturaldistance
purpose
Level 2 Logofsales
ROS
Cash
Time
CMMI
KMSW
wagebillperemployee
Level 3 BusinessGroup
The descriptive statistics (mean and standard deviation), correlation matrix, and VIF for the above
variables are calculated again and are given in the following tables.
Table 16: Descriptive Statistics
Mean Std. Deviation
Developed .734 .442
Rentry .408 .492
Purpose .785 .411
EntryMode .766 .424
LogofSales 7.176 1.706
ROS .164 .122
Cash 662.505 1402.333
96
CMMI .574 .495
KMSW .669 .471
BusinessGroup .474 .500
WageBillperemployee .076 .032
logofGDP 27.434 1.729
logoftax 3.428 .284
logofInflation 2.274 .441
TimeDiff 4.744 2.790
IsEnglishSpoken .642 .480
ExchangeRate 1.006 .114
CulturalDistance 1.474 .742
Out of 650 international locations chosen by Indian software companies, 510 locations have been
chosen for a market-seeking strategy, whereas 140 locations have been chosen for a labor-seeking strategy.
Out of the 32 firms, the maximum location decisions made by a firm between April 2000 and March 2009
are 66, while the minimum location decisions made are 3 by another firm in the same period. Out of 650
international location decisions, 498 have been done with a fully-owned subsidiary entry mode. Out of 32
firms, 13 firms belonged to a business group while the remaining 19 firms did not have such an affiliation.
Table 17: Correlation Table of Variables in HLM
1 2 3 4 5 6 7 8 9
1 Developed 1
2 Rentry .280** 1
3 Purpose .311** -.098* 1
4 EntryMode .029 .044 .117** 1
5 LogofSales -.314** .120** -.261** .061 1
6 ROS -.068 .106** -.058 .094* .322** 1
7 Cash -.217** .068 -.197** .094* .539** .210** 1
8 CMMI -.174** .095* -.217** -.094* .457** .207** .292** 1
97
9 KMSW -.224** .024 -.233** -.164** .501** .075 .255** .465** 1
10 BusinessGroup -.063 .159** -.042 .211** .228** .036 -.059 .120** -.027
11
WageBillperemp
loyee
-.132** .058 -.085* -.092* .237** .027 .250** .186** .233**
12 logofGDP .332** .464** .056 .015 -.226** -.124** -.056 -.074 -.090*
13 logoftax .251** .183** .168** .031 -.176** -.063 -.065 -.054 -.073
14 logofInflation -.326** -.066 -.051 -.047 .180** .014 .127** .162** .136**
15 TimeDiff .259** .403** .003 .076 -.130** -.069 -.018 -.079* -.026
16
IsEnglishSpoke
n
.138** .157** .116** .011 -.142** -.082* -.057 -.028 -.028
17 ExchangeRate .199** .055 .053 .012 -.018 .071 -.042 .012 .037
18 CulturalDistance .364** .111** .107** .083* -.062 .037 -.050 -.086* -.094*
Table 17: Correlation Table of Variables in HLM (Cont’d)
10 11 12 13 14 15 16 17 18
10 BusinessGroup 1
11 WageBillperemployee -.240
**
1
12 logofGDP -.037 .008 1
13 logoftax -.029 .005 .411
**
1
14 logofInflation .016 .116
**
-.167
**
.057 1
15 Tim e Diff -.01 9 .01 4 .609
**
.311
**
-.040 1
16 IsEnglishSpoken -.049 -.016 -.010 .048 .121
**
.368
**
1
17 ExchangeRate -.015 -.008 .061 .012 -.239
**
-.035 -.056 1
18 CulturalDistance -.019 -.016 .194
**
.017 -.245
**
.138
**
-.322
**
.149
**
1
It can be observed from the correlation matrix in Table 17 that none of the variables are highly
correlated with r < 0.7.
Table 18: Collinearity Statistics of Variables in HLM
Tolerance VIF
Rentry .635 1.574
Purpose .824 1.214
EntryMode .865 1.156
LogofSales .394 2.535
ROS .834 1.199
Cash .639 1.565
CMMI .671 1.491
KMSW .609 1.641
98
BusinessGroup .743 1.346
WageBillperemployee .819 1.220
logofGDP .410 2.441
logoftax .774 1.293
logofInflation .826 1.211
TimeDiff .435 2.298
IsEnglishSpoken .580 1.726
ExchangeRateVolatility .909 1.100
CulturalDistance .738 1.356
It can be seen from the collinearity statistics that none of these variables has VIF > 4.0 and
tolerance < 0.20 as shown in Table 18. So, these variables are not multicollinear with each other.
Result
The Laplace iterations of the intercept-only or the unconditional model showed reliability
estimates of greater than 0.1 (reliability estimates of random level 1 coefficient = 0.296, reliability
estimates for random level 2 coefficient = 0.316) and hence the intercept can be taken as varying randomly
(Raudenbush & Bryk, 2002). The p-value (p < 0.001) for fixed effect indicates that G000 is significantly
different from 0. The deviance for the unconditional model is 1917.6509 with number of estimated
parameters = 3 as shown in model 1 of Table 22.
Summary of the model specified in equation format is as given below:
Level-1 Model
Prob(Y=1|B) = P
log[P/(1-P)] = P0
Level-2 Model
P0 = B00 + R0
Level-3 Model
B00 = G000 + U00
99
Intra-class correlation (ICC) for the 3-level Bernoulli model was calculated using the formula
suggested for a 2-level Bernoulli model by Snijders & Bosker (1999, page 224) with the ICC calculation
approach adopted for 3-level models with continuous dependent variables. ICC for level 3 is 0.08, showing
that 8% of the variance is at level 3. The ICC for level 3 and level 2 is 0.26, showing that 26% of the
variance is explained at level 2 and level 3 together. The ICC for level 2 is 0.21, indicating that 21% of the
variance is explained at level 2 alone. These ICC values are high given the fact that in educational research,
ICC values between 0.05 and 0.2 are common (Snijders & Bosker, 1999). Thus, a 3-level model is
appropriate for analyzing the dataset.
Step 2 of the model-building:
All level 1 control variables as given in Table 19 are entered in a block in this step. The following
table shows the statistical significance of each of these variables as obtained from Laplace iterations.
Table 19: Level 1 Control Variables
Variable Coefficient T-ratio Degrees of freedom p-value
Intercept 2.321706 4.78 31 < 0.001
Reentry 1.283209 1.360 640 0.175
entrymode 1.985013 0.605 640 0.545
logofgdp 0.387813 0.837 640 0.403
logoftax 2.076442 1.584 640 0.113
logofinflation -3.024463 -3.284 640 0.001
Timediff -0.089852 -0.773 640 0.440
isenglishspoken 2.133039 3.467 640 0.001
exchangeratevolatility 5.604298 1.576 640 0.115
culturaldistance 1.501494 4.042 640 < 0.001
As can be observed from the above table, the statistically significant variables are logofinflation (p
< 0.001), isenglishspoken (p < 0.001), and culturaldistance (p < 0.001). For each one unit increase in log of
inflation, the odds that an Indian software firm is likely to choose a developed country over an emerging
100
market, on an average, decreases by 0.06, all else being equal. The odds that an Indian software firm is
likely to choose a developed country over an emerging market increases, on an average, by 6.16 if the
country happens to be an English speaking country, all else being equal. Similarly, for each one unit
increase in cultural distance between India and the host country, the odds that an Indian software firm is
likely to choose a developed country over an emerging market, on an average, increases by 4.63, all else
being equal.
For further model building, these 3 control variables are retained in the model and other level 1
control variables were dropped. The HLM analysis was run again with these 3 variables and the results are
given in the following table.
Table 20: Selected Level 1 Control Variables
Variable Coefficient T-ratio Degrees of freedom p-value
Intercept 1.936908 6.774 31 < 0.001
logofinflation -2.830607 -5.522 646 < 0.001
isenglishspoken 1.819233 5.791 646 < 0.001
culturaldistance 1.533392 5.741 646 < 0.001
The deviance is 1789.33 (chi-square = 128. 317, d.f. = 3, p < 0.001) as seen in Model 2 of Table
22.
Summary of the model specified in equation format is as given below:
Level-1 Model
Prob(Y=1|B) = P
log[P/(1-P)] = P0 + P1*(LOGOFINF) + P2*(ISENGLIS) + P3*(CULTURAL)
Level-2 Model
P0 = B00 + R0
P1 = B10
P2 = B20
P3 = B30
101
Level-3 Model
B00 = G000 + U00
B10 = G100
B20 = G200
B30 = G300
Step 3 of the model-building:
Level 1 independent variable namely purpose is introduced to the model in this step. As can be
seen in Model 3 of Table 22, purpose is statistically significant (p = 0.01) with coefficient = 1.788301 and
odds ratio = 5.9792. Thus, the probability of an Indian software company locating in a developed country
with a market-seeking strategy is 92%, all else being equal. On the other hand, the probability of an Indian
software company locating in a developed country with a labor-seeking strategy is 65%. Hence, hypothesis
1 is supported. The deviance for the model is 1764.246 (chi-square = 25.0877, d.f. = 1, p < 0.001). As the
deviance ratio goes down, it can be concluded that the model 3 is statistically significant.
Summary of the model specified in equation format is as given below:
Level-1 Model
Prob(Y=1|B) = P
log[P/(1-P)] = P0 + P1*(PURPOSE) + P2*(LOGOFINF) + P3*(ISENGLIS) +
P4*(CULTURAL)
Level-2 Model
P0 = B00 + R0
P1 = B10
P2 = B20
P3 = B30
P4 = B40
Level-3 Model
B00 = G000 + U00
102
B10 = G100
B20 = G200
B30 = G300
B40 = G400
This completes level 1 model, leaving level 2 and level 3 models unspecified. The model is built
further by introducing level 2 control variables in the next step.
Step 4 of the model-building:
The level 2 control variables namely logofsales, ROS, cash, and time are entered in this step. The
results as given below in Table 21 show that none of these variables are statistically significant.
Table 21: Level 2 Control Variables
Variable Coefficient T-ratio Degrees of freedom p-value
logofsales -0.657369 -0.731 196 0.466
ROS -1.485633 -0.278 196 0.781
cash -0.000186 -0.711 196 0.478
Time -0.181547 -0.799 196 0.425
All level 2 control variables except time are dropped. Time dummies are retained in the model for
further analysis of the longitudinal dataset. The HLM is re-run with time as a level 2 control variable. As
can be seen in Model 4 of Table 22, time becomes statistically significant (p < 0.001) with coefficient = -
0.396807 and odds ratio 0.6724. With each passing year, the odds that an Indian software firm is likely to
choose a developed market over an emerging country decreases, on an average, by 0.67, all else being
equal. Indian software firms, as they gain international experience, tend to move more into unchartered
territories. The deviance for the model is 1730.20 (chi-square = 34.0455, d.f. = 1, p < 0.001). It can be
concluded that the model 4 is statistically significant.
Summary of the model specified in equation format is as given below:
Level-1 Model
Prob(Y=1|B) = P
103
log[P/(1-P)] = P0 + P1*(PURPOSE) + P2*(LOGOFINF) + P3*(ISENGLIS) +
P4*(CULTURAL)
Level-2 Model
P0 = B00 + B01*(TIME) + R0
P1 = B10
P2 = B20
P3 = B30
P4 = B40
Level-3 Model
B00 = G000 + U00
B01 = G010
B10 = G100
B20 = G200
B30 = G300
B40 = G400
Step 5 of the model-building:
The level 2 independent variables serving as proxies for firm resources namely CMMI, KMSW,
and wagebillperemployee are introduced in this step. None of the three variables are statistically significant
(CMMI: p = 0.067; KMSW: p = 0.121; wagebillperemployee: p = 0.083) as can be seen in Model 5 of
Table 22. The deviance drops to 1720.605 (chi-square = 9.595, d.f. = 3, p = 0.02). The main effects of
resources are not dropped in further analysis as controlling for the main effects is required before
introducing the interaction terms in the model. This concludes the level 2 model, leaving the level 3 model
unspecified. The level 3 is introduced in the next step.
Summary of the model specified in equation format is as given below:
Level-1 Model
Prob(Y=1|B) = P
log[P/(1-P)] = P0 + P1*(PURPOSE) + P2*(LOGOFINF) + P3*(ISENGLIS) +
104
P4*(CULTURAL)
Level-2 Model
P0 = B00 + B01*(CMMI_MEA) + B02*(KMSW_MEA) + B03*(WAGEBILL) +
B04*(TIME) + R0
P1 = B10
P2 = B20
P3 = B30
P4 = B40
Level-3 Model
B00 = G000 + U00
B01 = G010
B02 = G020
B03 = G030
B04 = G040
B10 = G100
B20 = G200
B30 = G300
B40 = G400
Step 6 of the model-building:
The level 3 control variable namely BusinessGroup which is the only time-invariant firm-level
variable is introduced by grand-centering in this step of the model building. The variable is not statistically
significant (p = 0.79) as can be seen in Model 6 of Table 22. The deviance drops by small value to
1720.2570 (chi-square = 0.3485, d.f. = 1, p > 0.5). The variable is, however, retained in the model as it is
the only variable at level 3 and it is required to have at least one variable at level 3 to have a three-level
model. This concludes level 3 of the model. The three interaction terms between purpose and each of the
105
three resource variables namely CMMI, KMSW, and wagebillperemployee are introduced to the model one
by one as the last step in the model building process.
Summary of the model specified in equation format is as given below:
Level-1 Model
Prob(Y=1|B) = P
log[P/(1-P)] = P0 + P1*(PURPOSE) + P2*(LOGOFINF) + P3*(ISENGLIS) +
P4*(CULTURAL)
Level-2 Model
P0 = B00 + B01*(CMMI_MEA) + B02*(KMSW_MEA) + B03*(WAGEBILL) +
B04*(TIME) + R0
P1 = B10
P2 = B20
P3 = B30
P4 = B40
Level-3 Model
B00 = G000 + G001(BUSINESS) + U00
B01 = G010
B02 = G020
B03 = G030
B04 = G040
B10 = G100
B20 = G200
B30 = G300
B40 = G400
Step 7 of the model-building:
Cross-level interactions are introduced in this step. The conceptual model in the dissertation
proposes that firm resources moderate the direct relationship between firm strategy and location choices.
106
To test the hypotheses pertaining to the moderating effect of firm resources, interaction terms between
purpose and CMMI, purpose and KMSW, and purpose and wagebillperemployee are introduced one by one
to the model. The interactions between purpose and firm resources form cross-level interactions. With
introductions of all three interaction terms, the HLM model is built completely.
Interaction between purpose and KMSW: As can be seen in Model 7 of Table 22, the interaction is
not statistically significant (p = 0.68). The deviance for the model is 1719.8187 (chi-square = 0.4383, d.f. =
1, p > 0.50). The main effect of purpose remains statistically significant (p = 0.012). Thus, hypothesis 2 is
not supported.
Interaction between purpose and wagebillperemployee: As can be seen in Model 8 of Table 22, the
interaction is not statistically significant (p = 0.768). The deviance for the model is 1719.8533 (chi-square
= 0.4037, d.f. = 1, p > 0.50). The main effect of purpose remains statistically significant (p = 0.015). Hence,
hypothesis 3 is not supported.
Interaction between purpose and CMMI: As can be seen in Model 9 of Table 22, the interaction is
statistically significant (p = 0.047). The deviance for the model is 1711.2668 (chi-square = 8.9902, d.f. = 1,
p = 0.003). The main effect of purpose remains statistically significant (p = 0.005). The probability of an
Indian software firm without CMMI level-5 appraisal locating in a developed country with a labor-seeking
strategy is only 17%, whereas the probability of an Indian software firm with CMMI level-5 appraisal
locating in a developed country with a labor-seeking strategy is 85%, all else being equal. The probability
of an Indian software firm without CMMI level-5 appraisal locating in a developed country with a market-
seeking strategy is 91%, whereas the probability of an Indian software firm with CMMI level-5 appraisal
locating in a developed country with a market-seeking strategy is 94%, all else being equal. Thus,
hypothesis 4 is supported.
The complete model in the equation format is specified below:
Level-1 Model
Prob(Y=1|B) = P
log[P/(1-P)] = P0 + P1*(PURPOSE) + P2*(LOGOFINF) + P3*(ISENGLIS) +
P4*(CULTURAL)
107
Level-2 Model
P0 = B00 + B01*(CMMI_MEA) + B02*(KMSW_MEA) + B03*(WAGEBILL) +
B04*(TIME) + R0
P1 = B10 + B11*(CMMI_MEA)
P2 = B20
P3 = B30
P4 = B40
Level-3 Model
B00 = G000 + G001(BUSINESS) + U00
B01 = G010
B02 = G020
B03 = G030
B04 = G040
B10 = G100
B11 = G110
B20 = G200
B30 = G300
B40 = G400
In the final model, only P0 and B00 were specified as random and others were specified as fixed
effects. When coefficients other than P0 and B00 were specified as random, data analysis was slow to
converge, reliability estimates were lower than 0.10, and p-value for chi-square test for homogeneity was
greater than 0.5.
The summary of the statistical results of various models is given in the table below.
Table 22: HLM Results
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
Intercept 1.53** 1.93*** 2.02*** 2.00*** 2.01*** 2.01***
108
Logofinflation -2.83*** -2.85*** -2.64*** -2.53*** -2.52***
Isenglishspoken 1.81*** 1.81*** 1.80*** 1.80*** 1.79***
Culturaldistance 1.53*** 1.56*** 1.47*** 1.50*** 1.51**
Purpose 1.78* 1.68* 1.68* 1.68*
Time -.39*** -.42*** -.42**
CMMI 1.07 1.08
KMSW -1.31 -1.27
Wagebillperemployee -11.36 -10.20
BusinessGroup -0.25
Purpose*KMSW
Purpose*wagebillperemployee
Purpose*CMMI
Deviance 1917.65 1789.33 1764.24 1730.2 1720.60 1720.25
Estimated parameters 3 6 7 8