1

LL:
pvmaﬂ

‘ Ma.
r .
.3:
I ‘11:. .. .
12%.. WWW”.
«N. .y ,

‘ 'n
.s A ‘
aﬁwsusbvmmu:
lunar In...’ 1.1:
4, Pub". ,
X... 3.!
I” ‘ I
L
amu‘

 

 

. . . .

1.9:... r $4.1“; .T it . -»
LN... 3.... 3.6.9) .4. ‘.a..rm

4:. 1 ﬂ... ﬂan: .u.‘

25 ., $7 3

51. . ‘ I: l:
.47....vzm. Jane? «2%!
.11.. .x J

 

  

I?!"
..:.....,..:..:
1:93..." at... m .
.ﬁ, ur.«..rvvhu.'..,. I.”
‘.3o 3) ‘ . II n
.. a. 811. ft

 

 

...,..r..§...v......w.

1.?

 
 

«smiu

 

Rue.

 

‘ _ . _ . 13.75..“ ﬁiirﬁamv‘r

“a... .
A ~ PTO f

 

 

 

L iCiRItRY
l‘a’ilCl‘liGan State
Uni we”

 

 

 

This is to certify that the
dissertation entitled

Retail Format Diversiﬁcation Strategies in Foreign Markets

presented by

Carol Ann Finnegan

has been accepted towards fulﬁllment
of the requirements for the

Ph.D. degree in The Department of Advertising,
Public Relations, and Retailing_

 

 

 

Date

MSU is an afﬁrmative-action. equal-opportunity employer

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/ClRC/DateDue.indd-p.1

 

RETAIL FORMAT DIVERSIFICATION
STRATEGIES IN FOREIGN MARKETS

By

Carol Ann Finnegan

A DISSERTATION

Submitted to
Michigan State University
In partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

The Department of Advertising, Public Relations, and Retailing

2007

ABSTRACT
RETAIL FORMAT DIVERSIFICATION STRATEGIES IN FOREIGN MARKETS
By
Carol Ann Finnegan

Forced to seek out sustainable growth opportunities, retailers are
expanding internationally. Extant diversiﬁcation, resource-based view (RBV), and
international entry mode literatures provide us with a theoretical framework to
examine retailer foreign market diversiﬁcation decisions. Our goal is to explain
how the inﬂuences of compositional and contextual factors impact the
diversiﬁcation strategy and performance relationship.

Taking an intra-industry approach, this dissertation ﬁlls several gaps in the
literature. First, we look at key diversification decisions (i.e., entry mode and
diversiﬁcation direction) on performance (i.e., market share) of retailer foreign
market portfolios. The research ﬁnds that greater levels of diversiﬁcation
translate into higher market shares. While both between- and within-segment
diversiﬁcation is associated with positive performance, related diversiﬁcation
generally outperforms unrelated diversiﬁcation. Contrary to the literature, we did
not ﬁnd a curvilinear relationship between retail format diversification and
performance.

Second, we explore the relative efﬁcacy of three entry mode strategies
(i.e., joint ventures, wholly-owned subsidiaries, and acquisitions) at the format
level. We hypothesize that wholly-owned subsidiaries should perform better than

the other entry modes. In our initial analysis, our ﬁndings support the literature,

while, in our second analysis, acquisitions outperform the other modes. Mean
market share performance for joint ventures falls between wholly-owned
subsidiaries and acquisitions.

Third, we look at the performance of six combined diversification
strategies. Contrary to the literature (Busija, ONeill, & Zeithaml, 1997; Rumelt,
1982), we ﬁnd that, on average, related acquisition strategies outperform all other
strategy combinations. We develop an argument to explain our results in the
context of service marketing.

This dissertation employs a hierarchical linear modeling (HLM) approach
to hypothesis testing. This approach aids us in making more nuanced inferences
about our relationships, while avoiding common analytical pitfalls, such as
aggregation bias and ignoring the influence of higher level factors on key study
effects. We show how well the degree and constancy of foreign retail market
environments and corporate membership explain variation in foreign business
unit (FBU) performance. By pooling information across F BUs and controlling for
higher level covariates, we also develop more reliable parameter estimates.

Finally, we discuss the limitations, future research and managerial
implications. Our results suggest that some aspects of marketing and economic
theories may have limitations in a service marketing context. We also try to

provide some key insights from our research for industry managers.

Copyright by
CAROL ANN FINNEGAN
2007

I would like to dedicate this to my grandmother,
Rita — my friend and travel companion.

ACKNOWLEDGEMENTS

The doctoral program and this dissertation process has been a long and
winding road detouring into a few cul-de-sacs along the way. Honestly, I doubt
that I could have made it through this program without the help and guidance of
Dr. Brenda Sternquist. She demonstrated how to write an academic manuscript,
encouraged me to submit papers to conferences and journals, and gave me
advice on how to become a better teacher. My dissertation committee members
have also shown a great deal cf patience, humor, and support through this
process. lwould like to thank them — Dr. Linda Good, Dr. Kyoung-Nan Kwon, Dr.
S. Tamer Cavusgil, Dr. G. Tomas Hult, and Dr. Yeow Meng Thum.

Encouragement also came in droves from my friends and family. Along
the way, I have sent papers to them for comments, ate their food and slept on
their couches, and leaned on them for emotional support. They helped me
survive the transition from private sector to academia, occasionally reminding me
of why I embarked on this journey when it all seemed impossible. Though their
lives and priorities keep changing, their support has been unwavering. For that

(and many more reasons), I love them.

vi

TABLE OF CONTENTS

LIST OF TABLES ................................................................................................. ix
LIST OF FIGURES ............................................................................................... x
CHAPTER I .......................................................................................................... 1
INTRODUCTION .................................................................................................. 1
PROBLEM STATEMENT AND SIGNIFICANCE OF THE STUDY ....................... 9
CHAPTER II ....................................................................................................... 12
LITERATURE REVIEW ...................................................................................... 12
DIVERSIFICATION STRATEGY ........................................................................ 13
Direction of diversiﬁcation ........................................................................ 15
Governance ............................................................................................. 27
CHAPTER III ...................................................................................................... 39
METHODOLOGY ............................................................................................... 39
KEY VARIABLES ............................................................................................... 39
Endogenous Variable .............................................................................. 39
F BU Level Exogenous Variables ............................................................. 42
Corporate Level Characteristics ............................................................... 43
Country Level Characteristics .................................................................. 46
SAMPLE ............................................................................................................. 49
METHODS ......................................................................................................... 54
Conceptual Advantages ........................................................................... 54
Analytical Reﬁnements ............................................................................ 54
MODELS ............................................................................................................ 57
HYPOTHESIS TESTING .................................................................................... 60
CHAPTER IV ...................................................................................................... 63
RESULTS AND DISCUSSION ........................................................................... 63
Hypothesis 1 ............................................................................................ 66
Discussion of Hypothesis 1 Findings ....................................................... 89
Hypothesis 2 ............................................................................................ 94
Discussion of Hypothesis 2 Findings ..................................................... 111
Hypothesis 3 .......................................................................................... 114
Discussion of Hypothesis 3 Findings ..................................................... 126
CHAPTER V ..................................................................................................... 134
CONCLUSIONS ............................................................................................... 134
LIMITATIONS AND FUTURE RESEARCH ...................................................... 139
REFERENCES ................................................................................................. 143

vii

APPENDIX A: HISTOGRAMS, DESCRIPTIVE STATISTICS, ........................ 161

AND BIVARIATE CORRELATIONS ................................................................. 161
APPENDIX B: DECOMPOSING DIVERSIFICATION RESULTS ..................... 180
APPENDIX C: ADDITIONAL ANALYSES ........................................................ 183

viii

LIST OF TABLES

Table 1: Data Coding
Table 2: Corporationsinthe Sample.............................
Table 3: Countries in the Sample
Table 4: Diversification-Performance Relationship with Corporate Effects... ..
Table 5: Diversification-Performance Relationship with Country Effects... ..
Table 6: Format Governance Performance with Corporate Level Effects... ..
Table 7: Format Governance Performance with Country Level Effects... ..
Table 8: Strategy Frequency
Table 9: Format-Level Diversiﬁcation Strategy with Corporate Effects...
Table 10: Format-Level Diversification Strategy with Country Effects...
Table11: Summary ofResuIts
Table 12: Corporate Level Effects

Table 13: Country Level Effects.............. ..........

.47

64

.65

69

81

100

109

115

119

.124

135

180

182

LIST OF FIGURES

Figure 1: Global Powers of Retailing 2005 ......................................................... 52
Figure 2: International Diversification on Mean Corporate Market Share ........... ‘72
Figure 3: Corporate Size on Mean Corporate Market Share .............................. 74

Figure 4: Corporate Experience on Diversiﬁcation-Market Share Relationship .. 76

Figure 5: Corporate Size on F BU Size-Market Share Relationship .................... 78
Figure 6: Retail Labor Market Influence on Mean Market Share ........................ 83
Figure 7: Level of Development Impact on Mean Market Share ......................... 85

Figure 8: Level of Development on FBU Outlet — Market Share Relationship 87
Figure 9: International Diversiﬁcation Impact on JV-Market Share

Relationship ........................................................................................................ 97
Figure 10: Labor Market Size on Mean Market Share for Non-Majority
Acquisitions, JVs and WOS .............................................................................. 104
Figure 11: Industry concentration impact on Size-Market Share Relationship .105
Figure 12: Centralization Impact on Mean UACQ Market Share ...................... 117
Figure 13: Retail Market Size Impact on Format Size-Market Share

Relationship ...................................................................................................... 123

Images in this dissertation are presented in color.

CHAPTER I

INTRODUCTION

Multinational corporations (MNCs) are becoming increasingly important
players in global markets. The world’s 500 largest MNCs drive the globalization
process and represent 90% of foreign direct investment (FDI) and control 50% of
global trade (Rugman, 2000; 2004). In terms of global retailing, the top 250
companies control almost one-third of the $9 trillion dollar global retail market
(Deloitte Touche Tohmatsu, 2006). The complexities in international business
operations poses unique challenges and opportunities for MNCs in exploiting
existing strengths, while also exploring foreign environments for sources of new
capabilities (Ghoshal, 1987; Kogut, 1983; March, 1991 ). MNCs are particularly
motivated to enter foreign markets where they possess speciﬁc advantages
relative to host country competitors (Dunning, 1980; Hennart, 1982; Teece,
1983). Being exposed to different consumer markets and competitive challenges,
MNCs often learn new ways of achieving business objectives and innovating
product and service offerings. Achieving a balance between exploitation and
exploration of ﬁrm capabilities has been linked with success (Penrose, 1959).

The MNC has been conceived as an outcome of internalization processes
(Buckley & Casson, 1976; Dunning, 1980; Rugman, 1981) and strategic
motivations (Hamel & Prahalad, 1985; Kim & Hwang, 1992; Kogut, 1985) as well

as a dynamic organizational knowledge economy (Madhok & Liu, 2006). Recent

extensions to the literature make great strides in identifying unique aspects of the
MNC, a special case of the ﬁrm, and developing theory around the MNCs’
remarkable position of acquiring and utilizing knowledge from diverse
international sources (Dunning, 1988). For example, Madhok and Liu (2006)
theorize about the roles that subsidiaries, interacting with both the local
environment and amongst vertical and other MNC business units, are influenced
by and concerned with fundamentally different issues than their parent
companies. By drilling down into the various levels and inﬂuences of the MNC,
scholars have begun to identify structures, strategies, capabilities, and processes
at each level that contribute to the unique aspects of operating internationally and
how these advantages may lead to or hinder performance (e.g., Bartlett &
Ghoshal, 1989; Li, 1995; Ricart, Enright, Ghemawat, Hart, & Khanna, 2004).

We adapt a definition of the MNC from Madhok and Liu (2006): “the MNC
is both a single organization and a population of organizations, connected
somehow to a common legal entity, operating at different levels, but bound by a
common managerial vision” (p. 3). There are several important points in this
deﬁnition. First, organizations exist at multiple levels, each level with distinctive
strategies, objectives and performance implications that do not always have
exact parallels on other levels (Ginsberg & Venkatraman, 1985). At the highest
level, an MNC is considered a knowledge economy in which strategy is
formulated to exploit existing knowledge, explore its various operating
environments and engage with partners and customers to find new sources of

information and knowledge (Amit & Schoemaker, 1993). Because knowledge is

considered the most important intangible asset of a ﬁrm (Day, 1994), we are
interested in understanding the conditions that would support or limit the
development and transfer of knowledge between various parts of the MNC.
Second, at the business level, business units struggle with the question of
how to compete in a specific context given existing competitive and industrial
conditions (Bourgeois, 1980; Porter, 1987). Certain industries may have specific
strategic constmcts, sources of synergy, and measures that are not necessarily
comparable to other industries (Snow 8 Hambrick, 1980). Some FBUs can even
become important sources of knowledge and other competencies from which
other parts of the MNC can learn (Holm & Sharma, 2006). As the scope of
corporate activities widens, MNC business units operate in a greater number of
heterogeneous competitive environments. In other words, the corporate
managers increasingly strategize about operations across diverse environments,
whereas business managers are more concerned with local competitive
pressures. Consequently, it becomes important to be clear about the distinctions
between corporate and business level strategies (Ginsberg & Venkatraman,
1985). For example, a business unit is making strategic decisions to survive and
compete within the given industry, whereas the corporation is developing
strategies about the overall direction of all the businesses within its portfolio of
businesses. Examples of MNC forces could include home ofﬁce attributes, such
as, structure, age, and industry experience as well as internationalization
strategies. These factors would affect the amount and ﬂow of knowledge within

the MNC economy.

Third, because knowledge and other organizational capabilities can also
be developed with an industry or country, the context matters tremendously
(Misangyi, Elms, Greckhamer, & Lepine, 2006; Porter, 1990; Rumelt, 1991).
Within the international business literature, Toyne & Nigh (1997) note the
importance of identifying the appropriate level of analysis and theorize about the
boundary conditions they impose on lower hierarchical levels. Yammarino &
Dausereau (2004) also urge researchers to consider the most appropriate units
and levels of analysis critical to the study of a phenomenon. In our study, we
argue that a ﬁrm’s foreign business units (F BUs) are nested simultaneously
within their parent corporation as well as within the foreign retail industries in
which they operate. Therefore, industry and corporate factors necessarily have
an impact on F BU decision making and performance.

We focus on the strategic resource choices FBU managers make in
expanding within a foreign market. This represents a divergence from much of
the literature in how we examine diversiﬁcation. Instead of studying products or
foreign market choices, we report on format portfolio development decisions.
Because the choice of format is a critical element of retail expansion strategy, we
delve into the performance implications of foreign diversiﬁcation alternatives.
Our assumption is that F BU managers make decisions about which resources
and activities to pursue to foster development and reinvention of strategic
competitive advantages within the local market. In a RBV framework, resources
can be conceptualized in terms of services that are either directly or indirectly

(i.e., goods) rendered for consumers (Vargo & Lusch, 2004a, 2004b).

With respect to knowledge usage and development, ﬁrms have a choice
of exploiting existing knowledge or leveraging knowledge of partners or allies.
The fundamental question is how to use existing resources and when to
purchase others to gain economic rents and avoid losses. Effective and efficient
utilization of resources is concerned with the scope, relatedness, and
management of ongoing and new activities. Because diversification strategy is
critical to issues of resource deployment and management, it represents an
important area of study in explaining the differences in MNC business unit
performance.

In addition to diversiﬁcation choices, a key element of international
strategy is governance mode (e.g., Buckley & Casson, 1976; Cavusgil, 1980;
Dunning, 1980; Johansson & Vahlne, 1977). In the internationalization literature,
foreign entry modes can be considered as existing on a continuum from low
control to high control (Doherty, 1999). Representative low control modes for
service ﬁrms include franchising and in-store concessions, while a high control
mode is foreign direct investment in the host country (Dunning, 1988). Hybrid
modes are types of strategic alliances and joint ventures. The appropriate choice
of entry mode is a critical ingredient in foreign market entry strategy‘ (Root,
1994). The consequences of a wrong or sub-optimal choice range from
increased transaction costs to business failure.

Mode of diversiﬁcation refers to the governance mechanism chosen to

enter into the new line of activity. Traditional Transaction Cost Analysis (TCA)

 

‘ Interestingly, Calantone and Zhao (2000) ﬁnd that while Korean and American managers
positively link their ﬁrms” performance to issues of control, Japanese managers do not. This
opens up the possibility of cross-cultural differences in usage of formal control mechanisms.

5

predicts that the governance mechanism chosen to enter into a foreign market is
determined by the level of asset specificity, environmental uncertainty (i.e., cost
of transacting international business) and behavioral uncertainty (i.e., free-riding
potential and/or performance ambiguity). In general, when ﬁrms have large
investments in proprietary assets and uncertainties are high, firms would want to
retain greater control over their foreign operations (e.g., Anderson & Gatignon,
1986). As the investments in non-redeployable assets and uncertainties go
down, markets become a more efﬁcient means of exchange than ﬁrms.

A speciﬁc limitation of the diversification literature is the assumption that
the ﬁrm has to own knowledge to beneﬁt from it. Unlike other areas of
management and marketing which explore the value of a partner’s knowledge
(e.g., Bartlett & Ghoshal, 1989; Granovetter, 1985), the diversification literature
assumes that synergies are either created and utilized within the corporation or
acquired and brought into the organization. By only considering one end of a
governance continuum, extant diversification literature has only uncovered a
fraction of the potential strategic combinations. Also, while corporate governance
and product diversiﬁcation choices do not appear to be directly related (Busija et
al., 1997), it is unclear whether business unit diversiﬁcation and governance
choices are directly or indirectly related.

External forces influencing firm operations in foreign markets include
society, industry and competitors (Cavusgil, Ghauri, & Agarvval, 2002; Lenway &
Murtha, 1994; Porter, 1990). Instead of studying external factors that may

inﬂuence market selection, this study introduces a theoretical contribution by

conceptualizing the multiple moderating roles of institutional forces that affect
implementation of business level strategies. We develop an argument that
corporate managers must force business units to become more similar
(interchangeable) for ease of cross-market coordination and control. In
accordance with previous research, when the firm pursues a related
diversification strategy and organic growth, the ease in knowledge transfer
between business units should enable the realization of actual synergies required
for superior performance (Rumelt, 1974). This implies that firms strategize and
organize their business units to exploit firm advantages. Concurrently, the host
country legal-regulatory environment, industry associations, stock markets, and
others also reward ﬁrms for being superficially more similar to their competitors
(i.e., adhering to industry norms). For example, one explanation for Home
Depot’s failure in the Chilean market was that the company and its managers did
not operate in accordance with Chilean business or social expectations (Bianchi
& Arnold, 2004). Thus, corporate and environmental forces have the potential to
moderate the relationship between business unit strategy and performance. In
summary, business scholars are urged to use a broader lens in identifying the
salient factors impacting the relationship of interest (Toyne & Nigh, 1997).

This also suggests that straightforward explanations of diversiﬁcation and
governance strategies may need to be reconsidered. Because competition is a
local phenomenon, a ﬁrm’s ability to adapt to and thrive in local environments
may depend upon different strategy combinations. The moderating forces from

both the MNC and the local environment are anticipated to affect an FBU’s ability

to implement its growth strategy. When implementation is more difﬁcult and/or
costly than anticipated, the FBU may be unable to realize actual synergies from
resource sharing. The reverse is also possible. Certain strategy-performance
relationships can be enhanced by membership within a corporation and/or retail
industry. Thus, institutional forces may either suppress or magnify the potential
beneﬁts arising from synergies and affect performance.

Our conceptual framework is tested on a sample of the world’s largest
international retailers. Spurred on by global opportunities, reductions in
international information and communication costs, saturation in home markets,
and shareholder pressures for growth, many of the world’s largest retailers are
proactively seeking growth opportunities in foreign markets. For instance,
excluding Sam’s Clubs, in comparing 2005 to 2006, Wal-Mart’s international
sales grew by 11.4% and its store network grew by 43%, whereas domestic Wal-
Mart sales grew by 9.4% and its store network grew by less than 5% (Wal-Mart,
2006). In addition, the scale of international investments is unparalleled in the
history of retailing (Coe, 2004). Despite the enthusiasm for international growth,
anecdotal evidence also suggests that retailer performance in international
markets is, at best, uneven (Alexander & Quinn, 2002; Burt, Dawson, & Sparks,
2003). Discouraging evidence ranges from Ahold’s exit from Latin America
(Wrigley & Currah, 2003) and Wal-Mart’s recent announcements of its exits from
both South Korea and Germany to Boots’ stumbles in Japan. There have also

been successes, such as Carrefour’s storming of the Chinese market and Aldi’s

expansion in Europe. We are interested in discovering to what extent retailer
diversiﬁcation strategies can explain performance.

Unfortunately, scant research exists that examines the forces that
inﬂuence the relationship between diversification strategies and performance of
service ﬁrms (Nayyar, 1992, 1993b). Attempts to retroﬁt explanations from the
manufacturing sector into the service sector have proven inadequate, given
important strategic and structural differences between the sectors (Cortsjens &
Cortsjens, 1995; Zeithaml, Parasuraman, & Beny, 1985).

Marketing fast moving consumer goods as opposed to pure services,
retailer concerns and opportunities can also been distinguished from other areas
of service marketing (Ekeledo & Sivakumar, 1998). Our objective is to gain some
insights into how retailer diversiﬁcation strategies link to their performance in
foreign markets; marketing scholars from all disciplines agree about the

importance of exploring these relationships (Dunning, 1995).

PROBLEM STATEMENT AND SIGNIFICANCE OF THE STUDY

Due to mixed results in the research and business worlds, retailer
internationalization strategy has been identified as an area requiring further study
(e.g., Doherty, 1999; Pelligrini, 1994; Stemquist, 1997a). This study investigates
a critical sub-component of retailer business unit strategy - diversification choice.
The objectives of this study are to propose a conceptual framework to explain the
relationship between ex post business level diversification and governance

decisions and performance. Using a multi-level model, institutional forces (i.e.,

9

corporate and country) will be proposed and tested for relative strength and
importance on FBU strategy-perfonnance outcomes. Diversiﬁcation theory (i.e.,
resource based view) and Transaction Cost Analysis provide the theoretical
foundations for our examination of this topic.

We theorize about the way in which FBUs expand within host country
markets. This perspective differs from traditional international diversiﬁcation
strategy, which refers to corporate expansion across national borders into
different consumer markets (Hitt, Hoskisson, & Kim, 1997). While international
diversification literature is an important part of the strategic management
literature, by focusing on the firm’s scope of diversification, it generally treats all
businesses alike and ignores distinctive country strategies and performance. As
a consequence, the international diversiﬁcation literature can erroneously
attribute corporate success to aggregated FBU performance when the results
may be due to a handful of extremely successful FBUs.

The research questions pertain to the following areas: 3) understanding
the direct linkages between ex post diversification and governance choices and
foreign market performance; b) considering how combinations of diversiﬁcation-
govemance choices may inﬂuence foreign market performance; and, c)
identifying contextual inﬂuences that affect the relationship between strategy
decisions and foreign market performance. By exploring the strategy-
perforrnance link within foreign market operations, we hope to uncover
successful strategies across nations and corporations. Moreover, we separate

levels of analysis (i.e., FBU vs. corporation/country) for the purposes of

10

advancing retailing theory and accounting for and attributing sources of variance
between levels. Our purpose is to advance business unit strategy in order to
provide retailers with useful information to consider in making internationalization
decisions.

The rest of the dissertation is organized as follows. In Chapter II, we offer
a comprehensive literature review of the extant literature, synthesize our ﬁndings,
and develop our hypotheses. In Chapter III, we discuss our methodology,
providing a summary of our data, analytical techniques and models. Chapter IV
contains our results with a discussion of our ﬁndings. Chapter V concludes with
an interpretation and implications of our ﬁndings and a discussion of the study’s

limitations, future research and conclusions.

11

CHAPTER II

LITERATURE REVIEW

In this chapter, we provide a synthesis of the literature, develop our
hypotheses and present the conceptual framework. Our objective is to explore
performance consequences of international retail expansion strategies.

Expansion is a primary business objective for retailers. Firms pursue new
market opportunities wherever available at home and/or abroad. In pursuing
market opportunities, corporations seek to leverage proprietary and other
resources and capabilities in order to create synergies between existing and new
resources (Dunning, 1980; Rumelt, 1974). These synergies are often expressed
in the form of economies of scope, scale and learning (Kogut, 1985). By sharing
core competencies across business units, firms can outperform their competitors
and identify new business opportunities (Burgelman & D02, 2001; Hamel &
Prahalad, 1990).

lntemational expansion can be costly. Business risks associated with
operating in a foreign environment (e.g., country risk) tend to be higher than
domestic expansion. Currency exchange rate ﬂuctuations or risks of
appropriation of property represent some of the risks. Environmental
uncertainties due to lack of business experience or cultural differences can
increase operational costs (Johanson & Vahlne, 1977; Kogut & Singh, 1988). In
contrast, by spreading the activities across different geographic regions, firms

can reduce their business risk (Kim, Hwang, & Burgers, 1993).

12

In the proceeding section, we discuss salient elements of diversiﬁcation
strategy. Synthesizing and building upon extant literature, we develop our
hypotheses. Our objective is to examine ex post foreign market diversiﬁcation
decisions, while accounting for the differences between corporations and host
country environments that may affect strategic implementation.
DIVERSIFICATION STRATEGY

Farjoun (2002) refers to strategy as a posture, noting two main types of
postures - “position (e.g., differentiation strategy) and scope (e.g., vertical
integration)” (p.563). We follow Ramanujam and Varadarajan’s (1989) argument
that these two types of postures are actually two key elements of corporate
diversiﬁcation strategy. After the expansion decision has been made, position
and scope jointly determine how the strategy will be implemented in the host
country.

Diversiﬁcation refers to “the entry of a ﬁrm or business unit into new lines
of activity entailing changes in its administrative structure, system, and other
management processes” (Ramanujam & Varadarajan, 1989, p. 525). Within the
diversiﬁcation literature, several explanations have been put forth that examine
how, why, and where firms diversify (See Datta, Rajagopalan, & Rasheed, 1991;
Hoskisson & Hitt, 1990; Ramanujam & Varadarajan, 1989 for excellent reviews).
Theoretical foundations applied to these questions include industrial organization
paradigms (e.g., Schmalensee, 1985), agency theory (e.g., Jensen & Meckling,
1976; Trautwein, 1990), resource-based view (e.g., Rumelt, 1974), resource

dependence theory (e.g., Pfeffer & Salancik, 1978), dominant logic (e.g., Bettis 8

13

Prahalad, 1995) and finance theory (e.g., Lubatkin & O'Neill, 1987). Common
propositions of these studies are that when a ﬁrm pursues a related
diversiﬁcation strategy, superior performance results from internal ownership.
Conversely, when a ﬁrm pursues an unrelated diversification strategy, superior
performance results from acquisitions. The underlying premise for these
propositions is that if causal ambiguity is low enough and absorptive capacity is
high enough, the costs of transferring information within the firm will be lower
than potential cost advantages of buying the knowledge (Madhok & Liu, 2006;
Shanna,1998)

For long-term survival and performance, corporations need to create a
balance between exploration and exploitation activities to continuously bolster
and renew their capabilities (Ansoff, 1969; Madhok, 1997). An implicit
assumption in this literature is that continuous learning enables ﬁrms to develop
knowledge capabilities required to provide products and services desired by
customers. Strategy literature centers on exploitation of existing resources as
the primary reason to diversify (i.e., related diversiﬁcation), while finance
literature focuses on whether building diverse corporate portfolios can reduce
overall business risk (i.e., unrelated diversiﬁcation). However, one could also
argue that unrelated diversification can be treated as an exploration activity since
the firm is experimenting in a new line of business or customer group. For
example, by virtue of crossing national borders, firms expose themselves to new
sources of information and customer groups. Furthermore, within the same firm,

a combination of different diversiﬁcation strategies may achieve a balancing

14

objective in order to maintain and renew sources of competitive advantage over
time. Our purpose is to explicate the manner in which the dual exploration-
exploitation goals are achieved by allowing for different combinations of

strategies.

Direction of diversiﬁcation

The direction in diversiﬁcation refers to a modiﬁcation of the business to
meet the needs of targeted customer group(s) (Abell, 1980). Much of the
diversiﬁcation literature measures diversiﬁcation by the product-market
proliferation choices. But firms often have difﬁculty satisfying customer groups
with divergent product needs. For example, Carroll and Swaminathan (1992)
observe the challenges of beer producers trying to switch from mass production
and distribution of light beer to volume premium beer. Because brewers had
previously invested their resources in a significantly different market (e.g., light
beer), they did not have the capabilities necessary to win in the premium beer '
business. Aside from manufacturing issues, marketing issues varied signiﬁcantly
different between the two businesses. This result is similar to Dupont’s situation
in the early 20th century where differences between two markets reduced some
of the manufacturing economy of scale benefits (Chandler, 1990).

Some retailers have shown a tendency to stick with what they know.
Wrigley (2000) describes Sainsbury’s US diversification strategy that tended to
‘Irely on exporting core capabilities, such as merchandising, logistics, systems
and store development competencies, of the UK home ofﬁce to its US

businesses” (p. 904). Although businesses appear to be fairly similar, strategic

15

capabilities are only transferable to other businesses that have the similar core
operating principles and customer group characteristics.

Typically dealing with thousands of SKUs, retailer diversiﬁcation strategy
is not necessarily comprehensible at the product level2. But, at the format level
where customer service level expectations are widely deﬁned by format type,
ﬁrms tend to adhere to these expectations. Subsequently, each format offers a
distinctive retail mix (Colla, 2004). Within mass market retailing, ﬁrms that are
geared toward one end of a value/price continuum, which emphasizes high
volume, limited assortment and low price, should have substantial difﬁculties
shifting their business operations toward the companies on the mid to high end of
the value/price continuum, which emphasizes lower volumes, wider assortment
and variety and higher prices (Levy & Weitz, 2007). Each format offers
distinctive functionality for differentiated customer groups and requires
specialized resources in order to compete effectively with other formats (i.e.,
inter-type competition). This makes the choice of how to diversify particularly
important for retailers.

Related diversiﬁcation is a means by which ﬁrms modify and expand their
business to offer new functionality to existing customers or to capture new
customers. Firms tend to move into related areas of a business segment with the
intention of leveraging their unique resources in new markets by developing

economies of scale, scope and knowledge (Chandler, 1990; Teece, 1980). While

 

2 This conceptualization contrasts with Ansoffs (1988) depiction of ﬁrm strategy in which he urges
business managers not to over-emphasize customer needs relative to existing product needs. He
illustrates the dangers of too much customer focus by speculating that individual customer needs
could probably not be met for food and TVs in the same retail store because these two products
have such different customer markets (p.77).

16

resources are defined as all assets, specific processes and knowledge are
required by ﬁrm managers to design and implement efﬁcient and effective
strategies (Daft, 1982)

Cross-exploitation of excess resources supplies one rationale for
diversiﬁcation direction decisions. Because ﬁrms are concerned with the most
efﬁcient use of excess capacity (Chandler, 1990; Montgomery & Wemerfelt,
1988), diversiﬁcation strategies create opportunities to fully utilize slack
resources in existing and new markets. For example, an excess capacity of
human expertise provides an important motive fOr diversiﬁcation (Farjoun, 1994).
Benefits accrue when resources are shared across business units. Ansoff (1969)
argues that synergy can be manifested through joint usage of common skills and
facilities as well as through new applications/activities created with
complementary skills. By strategizing about connections between related
resources, other applications or synergies3 may be uncovered and the costs of
utilizing these joint resources may be reduced (Li & Greenwood, 2004). At the
same time, firms not following a diversiﬁcation strategy should not be able to
access the resource at a similar or lower cost than the diversified company
(Markides & Williamson, 1996; Porter, 1987). This efﬁciency argument is geared
towards achieving the maximum productivity from the corporation’s assets
relative to the market.

Drawing from resource-based view, the synergy creation hypothesis

complements the cross-exploitation explanation. Synergies are created when

 

3 Critics argue that by treating potential synergies as actual synergies, causal attributions become
difﬁcult to support in part due to a tautology created by examining constructs that are deﬁned in
terms of related outcomes (Nayyar1992; Tanriverdi and Venkatraman 2005).

17

resources are combined to produce an effect greater than either resource alone
could create. Sometimes referred to as core competencies (Hamel & Prahalad,
1990) or dynamic capabilities (e.g., Eisenhardt & Martin, 2000), these higher-
order resources produce synergies by inducing interdependencies between
critical resources, particularly knowledge. Knowledge can be embedded in
products/process technologies, individuals, or organizations (Robins &
Wiersema, 1995).

We focus on organizational knowledge. Through usage in problem-
solving and socialization processes, individuals utilize their skills to create an
organized body of knowledge that is unique to the firm (Nelson & Winter, 1982).
For example, ﬁrms are able to codify their operational know-how and host
country market information into routines, decision support systems, and other
processes (Morgan, Zou, Vorhies, & Katsikeas, 2003). However, the more tacit
and complex the knowledge, the more difficult it is to codify into organizational
routines and transfer within the corporation (Szulanski, 1996). Organizational
routines are integrated within the ﬁrm to create unique firm capabilities which
when utilized lead to valuable outcomes (Makadok, 2001 ). In part, stores of
organizational knowledge are utilized in the formulation of organizational strategy
(Crossan & Berdrow, 2003; Mintzberg, Ahlstrand, & Lampel, 1998). Therefore,
the unique capabilities being utilized in diversiﬁed companies should create both
revenue-generating and cost-reducing opportunities, translating into higher

profits.

18

Economies of scope and potential synergies are the primary benefits
arising from diversiﬁcation choices (Penrose, 1959; Teece, 1982). In retailing,
economies of scale come through handling large volumes of products at lower
unit costs, while economies of scope arise by distributing large numbers of
similar national brand and/or store brand product lines through common
warehousing and/or distribution facilities (Chandler, 1990). Economies of scope
are also accumulated to the extent that ﬁrm’s can coordinate the access and ﬂow
of knowledge across and within business units. While economies of scope are
sometimes observable, the resources that create these economies are not
necessarily tangible. For instance, more visible aspects of retail management
expertise can be transferred abroad in a wide range of manifestations, including
store formats, in-store design, supply chain management systems and customer
service innovations. But Coe (2004) argues that more opaque knowledge
capabilities of retailers tend to be embedded in back ofﬁce operations. Because
knowledge is an unobservable asset, competitors have difﬁculty imitating and/or
duplicating assets (Barney, 1991; Wemerfelt, 1995). Causal ambiguity also
restricts competitors’ ability to understand a sequence of events or actions
required to yield similar results (Dierickx & Cool, 1989; Grant, 1996). Since
strategic resources are rare, inimitable, valuable, and non-substitutable, and can
be used to create competitive business strategies (Peteraf, 1993; Wemerfelt,
1984), knowledge represents a long-term potential advantage and source of

abnormal returns and proﬁt (Barney, 1991).

19

As compared to related divisions in different countries, related businesses
within the same country can more easily combine marketing activities (i.e.,
logistics, advertising, distribution, etc.), allowing units within the same country to
develop economies of scale and scope (Chandler, 1990; Davis, Robinson,
Pearce ll, & Park, 1992). For example, building up overlapping networks of retail
outlets creates the opportunities to reduce in-bound and out-bound (i.e., reverse
logistics) transportation costs and extract media discounts. Doherty (1999)
argues that retailers possess important intangible capabilities in the form of
formats and managerial technology. In addition, synergies can also be
transported across national borders. For instance, expatriate staff can import
their expertise to the host country operations (Farjoun, 1994). F irrns primarily
pursuing related diversiﬁcation are able to realize horizontal synergies across
their businesses that lead to reduced costs, increased margins, and higher
proﬁtability (Porter, 1985).

Primarily based on the manufacturing sector, empirical support for related
diversiﬁcation-performance relationship has been mixed (Hoskisson and Hitt
1990; Ramnujam and Varadarajan 1989), theory suggests that there should be a
positive relationship. Firms pursuing related diversiﬁcation strategies should
outperform firms with unrelated diversification strategies (Bettis & Hall, 1982;
Markides & Williamson, 1994; Miller, 2004; Montgomery & Wemerfelt, 1988;
Palepu, 1985; Rumelt, 1974). Within the corporation, firms can leverage resource

complementarities to improve performance (Milgrom & Roberts, 1995). Higher

20

firm growth rates are also associated with related diversification strategies
(Jacquemin & Berry, 1979).

Some researchers find a non-significant relationship between
diversiﬁcation and performance (Bettis & Hall, 1982; Christensen & Montgomery,
1981 ). For example, in dividing their sample between related and unrelated
diversifiers, Am it & Livnat (1988) do not ﬁnd signiﬁcant differences in
performance between companies employing different strategies. Bergh (1995)
demonstrates that some longitudinal research was conducted without regard to
violations of statistical assumptions or modeling, which have resulted in
potentially erroneous ﬁndings. For example, he ﬁnds no relationship between
diversiﬁcation and performance after correcting for statistical violations, but also
ﬁnds some evidence of changes in the relationship over time from negative to
positive. Finally, other researchers observe a negative relationship between
diversification and performance. .In a simulation study of diversiﬁcation moves,
Gary (2005) shows that a higher degree of asset relatedness may negatively
impact performance when ﬁrms make inadequate investments in shared
resources and overstretched resources.

But it is unclear whether these results hold in a foreign market context.

F BUs may not be able to capture the same synergies as domestic businesses.
For example, when examining performance differences in MNCs, Palich,
Cardinal, & Miller (2000) do not find a difference between product-related and
unrelated diversifiers. However, Geringer, Beamish and daCosta (1989) ﬁnd that

MNEs that have a higher degree of internationalization experience and leverage

21

related products outperform MNEs operating in fewer markets with unrelated
products. In addition, Nayyar (1992) finds that focus on customer segments
leads to greater performance in service ﬁrms, whereas focus on internal
capabilities or geographic regions does not improve performance. But Nayyar
tests his hypotheses on a range of service industries and services‘. Industry
effects are known to confound corporate effects (Rumelt, 1991). In summary,
theory needs to be tested at the FBU level.

Unrelated diversiﬁcation refers to entry into completely different
businesses than the ﬁrm’s core business. This can be achieved through entering
into completely unrelated business in the same industry, sector, or through
vertical integration. Unrelated diversiﬁcation offers another opportunity for
retailers to expand into new lines of business. By exploring new lines of
business, ﬁrms are exposed to new sources of information. This creates learning
opportunities, which help develop a greater store of knowledge and can lead to
the development of new capabilities (March, 1991). Moreover, different sources
and types of information may also spark organizational innovations (Madhok,
1997; Nelson & Winter, 1982). These innovations may provide the basis for new
or renewed sources of sustainable competitive advantage (Barney, 1991).

The costs of unrelated diversiﬁcation have also been discussed in the

literature. When a ﬁrm has no previous experience in a line of business, upfront

 

4 ln Nayyar (1993), information asymmetries are found to have a differential impact for experience
and search products. Economies of scope are found to be positively associated with search
services (eg. retailing) as opposed to experience services. In addition, the CEOs are the
informants for the survey. It is questionable whether they possess the depth of information and
knowledge to comment on fairly detailed and speciﬁc operational matters. Finally, international
diversiﬁcation was tested at the regional, not country level, raising questions about the
appropriate level of aggregation.

22

investments in learning about market operations can be costly. Firms may also
need to transfer or acquire capable personnel to guide market entry, representing
opportunity costs of managerial time and effort. Finally, performance outcomes
are less certain.

In the management literature, extensive diversity is considered an inferior
strategy to a more focused strategy (e.g., Rumelt, 1982). Some studies
contradict these findings by providing evidence of domestic companies that
manage to perform adequately and survive, despite the jump into unrelated
products (Leontiades, 1986; Ramanujam & Varadarajan, 1989). Several
explanations have been put forth in the literature. First, as R&D investment and
technology diversity increase, diversiﬁed ﬁrms outperform undiversiﬁed ﬁrms
(Miller, 2006). Second, failures in external markets may also create opportunities
for economies of intemal capital markets in some environments (Williamson,
1985). This assumes that firms’ internal resources are more efﬁcient than the
market under certain conditions. Third, managers who are exposed to different
kinds of challenges may be more adaptable than managers who never face any
challenges. For instance, management teams that are equipped to handle high
administrative complexity and closely attend to capital allocation and use
decisions have been known to outperform competitors (Dundas & Richardson,
1982). Moreover, bureaucratic costs are also posited to be the lowest for
unrelated diversifiers, who do not need to engage in the same level of intra-firm
coordination and cooperation as related diversiﬁers (Jones & Hill, 1988). Thus,

high diversity may not necessarily be a sub-optimal strategy in all contexts.

23

But the lntemational context certainly complicates the management task.
Subsidiary failure rates tend to be higher when pursuing diversiﬁed products in
foreign markets (Bane & Neubauer, 1981). Compounding uncertainties
associated with foreign market entries, firms that pursue diverse products
decrease their potential for survival (Li, 1995). Consequently, unrelated
businesses tend to be ﬁrst on the chopping block when the firm is under severe
pressure from foreign competitors in order to best utilize ﬁnite managerial
capacity (Bowen & Wiersema, 2005).

Vertical integration is another form of unrelated diversiﬁcation. For
retailers, backward integration into distribution services and/or manufacturing
represent the primary opportunities for unrelated diversiﬁcation. Diversification
strategies are typically undertaken because managers perceive performance
beneﬁts (i.e., cost minimization and/or value creation) in extending their
investments in their supply chains. Backward integration of distribution services
and/or manufacturing is a way to diversify away systematic business risk in
uncertain environments as well as to provide a platform for ensuring stable
supplies for retail operations. For instance, in developing economies, Aldi tends
to purchase or lease warehousing/distribution facilities ﬁrst, and then expands its
outlets based upon the capacity of distribution center (e.g., 50-100 outlets per
distribution center). This suggests that inefficiencies or failures in the supply and
distribution markets induce certain retailers to integrate these functions to reduce
overall transaction costs. However, because operating a distribution center

requires specialized management skills that are different from retail operations,

24

ﬁrms must either hire experienced managers at a premium or invest in
management training to derive similar productivity beneﬁts associated with their
core business. Arguably, by developing private label programs and investing in
product design staff and outsourcing manufacturing, retailers are also engaging
in a form of backward integration. While ﬁrms may be willing to invest in human
and physical resources in order to enter a market, by taking on another role in
the supply chain, on average, their ﬁxed cost structure should be higher than for
retailers pursing a related diversiﬁcation strategy. In turn, the expected payoffs of
unrelated diversification should come in terms of higher margins and profits
relative to related diversification.

Several factors limit the potential benefits of diversiﬁcation. Foreign
subsidiaries that diversify into product areas that are different from their parents
have lower long-term survival rates (Li, 1995). Spatial dispersion, the extent of
differences between businesses within a corporation, and the complexities of
coordinating activities in a large portfolio can increase marginal management
costs (Bergh & Lawless, 1998). As businesses and activities become more
diffuse, information processing capabilities become strained, increasing the costs
of internal governance (Tallman & Li, 1996). Jones and Hill (1988) also argue
that the bureaucratic costs associated with coordination and cooperation can be
higher for related diversiﬁers relative to unrelated diversifiers due to increasing
agency costs associated with resource sharing and joint production as
corporations grow. Rugman and Verbeke (2005) also argue that MNCs succeed

only when they have the ability to adapt the ﬁrm’s speciﬁc advantages to the

25

requirements and opportunities present in the host country. Moreover, they argue
that adaptation is less costly in home regions versus outside the MNC’s home
region. As the costs of coordinating activities within the ﬁrm rise, the potential
beneﬁts of diversification are nullified. Thus, although the empirical evidence
suggests that, on average, related diversiﬁers should outperform unrelated
diversifiers, Varadarajan & Ramanujam (1987) caution that this may not be a
prescription for every type of firm.

Yet others argue that there is a curvilinear relationship between degree of
diversiﬁcation and performance (e.g., Markides & Williamson, 1996). Ansoff
(1969) traces product-market life-cycle curves, arguing that strategic
diversiﬁcation decisions can help ﬁrms stimulate and expand demand in existing
and new markets. Peteraf (1993) outlines the economic arguments for why ﬁrms
may face optimal levels of diversiﬁcation, depending upon the speciﬁcity of their
resources and the resultant range of strategic options possessed. The argument
is that beyond this optimal level, ﬁrms should face diminishing economic rents as
they enter additional product markets. Combining product life cycle curves and
optimal points of diversification on rents, ﬁrm performance should initially
increase with higher levels of diversiﬁcation. At some mid-point in the extent of
diversiﬁcation, performance will peak and then begin to decline.

Mixed empirical results also suggest limits to diversiﬁcation. In a meta-
analysis exploring some of the divergent ﬁndings in cross-sectional studies,
Palich et al. (2000) ﬁnd a curvilinear relationship in product diversiﬁcation and

performance. These studies suggest that diversification helps to a point, and then

26

the costs of diversiﬁcation become higher than the realizable beneﬁts. In
studying corporate refocusing, Markides (1992) ﬁnds that firms can diversify
beyond an optimal point, after which they see positive, but declining marginal
retums. Declining returns often prompt ﬁrms to disengage in parts of their
business that are distant from their core competencies. Gary (2005) provides
some evidence about how shared corporate resources may become

. overstretched at a certain point leading to the lack of value creation, particularly if
, investments in these shared resources are not maintained adequately. But it is
unclear whether the curvilinear relationship would hold in a customer segment
and/or retailing context.

H ypothesis1 : At lower levels of diversiﬁcation, FBU performance should increase.

At middle to high levels of diversiﬁcation, FBU performance will peak and then
begin to decline.

Governance

Governance decisions represent a critical element in business strategy.
Mode of diversiﬁcation refers to the governance mechanism chosen to enter into
the new line of activity. Firms choose the most efﬁcient mode to protect their
speciﬁc assets in the presence of environmental and behavioral uncertainties
(Williamson, 1985). The range of control mechanisms run from the low control
mode (i.e., market) to high control mode (i.e., hierarchy) with hybrid choices of
licensing, franchising and joint ventures.

Management of a particular mode may also represent a unique capability

(Barkema, Shenkar, Vermeulen, & Bell, 1997). For example, Carrefour's

27

President of China Operations, noted the ﬁrm’s expertise in organic (wholly-
owned subsidiaries) growth, and made a distinctive departure from past strategy
to institute an expansion place that features mergers and acquisitions (M&A)
(Child, 2006). In summary, ﬁrms can build expertise in managing governance
modes.

Service ﬁrms have a range of international entry mode choices and
international business strategies. Arguably, services have the same basic set of
entry mode choices as manufacturers (Erramilli & Rao, 1993). Agarwal and
Ramaswami (1992) argue that hard services have the same type of entry mode
choices as manufactured goods. Critics would argue that services have a
constrained choice set because of necessity of physical presence (e.g., export
not an option) (Ekeledo & Sivakumar, 1998). Further, it is unclear whether mode
choices are stable. Burt (1991) shows that in the European grocery sector, ﬁrms
tend to stick to the same modes, whereas Picot-Coupey (2006) documents
variations in mode choices in the French clothing and accessories sector. Firms
also have an opportunity to learn from their experience. Lu (2002) ﬁnds that
previous entry modes inﬂuence future entry mode decisions. In the service
sector, businesses appear to prefer high default modes, such as subsidiary
ownership, instead of a market default assumed for manufacturers (Doherty,
1999; Erramilli & Rao, 1993).

Once made, governance decisions are difficult to change and erroneous
controls are costly to manage (Vlﬁlliamson, 1996). But, on balance, normative

prescriptions found in the TCA-based literature on entry mode choice appear to

28

lead to higher performance (Brouthers, 2002). However, these mode choices
only represent a component of diversification strategy.

Entry mode is often underplayed in much of the strategic diversification
research (Datta et al., 1991; Hoskisson & Hitt, 1990). Earlier work on
diversiﬁcation put a considerable amount of focus on M&A. Within the foreign
market context, this process exists when a foreign firm acquires its actual and/or
potential competitors. Within an organizational leaming framework, both
acquisitions and joint ventures can be viewed as opportunities for exploration in
which ﬁrms attempt to learn from the experiences of others (Slater & Narver,
1995). Vlﬁthin efficiency motives cited for M&A, Trautwein (1990) highlights
ﬁnancial, organizational and/or managerial synergies that are presumed to result
from M&A activities. As such, M&A’s are enacted to gain access and control of
another ﬁrm’s resources and capabilities in anticipation of the creation of new
and/or expanded synergies with an acquirer’s existing capabilities.

Much of the diversiﬁcation literature focuses on mergers and acquisitions
and organic growth options. Because diversification is fundamentally about how
to share internal capabilities to create synergies between intemal businesses or
products, full control modes have been emphasized in the literature. In part, this
is because competencies are considered difficult and costly to transfer. However,
Lord and Ranft (2000) ﬁnd that managerial incentives tied to corporate
performance encourages transfers of local market knowledge from other
divisions already operating in the host country to a newly-entering division. In

addition, Tanriverdi and Venkatraman (2005) conceptualize how multi-business

29

ﬁrms can utilize their production cost synergies generated from several kinds of
knowledge relatedness (i.e., about internal products, external products,
customers, risk and investment, and alliance management) to create value
synergies leading to greater ﬁrm performances.

Empirical research is equivocal. Stock market analysts and investors tend
to positively value mergers, although, on balance, the acquired company’s
shareholders seem to gain signiﬁcantly more than the acquirer’s shareholders
(Caves, 1989; Ravenscraft & Scherer, 1987)6. Often measuring short-term event
windows, long-term performance is not considered, despite the assumption that
synergies require time to bear fruit (Woodcock, Beamish, 8. Makino, 1994).
Other researchers demonstrate that acquisitions do not result in increased
profitability for the acquiring ﬁrm (Fowler & Schmidt, 1989; Herman &
Lowenstein, 1988; Rumelt, 1986). Merged ﬁrms are often unable to realize the
synergies envisioned in the pre-M&A process due to unanticipated integration
challenges (Tanriverdi & Venkatraman, 2005). For instance, divergent
organizational cultures make lines of communication and coordination less
effective than required for synergy creation (Chatterjee & Blocher, 1992). But, it
is unclear whether M&A strategies may work better for some industries as

compared to others. Since retailers often choose M&A as a favored form of

 

5 Interestingly, the authors acknowledge the importance of alliances in generating synergies; yet
they only refer to the management of these alliances as the important knowledge component (i.e.,
staying within the borders of the ﬁrm). In other words, they stop short of recognizing knowledge
(or other) complementarities may reside within the alliance partner.

But these ﬁndings should taken with some caution. A potential limitation in extending these
ﬁndings in an lntemational context is that a fundamental assumption in the ﬁnance studies is a
competitive capital markets, which may not necessarily exist in all emerging markets.

30

governance, it is important to determine whether M&A is an effective strategy
within this industry in a foreign market context.

However, there is evidence that corporations do not consistently follow
“pure” diversiﬁcation strategies across their FBUs (i.e., either all organic growth
or acquisition) (Lamont & Anderson, 1985). In other words, managers assess the
best mode given internal and external factors. Moreover, ﬁrms pursuing mixed
strategies perform no worse than ﬁrms following “pure” strategies (Lamont &
Anderson, 1985).

Either by regulatory design or by choice, ﬁrms may also choose to enter
into hybrid governance forms. Hybrid forms represent a mid-point on the
governance continuum between wholly-owned subsidiaries and the market
(\Mlliamson, 1985). Since synergies are tradeable (Kay, 1982), firms can
purchase access to knowledge and knowledge-based resources through

alliances and partnerships (Markides & Williamson, 1996; Simon, 1991). Over
time, ﬁrms acquire knowledge from their partnerships and this knowledge
becomes a potential source of synergy (Tanriverdi & Venkatraman, 2005). For
example, franchise networks demonstrate the value of external partners to adapt
to diverse markets, provide learning opportunities, and achieve greater returns
for the organization (Sorenson & Sorenson, 2001). Hybrid governance forms are
chosen when: a) firms have a resource deficiency, b) suitable partners are
available, and c) when the firm trusts that either the partner does not have the
ability to acquire its core resources and/or would not act opportunistically

(Woodcock et al., 1994). In sum, ﬁrms utilize hybrid modes with the intention of

31

combining complementary capabilities in an effort to generate new
competencies, solve resource deﬁciency challenges, and potentially develop new
applications from these capabilities (Teece, Pisano, & Shuen, 1997; Woodcock
et al., 1994).

Although joint ventures also represent a common expansion strategy for
retailers, scant research exists on this governance strategy (Palmer, 2006).
When retailers are legally obliged, lack an important resource, or a host country
partner has a high value asset speciﬁc to its particular national market,
multinational retailers may choose to enter into a joint venture with the local
partner (Hennart, 1988). For example, multinational retailers may not have
access to the best locations, which are controlled by local or other foreign
competitors. Potentially attractive partners can also augment weak or non-
existent MNC capabilities (Mowery, Oxley, & Silverman, 1996). This can provide
invaluable assistance since one of the largest costs of entry is the investment in
knowledge about the local market (Greve, 1996). Barkema et al. (1997) ﬁnd that
ﬁrms can incrementally learn from their lntemational joint venture experience as
long as their foreign operations are related to their core business. Since
acquisition, absorption and integration of “unfamiliar” information is a costly and
gradual process, which may put the ﬁrm at a competitive disadvantage in the
new market relative to existing competitors (Penrose, 1980), foreign partners with
complementary knowledge enables firms to pursue internationalization.

Furthermore, Palmer and Owens (2006) argue that joint ventures may be an

32

attractive governance mechanism precisely because it is a lower level
commitment of resources as compared to organic growth strategies.

The dangers of partnership must also be mentioned. By engaging in
business activities with a partner, one side may appropriate a disproportionate
share of quasi-rents by discovering the other side’s knowledge over time (Baum,
Calabrese, & Silvennan, 2000; Cohen & Levinthal, 1990). Studying alliances in
North America, Europe and Asia, Dussauge, Garrette and Mitchell (2000) ﬁnd
that alliances in which partners contribute complementary capabilities result in
higher levels of Ieaming and capability acquisition as well as higher probabilities
of partner takeover or reorganization than in alliances in which partners possess
similar capabilities. Without systems for protecting trade secrets, these same
partners may also become future competitors. Another potential hazard is
multiple, key stakeholders. The FBU may experience potentially conﬂicting
demands from its owners that may be detrimental to its efficiency and
effectiveness (Prahalad & D02, 1987). In summary, by sharing control and trade
secrets with another business, foreign retailers also face some risks.

Numerous authors have studied the performance beneﬁts associated with
various entry modes (e.g., Erramilli & Rao, 1993; Woodcock et al., 1994).
Eramilli and Rao (1993) propose that the default mode of service ﬁrms is
ownership. This is due to the nature of the intangible assets producing the
monopolistic advantages that motivate ﬁrms to intemationalize. The difﬁculties in
codifying and transferring (intangible) knowledge also create the need for

internalization because licensing through market transactions involves high

33

costs. However, Beamish & Banks (1987) argue that trusting joint venture
partners can outperform wholly-owned subsidiaries. In addition, controlling for
product diversiﬁcation in a domestic context, Simmonds (1990) finds that new
ventures outperform acquisitions. In a study of Japanese entrants into the US.
market, Woodcock et al. (1994) ﬁnd that new ventures outperform joint ventures,
but that joint ventures outperform acquisitions. Moreover, Li and Guisinger (1991)
ﬁnd that joint ventures experience failure rates at a level between the more
successful new ventures and less successful acquisitions. In summary, the
evidence seems to suggest that wholly-owned subsidiaries should outperform
joint ventures, but that joint ventures should outperform acquisitions.

Why would one mode outperform another? While most foreign entry mode
research focuses on ex ante costs that may determine mode choices (e.g.,
selection costs), few entry mode studies focus on ex post costs associated with
operating F BUs after the mode selection is made. Woodcock, Beamish and
Makino (1994) explain the performance differences in terms of the costs
associated with each entry mode. The authors make an argument that new
(organic) ventures have low costs of procuring additional resources as well as
low ownership and control 'cost. The new venture utilizes existing resources
required for success in the new market and appoints managers, who understand
and can operate effectively within the firm’s culture and bureaucracy. In contrast,
acquisitions have high levels of both costs since, despite aggressively optimistic
assumptions to the contrary, acquired firms rarely integrate ﬂawlessly with the

acquirer, severely limiting the realization of synergies, often creating resource

redundancies (Chatterjee, 1990) and complicating management control tasks.
Joint ventures are posited to have medium procurement and high
ownership/control costs. Although both ﬁrms must pay for the acquired
complementary resources, the confusion and difficulties created by multiple
stakeholders with potentially conflicting goals makes the costs of ownership and
control higher than wholly-owned subsidiaries. For both acquisitions and joint
ventures, if the rationale for purchasing synergies is the primary reason,
managing heretofore unknown businesses should increase the costs of control.
Efficiency arguments suggest that total costs should be the highest for
acquisitions, all else equal.

Retail internationalization literature suggests that mergers and acquisitions
represent the most common modes of entry for international retailers into key
emerging markets in Latin America, Asia and Eastern Europe (Coe, 2004). In
part, this is attributed to limitations in some countries imposed upon organic
expansion of large-scale formats, which forces retailers to choose other means to
expand. Firms tend to prefer acquisitions when partners with salient capabilities
are not willing to risk sharing their source of advantage with an outside partner
(Woodcock et al., 1994). Large-scale retailers with core capabilities that tend to
be in the form of back office functions and supply chain partnerships may be less
willing to risk the sources of their advantage by engaging in joint ventures. But
the ﬁndings on performance of post-acquisitions and synergy sharing are
disappointing (Caves, 1989). For instance, Chatterjee (1990) finds that, instead

of synergies, acquisitions create redundancies, increasing the costs of intra-ﬁrm

35

coordination. This may help to explain why retailer internationalization results
are often disappointing.

Hypothesis 2: In foreign markets, wholly-owned subsidiaries should outperform
joint ventures, which should outperform acquisitions.

Previous diversiﬁcation researchers studied groups of businesses
following similar strategies and examined their performance (e.g., Busija et al.,
1997). We extend this idea by looking at ﬁrms that utilize similar strategic
diversiﬁcation conﬁgurations (e.g., direction and mode) and examining how these
conﬁgurations lead to performance.

By expanding the governance continuum, interesting possibilities arise in
terms of interactions between governance and direction of diversity. The
underlying assumption is that we have three sources of risk with which to
contend: 1) the risk of not competently competing in areas unrelated to the
corporation’s core competencies, 2) the risk of not effectively managing with
unfamiliar governance modes, and 3) foreign market operational risks. By not
leveraging existing resources, the FBU misses the opportunity to create
synergies, such as economies of scope, among their businesses. The lack of
interdependence between businesses creates a weakness relative to other ﬁrms
that are able to leverage their excess resources across business. Additionally,
given that managers tend to possess imperfect information and analyses about
foreign markets and partners (Aharoni, 1966; Kobrin, 1982), the risk of employing
low control modes increases. Furthermore, there is an interaction effect between

governance and diversiﬁcation strategies. We would expect that FBUs pursing

related markets with organic growth should be the highest performers. Busija et
al. (1997) provide some evidence that ﬁrms pursuing related-constrained
diversification strategies and internal growth outperformed other strategy mix
types. At the other end, due to the multiple sources of risk, we would expect the
lowest outcomes to be associated with unrelated diversification strategies with
acquisition. In a study of large US. firm acquisitions, Porter (1987) ﬁnds that
three out of four unrelated acquisitions failed. The costs of absorbing another
entity with its own organizational culture and trying to manage a new line of
business should be a lethal combination. Finally, there is evidence that joint
ventures may work well in both related and unrelated diversification. On one
hand, the corporation could be buying access to complementary synergies held
by its partner. On the other hand, parents may be able to manage and share its

resources in a joint venture that operates in its line of business.

Hypothesis 3: Related diversiﬁers that expand via organic growth should
outperform all other strategy mixes. Conversely, we expect that unrelated
diversiﬁers that expand via partnership or acquisitions to be the lowest
performers.

When we consider the effectiveness of diversification decisions at the
domestic level across nations, forces that would impede or strengthen the local
level implementation of these decisions would be of interest. In retailing,
competition is primarily a local phenomenon (Porter, 1987). Although multi-
market competition has been discussed in the literature (Fuentelsaz & Gomez,
2006), for a largely location-bound retailing function this appears to be less

relevant. This has important implications in studying the strategy-perfon'nance

37

relationship. First, theory and measurement should take into account multiple
levels of analysis (Klein, Dansereau, & Hall, 1994; Mishra, Heide, & Cort, 1998).
The competition takes place in host country markets, which have important
contextual variations that may affect the implementation of the diversiﬁcation
strategies across nations. In addition, the firm’s overall international
diversification strategy should be treated as a portfolio of its country market
strategies. In other words, the firm’s foreign portfolio consists of individual market
operations, each inﬂuenced by varying levels of institutional factors. If we are to
avoid bias in our parameter estimates and inaccurate interpretations from our
models, we need to separate the foreign business unit (F BU) country strategy
from the corporate and national factors. Theoretically, corporate and national
conditions are more appropriately considered moderators of the relationship
between strategy and performance as they hinder/help the implementation of the
diversiﬁcation strategy.

In summary, we hypothesize about within-country/corporate group and
format performance. Our ﬁrst hypothesis addresses the limitations of
diversification strategy. After initial gains from diversifying operations, retailers
should see a plateau and drop off of performance outcomes. Our second
hypothesis relates to the relative efficacy of separate governance modes,
predicting WOS outperforming joint ventures and acquisitions. Finally, our third
hypothesis looks at the relative efﬁcacy of six diversiﬁcation strategy

combinations.

38

CHAPTER III
METHODOLOGY

In this section, we discuss our key variables, sampling, and research
method. Our construct operationalizations are deﬁned and measures described.
Descriptive Statistics of our key variables are contained in Appendix A. Our
sampling plan is outlined. Finally, we describe the analytical tools and

procedures, supporting hypothesis testing.

KEY VARIABLES

Our choice of key variables to test our hypotheses was guided by four
criteria. First, within the MNC research context, these measures represent
important constructs for our model. Second, we are guided by the extant
business literature. Third, we expect to find significant differences between the
retailers and countries on the selected variables. Finally, we have reason to

believe that relationships between constructs should exist.

Endogenous Variable

Previous studies of country, industry and ﬁrm affects on business
performance have used various ﬁnancial (e.g., ROA, ROS, profitability) and
strategic (e.g., market share) measures (Mascarenhas, 1992; McGahan & Porter,
1997; Rumelt, 1991; Schmalensee, 1985). While Schmalensee (1985) throws
down the gauntlet in support of the I-0 view that industry structure drives ﬁrm

performance (via firm conduct), more recent studies, such as Rumelt (1991),

39

provide convincing arguments and evidence suggesting that business-level
effects are much more important than industry factors in explaining variance in
ﬁrm performance. In studying the performance of foreign affiliates, Makino, lsobe,
& Chan (2004) find that foreign affiliate and corporate effects explain the most
variability in performance, but that country and industry affects also help explain
performance, particularly in developing country contexts. Moreover, the authors
ﬁnd signiﬁcant interaction effects between industry and country.

Two main difficulties in using this measure in an F BU-performance context
stem from differences in national accounting systems and structural differences
between industries and across countries7. We control for the second problem by
limiting our examination to the retail industry. The ﬁrst problem can only be
addressed by finding another performance measure. The one measure that
exists across countries, ﬁrms, and formats is market share. Absolute market
share is appropriate in studying a single industry because the sum of shares
would equal to 100% and the range would be between 0 to 100%. Several
variants of market share can be tested, such as unit vs. sales figures, and single
year vs. multi-year average. Szymanski et al. (1993) find that single year
estimates of market share elasticity tend to be lower than multiple-year averages.
In addition, the authors note that depending upon industry and product category,
price (unit) measures may not produce similar results, but they did not have
enough data to test potential moderating effects. Moreover, we use an outcome

variable that proceeded in time data collected for the exogenous variables.

 

7 Other approaches to modeling format relationships are to create binomial or tn'nomial choice
models, examine store patronage, or household expenditures (e.g., Fox, Montgomery and Lodish
2004; Goldman, Ramasaswami and Kn‘der 2002; Morganosky 1997).

40

Because the effects of diversification decisions would not be immediate, it is
prudent to allow for the passage of time into the model. Although temporal
separation between independent and dependent variables does not signify
causality, it is a necessary condition for drawing inferences. We also test the
time-invariance of our variables by testing the models with a dependent variable
from the next year. For example, if we are testing performance in 2003 related to
strategy choices in 2002, we will also test the ﬁnal model against performance in
2004. For the covariates, this helps provide information about the lasting impact
of speciﬁc strategic and structural characteristics. In summary, we can test
market share using multiple measures to determine whether the temporal nature
or unit of our measures create differences in the parameter estimates.

Critics of a market share performance measure note contradictory ﬁndings
on the relationship between market share and profitability and suggest that the
relationship is a measurement artifact (e.g., Jacobson & Aacker, 1985). In a
meta-analysis of the market share-proﬁtability relationship, Szymanski,
Bhaaradwaj and Varadarajan (1993) find a positive relationship, but also show
signiﬁcant moderators, such as specification errors, sample characteristics, and
measurement characteristics. Using a multilevel model on retail speciﬁc data
corrects for some of these issues. Ultimately, their analysis did not ﬁnd deﬁnitive
evidence disproving the positive relationship.

One of the criticisms of the diversification literature is that many of the
previous studies have not fully speciﬁed their models, creating uncertainties

about the generalizability and validity of the body of work (Datta et al., 1991). We

41

attempt to identify variables at the FBU, corporate and country level that may

impact market share and may potentially also impact the diversification decision.

FBU Level Exogenous Variables

Diversiﬁcation is measured by an entropy measure for the current retail
operations of corporations by country. We adapt the Jacquemin-Berry entropy
index by categorizing retail formats using the US. Census Bureau’s 6-Digit
NAICS Code system. Because we are studying intra-industry diversification, we
need a ﬁner grained approach to differentiate between businesses. According to
Palepu (1985), the entropy measure is directly adapted to the text found on
pages 14-15. In general, total diversification scores range from 0 (no
diversiﬁcation) to approximately 3 (high diversiﬁcation) (Bergh, 1995, p. 1698).

Operational definitions of governance modes are designed to keep the
modes conceptually distincta. Wholly-owned subsidiaries are ﬁnanced, owned,
and operated by one parent ﬁrm from the beginning of its existence. This is
sometimes referred to as “organic” growth. Acquisitions are defined as local
companies where more than 50% equity was purchased by the foreign
corporation. Joint ventures (JVs) are defined as new entities created, owned, and
operated by more than one parent. JVs are not necessarily assumed to be

‘balanced’, whereby there is a perfect sharing of resources among the owners.

 

8 In an interesting test of endogeneity in entry mode choice, Shaver (1998) shows that ﬁrms
actually select entry modes based upon their assessment of observable and unobservable
capabilities and expected performance outcomes. In contradiction to the entry mode literature, he
refutes the notion that one type of entry mode is necessarily better than another (e.g., organic
growth vs. acquisition) based upon observable capabilities.

42

The JV partners must include at least one foreign corporation, but other partners
can be either foreign or domestic.

Other F BU level covariates are the size of operations within the country,
format types, and age. Size of operations relates to total number of outlets, while
formats (i. e., h ypennarkets, supermarkets, convenience stores, and discount
stores) are operationalized as dummy variables. ln-country experience (i.e., age)
refers to the number of years that the format has been operating in the host
country. Two other dummy variables will be used. The dominant business of the
FBU refers to the business that brings the highest percentage of sales within a
speciﬁc country. Relatedness refers to the membership in the dominant segment
for an FBU in a speciﬁc country. It is not assumed that the dominant business in
a foreign country has to be the dominant business in the home country, but this is

considered likely.

Corporate Level Characteristics

RBV posits that resources and capabilities are heterogeneously
distributed across ﬁrms and long lasting (Barney, 2001). Thus, we need to look at
important knowledge capabilities in the parent organization as well as how these
capabilities may be transferred to the FBUs.

Level of Centralization is an important structural factor inﬂuencing the
implementation of FBU strategy. The Resource Based View (RBV) argument
hinges on the assumption that information is passed between business units

coordinated by corporate managers. In other words, corporate managers can

43

manage the vertical information flows necessary to recognize and realize
synergies across F BUs. Gupta and Govindarajan (1986) argue that the
effectiveness in which linkages between business units is managed will affect the
potential realization of corporate synergies. Subsidiaries of MNCs have common
reporting standards and are periodically audited (and rewarded) for compliance
to these standards. Reuer and Leiblein (2000) show that management systems
moderate the relationship between international investments and organizational
downside risk. Corporate policies of maintaining the appropriate levels of
organizational slack before and after a diversiﬁcation move is an important factor
in fully realizing synergies of shared resources leading to business unit success
(Gary, 2005). Moreover, organizations with centralized decision-making and
funding should exert a stronger influence on local operations than decentralized
organizations (Scott & Meyer, 1983). Synergy transfer requires a sufﬁcient
degree of centralized management systems and control in order to function.
Although some fairly decentralized companies can employ knowledge
management systems similar to Ahold’s, corporate synergy creation and transfer
is still assumed to be weaker for decentralized companies. A dummy variable will
be constructed from the corporate reports and Euromonitor reports to distinguish
between centralized and decentralized firms.

International Experience (i.e., international diversiﬁcation) represents a
source of knowledge on how to conduct business overseas. Operating in diverse
environments decreases the firm’s liability of foreignness by providing exposure

about constituent aspects of the environment and how they may affect

implementation (Hymer, 1976; Zaheer, 1995). Moreover, the ability to manage
across diverse cultures can aid in the development of new capabilities (Barkema,
Bell, & Pennings, 1996). These new capabilities can be used to compete more
effectively against local competitors (Delios & Beamish, 2001) and develop new
market positions (Greve, 1996). Finally, more internationally experienced ﬁrms
are inﬂuenced less by institutional forces. Experienced firms may learn how to
select more favorable environments by knowing which institutional hazards to
avoid (Henisz & Delios, 2001).

The measure of international experience needs to account for the potential
of differences in effects with varying levels of time, i.e., potential curvilinearity.
Although the moderating effects have not been studied, the cognitive limitations
argument could be applied also for managing across many different countries. In
other words, the beneﬁts of international experience should decline after a point.
This would indicate to an inverted ‘U’ shaped effect.

Firm size moderates the relationship between degree of diversiﬁcation and
performance (Jahera, Lloyd, & Page, 1987). Firm size is measured by the
number of employees worldwide. Size positively affects the types of modes firms
can choose, but may also encourage managers to take more risks than if they
were making decisions for a resource-constrained firm (Morck, Schleifer, &

Vishny, 1990).

45

Country Level Characteristics

To counter arguments that diversiﬁcation performance is only an artifact of
market or industry factors, we need to include four important covariates. First, we
should introduce industry concentration ratio to determine whether there are a
few companies controlling the market. A four-firm concentration ratio captures
the percentage of sales controlled by the four largest ﬁrms in the retail industry.
High concentration implies potential barriers to entry for new ﬁrms. There is some
speculation that larger formats might be in a position to grab higher market share
in countries with fragmented retail markets (Colla, 2004). Second, higher market
growth may attract more firms into the country. To make sure that our results are
not due to industry artifacts, we need to control for the differences between
different market environments. Format Size is measured by the total number of
outlets within a format for a country, where as Retail Employment relates to the

total number of people employed in the retail sector.

 

3.3:? E64. 86.2.1: £25233...” 3:26.262; 6292:: £552.63
86.9 £92.19 E2 6239 6966393 8:26.265; 621.2 183.2%? 2:830

32me
8235295

 

.§§_e§8ﬁcgﬁem§m§88§§.§§_88§$mﬁ
63811138832 wgggggecmoegmegﬁgvmoaz
.gmgbmmﬁgmuorugmmgﬁg Eggnogtgtmga
l
A1:r_1Nu8”8§8mm5
2
.ﬁcuzﬁcsssﬁsegeuowagbmmmmegecsgemo
1
1§mu19§
znzammémrmcoaaﬁﬁcéozmgwgéz 383815 568:.

._558§_5c§§§m§88_§§§§§8§8m§§8 E 3
4883689823; Emcoabﬁcecbmmuge 1

A11Fr_1NuE

.mgmgs$g_§b:5§o_§

989813818925; wgcﬁbﬂmugsguabosmaﬁéﬁ
8§§§_§c§=§bﬁmm§§;§§§b§mﬁ 113s

8+8” A151Nn E

cozmoEmLmzo

 

Dmn.

 

 

:ozmﬁﬁcozﬂono

 

. ouoo

 

 

9.68 28 up 2%»

47

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3250 :31: :0 52:3: ESP $250 EBay:
mum
Amucwmsos :c EoEBEEm :91: :38. 63.5.2 :83
Sims.
oz 8 weir ”139:9 >EE30 m:_m:o::m
#5550
91:30 :91: :0 55:5: .90... 91:30
Amucmmaoé ca 3983::1 v6 59:5: ESP mum
5212:2120
mcozmﬁao :91: 5:5 mocEsoo v6 :wnEzz 5:033:25
oz 8 3% w Hm_nm_:m> >EE30 :o_:mN__m:Eoo
mo:m_:maxm
31:30 :91: 9:38on 98> .6 59:32 =Smm
:oonEoO
Beam
woﬁm :mx:m:: :91: .29 .6 39:10:11 m mm 3m“. :0: 86m :32. Swims. =93:
ucmwzoc: :_ :mm; 3 11319250 :0 :onEzz £250
92w
mo:m_:m>:oo _:m:::oom_u Em: .:9::oom_u tow .:mx:mE:mQ:m 435851? 61339:? >EE:Q BEE“.
mo:m_:qum
F + 3:61 tSw bacsooé -Noomv 31:80:.
$12me
851: 62 8 $263: EmEEou 2 86.1: :_ C uo_nm_:m> >EEso 86.1w.
815me
oz 8 woke -- 13:38 :_ 928 S393 Uo_nm_:m> >EE30 EmEEoD
3:22: .cw 000.: mew 3.3 :38. mm:< m:___mw
mcozﬁsgm .mm:3:m> E_o_ .mmtmﬁﬁnzm 683926;; :2 mm_nm_:m> 15:50 mocm::m>oo

 

 

 

F 035... “5:00

 

48

SAMPLE

This study considers the diversiﬁcation choices of multinational food
retailers. This group is appealing for several reasons. First, food retailers
represent almost 60% of the world’s largest retailers and market their products
via a range of formats (Deloitte 2006). In addition, to avoid large variations in
product categories and markets, potential confounds, this allows us to deﬁne our
sample on the basis of a common market.

Second, although intra-industry diversiﬁcation studies are rare (Li and
Greenwood 2004), we follow previous research that considers organizations in
an industry as constituting a population (Hannan & Freeman, 1987). Moreover,
we argue that MNC food retailers represent an important sub-population in which
members possess similar strategic advantages and compete for similar
resources in the environment. By doing so, we avoid some of the commonly
stated methodological shortcomings in conventional diversiﬁcation literature
associated with cross-sectional studies, such as inappropriate pooling (Bass,
Cattin, & Wrttink, 1977) and wide variation in resource bases or investments
between sectors (Bettis & Hall, 1982). This is particularly true in previous
diversiﬁcation studies that often exclude service industries in their samples
(Nayyar, 1992).

Third, although global retail diversiﬁcation is an under-researched topic,
MNC retailers represent a growing presence and power in every developed and
many emerging markets (Doherty, 1999). In the 2006 Fortune 500 list, Wal-Mart

is the second largest US. company after Exxon-Mobile with 97 other specialty,

49

general merchandisers, and food & drug retailers representedg. However, little
research exists on how service-based MNC strategies function in an international
context as opposed to a domestic, manufacturing context.

In explaining how retailers begin to achieve customer loyalty, Cortsjens &
Cortsjens (1995) elucidate the four major differences between manufacturers and
retailers. First, most fast-moving consumer goods (FMCG) retailers are bound to
their various locations. For bricks-and-mortar retailers, the circumference around
each location deﬁnes both the majority of their customers and their competitors.
Second, manufacturers have traditionally been able to generate cost synergies
across product brands, while this is a challenge for retailers. Related
diversiﬁcation of retail formats is one avenue retailers can take to attain some
synergies from their operational know-how, private labels, and/or exclusive rights
to manufacturer brands. Third, retailers typically have high fixed costs and low
margins, whereas F MCG manufacturers have lower fixed costs and higher
margins. Consequently, as opposed to manufacturers, traditional retailers are
more focused on sales volume and pricing to make the difference in their
proﬁtability. Understanding how to get optimal retail mixes for many different
environments is one of retailers’ biggest challenges. Fourth, pricing strategy can
inﬂuence consumer perceptions about the competitiveness and value of the
store. Consumer perceptions are inﬂuenced by many factors and are not

constant across national borders (Cortstjens, Corstjens, & Lal, 1995).

 

9 www.fortune500.com

50

Since we are concerned with retailing, our approach is to consider only the
foreign market operations of retailers, as opposed to a broader cross-sectional
study of many sectors. We describe these FBU’s in two ways. First, we
aggregate FBUs for a particular ﬁrm in a country to test group-level hypotheses
on diversiﬁcation. If a group operates four formats, its total diversiﬁcation score
captures and weighs relative contributions of each format. The outcomes studied
become diversification strategies across countries and corporations. Second, we
look at each FBU with respect to the entry mode chosen for each format in a
country. In this case, we use a relatedness measure to capture diversiﬁcation
within and away from the dominant industry segment for that F BU. We are
interested in being able to sort out the entry mode-performance relationship,
controlling for other effects. Finally, we test the relative efﬁcacy of six format
diversiﬁcation strategies, controlling for important covariates.

To test the proposed models, we gather secondary data from several
sources. Our primary source of data is from Euromonitor, which has detailed
information about 85 global retailers operating in 52 countries. Euromonitor uses
a combination of surveys and national statistics to derive their databases. These
retailers typically represent the largest retailers in the country. We also
supplement our Euromonitor data with corporate annual reports, MINTEL and
GMID, as necessary and when available.

Our sample is drawn from Deloitte Touche Tomatsu’s 250 Global Powers
in Retailing 2005. Since 104 firms on this list do not have an international

presence, our sample reduces to 146 firms. Of these 146 ﬁrms, Euromonitor has

51

information on at least 29 global corporations that sell food internationally. Our
observations are the business level entities of these global retailers in the
countries in which they operate. Within a country, a corporation may have
multiple brands, started at different times. Thus, our sample should contain a

range of business units established at different times.

Figure 1: Breakdown of Global Retailers

TOP 250 RETAILERS. BY SALES

United States
4 7 7%

Germany
11 0%

France
9.7%

 

Source: Global Powers of Retailing 2005

Some authors argue that archival industry data is inappropriate for use in
measuring strategic competitive advantages. For example, Levitas and Chi
(2002) argue that measurable tacit data quickly dissipates in an informed
marketplace, thus observable data cannot represent strategic competitive
advantages. However, Rouse and Daellenbach (2002) counter thatjust because
some tacit knowledge can be known and measured does not infer that others

could reproduce the multiple capabilities required to achieve the knowledge.

52

While we are cognizant of potential limitations of archival industry data, we
believe that the data are reasonable proxies for our constructs.

Because accurate global retail data are difﬁcult to collect, we were limited
in the number of countries included in the Euromonitor database. Of general
concern in a limited dataset is whether we would have a truncated data problem.
This is a specific kind of missing data problem when both exogenous and
endogenous variables are missing. For example, none of the countries in the
study would be considered for the United Nation’s Least Developed Countries
Fund. Instead, our sample over-represents the richest countries and the largest
emerging markets and does not represent the poorest or smallest countries at all.
But given the expansion policies under examination, the dataset offers a range of
realistic investment destinations. In other words, our countries are systematically
sampled and highly representative of the markets in which multinational retailers
operate. Thus, for our purposes, a full range of all possible countries is not
necessarily desirable. Adding countries to the sample that are not considered
likely investment choices for our corporations would pull our parameter estimates
away from their true mean, creating the backdrop for ﬂawed inferences and
policy recommendations. Our sample represents the most likely range of

countries in which MNCs reasonably invest with few exceptions or omissions.

53

METHODS

Conceptual Advantages

V\ﬁthin the extant research, there is a rich history of conceptualizing multi-
level inﬂuences in a variety of contexts and fields. In education, Lee and Bryk
(1989) examine whether Catholic schools provide a more socially equitable
distribution of school achievement than public schools. In human resource
management, McComb, Barringer and Boume ( 2004) demonstrate how
individual, work and organizational factors inﬂuence employee job-related
attitudes and behaviors. In management, Dranove, Peterof, and Shanley (1998)
develop multilevel theory and testing to prove the existence of strategic groups.
Addressing criticisms of single level models, data were chosen to provide clarity
in isolating the relative compositional and contextual effects that inﬂuence retailer

FBU performance (e.g., Duncan, Jones, & Moon, 1998).

Analytical Reﬁnements

Heeding calls for multilevel research from diverse literatures (Klein et al.,
1994; Madhok & Liu, 2006; Palmer & Owens, 2006; Rumelt, 1991) and driven by
theoretical concerns, methodological issues are also addressed. Because we are
considering FBUs partially nested within ﬁrms and countries, our study requires a
ﬂexible multivariate tool that can adjust for dependencies in the data. In addition,
Hierarchical Linear Modeling (HLM) tools can provide us with richer information
about speciﬁc corporate and country means and slopes within and between

markets by allowing country and corporate factors to vary (i.e., instead of being

54

fixed). Instead of summarizing the performance of all corporations in the sample
within one mean, we estimate separate means for each corporation. These
estimates are weighted by the amount of information we possess for each
corporation. In other words, instead of one market share mean for all
corporations with a wide standard error, we have specific means for each
corporation that vary according to the number of observations we have for the
corporation”. In weighting individual corporate means by their precision, HLM
models represent a signiﬁcant improvement upon conventional approaches to
linear modeling techniques that aggregate the behaviors of individual
corporations.

Biased estimates of parameters and standard errors result from not
considering differences in units of analysis and dependencies in the data
(Raudenbush & Bryk, 2002a). This potentially changes interpretations of the
evidence in the hypothesis testing stage. In addition, we can explore unique
business level strategies impact on performance, adjusting for salient
characteristics of business units, corporations and countries. Aggregation bias is
a serious concern. Simply pooling data from all F BUs in all corporations and
countries would likely produce an uninterpretable estimate, which is a blend of
many contributive factors. Thus, a study of F BU must account for heterogeneity
across corporations and countries. Another benefit arises in using FBU level
data. By classifying our observations by country and corporation, both geo-social

and organizational effects are considered.

 

1° According to statistical principles, if we only have one observation for a corporation, our mean scores are
less precise than if we have 30 observations for that corporation.

55

Given the risk that we may not have enough information to use a cross-
classified model, we can collapse our model and consider country and corporate
effects separately. While this does not allow us to directly compare the relative
contributions of corporate and country contexts, it does represent an
improvement on traditional linear modeling techniques. Thus, our model
represents a more realistic framework to study the degree to which compositional
(i.e., corporate) and contextual (i.e., country) factors explain the variation in FBU
performance (Zaccarin & Rivellini, 2002). By pooling information across FBUs
and contrOlling for country and/or corporate effects, we can develop more reliable
parameter estimates (Hahn & Doh, 2006). By using more powerful tools, we can
study retailing at a deeper level, gaining more information about how business
units operate.

Despite its usefulness, hierarchical linear modeling techniques are rarely
used in international marketing research. In a survey of international business
journals, Hult, Ketchen, Jr., Griffith, Finnegan, Gonzalez-Padron, Harmancioglu,
Huang, Talay and Cavusgil (2007) ﬁnd that less than 2% of the international
business articles employed multi-level data analysis techniques. However,
variance components tools have a relatively long tradition in the strategic
management literature, in part as a direct result of trying to account for multi-level
effects in the diversification research (Rumelt, 1991). Despite this history, multi-
level models have not yet become well known in business research. This study
uses cross-nested and Two-Level HLM frameworks in order to more realistically

model our phenomenon of interest and to isolate the unique compositional and

contextual forces inﬂuencing FBU performance. To our knowledge, multi-level
modeling remains rare in the marketing literature and cross-classified models are
not yet used.

By concentrating on one industry, this study will mitigate a complication
typical to cross-industry studies, which is that the heterogeneity of industry
structures makes it difficult to compare variance components among ﬁrms across
industries (Rumelt, 1991, p. 168). Although questions arise about the time-
invariance of diversification models, we also consider how to test this issue. By
using market share information from two subsequent years, we explore the
temporal nature of our key relationships. To our knowledge, there are no
empirical studies in the international marketing literature that examine how ﬁrm
and country factors inﬂuence the relationship between FBU factors and

performance in a multi-Ievel framework.

MODELS

Random effects models are more ﬂexible than classical analyses of
variance in that they have the ability to handle unbalanced data, covariance
components, and discrete as well as continuous covariates at multiple levels
(Raudenbush, 1993). In addition, HLM is more appropriate than other regression
based techniques because of the nested nature of the data. Retailer FBUs are
cross-nested within the corporations and the countries in which they operate.
Although FBUs of the same company share the same parent, they also need to

adapt to forces within the local retail market. We can also model the nesting

57

within either of these contexts separately, and it would be an improvement upon
traditional models.

We begin with a One-Way ANOVA with Random Effects to illustrate the
basic difference between this class of models and conventional, Ordinary Least
Squares (OLS) approaches. We start with this model, so that we can establish

the basis for embarking down this analytical path. In our model, this represents

the base model. At level one of the model, Y, = ,Boj + r”. where r”. ~ N(0,0'2) for

i=1,..., n]. FBUs in corporation j, and j=1 ...34 corporations. The 02 represents the
FBU-Level variance. At level two, or the corporate level, ,6,” provides a

benchmark of the average market share value for each corporation, which is

composed of the grand-mean for all corporations ( 700) and random error (qu ).

Thus, the basic model has a fixed effect (i.e., the grand-mean) and two random
effects at the FBU and corporate levels. In contrast to the One-Way ANOVA with
Fixed Effects, this model partitions variance between the two levels.

From this base model, researchers can choose whether to model the
second level more fully in terms of moderating factors to explain both the
intercept and/or slope terms as well as whether to allow slope coefficients to
vary. The modeling decisions are based upon theory and data speciﬁcations.

We chose to test the inﬂuence on FBUs of being nested in countries and

corporations separately". We employ a Two-Level HLM. We free the mean to

 

” We attempted to use a cross-classiﬁed model to determine the relative contributions of
compositional and contextual effects. While this model best ﬁts our conceptualization, our data
was too sparse data for reliable usage of this analytical tool.

58

vary across corporations, but fix our slope coefﬁcients. In fixing our slopes, we
show our expectation that the moderation effect is the same across corporations.
In sum, we are modeling the impact of higher level factors inﬂuence on the

means and relationships at the FBU level.

The General Model

Level 1: FBU Level
Yr) = £011” ﬂ1jx1ij+ ﬂ2jX2r'j +--- 1’ ﬂQjXQij + rt]

Y3. is the performance of each FBU i in country or corporation j;

,Bq. is the mean performance across countries or corporations ( [30}. ) and
regression coefﬁcients relating X J to nyor cell i

de are the FBU level strategies or characteristics for groupj in cell i

5.]. is the within-cell random effect, i.e., the deviation of FBU ij’s performance

from the FBU mean.
02 is the within-cell variance of rij

i =1,..., nj FBU within cell ij;
j=1 J: the number of countries or corporations, depending upon the model

Level 2 Country or Corporate Effect

ﬂqj = 7q0 +7q1WIj +7q2vv2j +"' +quqWqu +qu

700 is the model intercept, E(7q0) when all the explanatory values are zero
ka is the characteristic or strategy effect on the distribution of outcomes

within country or corporation j
qu is the random effect associated within country or corporation j

CENTERING
A choice of centering of predictors helps determine the intercept term for

each set of relationships in our model. The primary reason is that, without

59

centering, the intercept is an average value of the outcome variable when the
predictor is zero, which may not make sense (Luke 2004). At Level 1, we do not
center our data because a diversification score of zero is meaningful in
diversiﬁcation research. For example, Aldi and Lidl typically do not deviate from

' their hard discounter format, which would give them diversiﬁcation scores of zero
for sticking to one format. At Level 2, we use grand-meaning centering for ease
of use and interpretation (Raudenbush & Bryk, 2002b). This means that we

would interpret our results as deviations from the grand means of all FBUs.

HYPOTHESIS TESTING

We test the specific relationships utilizing the general models above and
using Maximum Likelihood Estimation. Our investigations require access to
SPSS 15.0 and HLM.6.0 to perform the statistical analyses. Given the speciﬁc
research questions asked and degree of freedom constraints, at a minimum,
three separate models need to be constructed. Furthermore, individual
parameter estimates can be adjusted with Bayesian shrinkage methods.

The key relationships are tested with the control variables added
incrementally. The Level-1 model is built first with ML deviance statistics
examined to see if model improvement occurs with the introduction of the latest
variable. Since we have greater than 30 observations at the higher levels, ML
estimates should produce similar results to REML. In addition, we provide a
diagnosis of the data visa vie HLM assumptions about errors and random effects.

For instance, Boxplots, Scatterplots and 0-0 Plots at each level should be

60

examined to determine whether there are any transformations required or any
abnormalities in the data structure (Luke, 2004).

For the ﬁrst hypothesis, we examine the aggregated diversiﬁcation-
performance relationship for F BU. in corporation; or country,-. F BU consists of all
entities operating in a speciﬁc country for a particular corporation. We test a
squared diversiﬁcation term to determine if there is a curvilinear (inverted U)
relationship. If there is a positive main term and a negative squared term, this
indicates the presence of curvilinearity. FBU level covariates include the number
of governance modes and total number of outlets. While we have no theoretical
justiﬁcation for predicting the direction of the governance covariate, we expect to
see a negative relationship between total number of modes and market share
due to the increasing difﬁculty in trying to control different modes. We also expect
to see a positive relationship between network size (i.e., total number of outlets)
and market share. The corporate level control variables are experience, size, and
centralization. We would expect to see a positive relationship between
experience, size and centralization on Level-1 means and relationships. All
three variables will be tested to see whether they explain the variance in the
intercept and diversification parameters. At the country level, industry growth and
concentration are the covariates expected to affect the intercept and
diversiﬁcation terms.

For the second hypothesis, we examine each FBU format in country
and/or corporate contexts. We are interested in whether particular control modes

are associated with higher performance. We also need to control for format

61

(type) differences, dominant business within the country, and whether the
particular entity is related to the dominant business. The corporate level control
variables are experience, size, and centralization. We would expect to see a
positive relationship between experience, size and centralization on Level-1
means and relationships. All three variables are tested to see whether they
explain the variance in the F BU Level parameters. At the country level, industry
growth, format size, and industry concentration are the covariates expected to
affect the F BU level terms. We expect to see a positive relationship between
experience, size and centralization on means and slopes.

For the third hypothesis, we examine six diversification strategy-
performance relationships. FBUs consist of all formats operating in a specific
country for a particular corporation. We also create a proﬁle for the FBU by
looking at the dominant governance mode along with the diversiﬁcation strategy.
The corporate level control variables are experience, size, and centralization. All
three variables are tested to see whether they explain the variance in the
intercept and diversiﬁcation parameters. At the country level, industry growth and
concentration are the covariates expected to affect the intercept and
diversification terms.

For each hypothesis we have estimations of the FBU, corporation or
country effects. The signiﬁcant value in partitioning the variances is to show the
relative contribution in explaining variance in the dependent variable for each
level of the model. Graphical presentations of the results are used to illustrate

the relationships.

62

CHAPTER IV
RESULTS AND DISCUSSION

In this section, we report our findings from the statistical analyses and
discuss our finding in the context of our theoretical model. Our results from the
statistical analyses are described below. The ﬁrst section will report descriptive
statistics and correlations. The second section will be broken down by the

analyses related to each hypothesis. The last section summarizes our results.

THE SAMPLE

In Appendix A, we show the descriptive statistics, histograms, and
bivariate correlations for our sample. We have collected data on 351 format
observations belonging to197 FBUs. In turn, these F BUs belong to 34
corporations and operate in the 49 countries in Tables 2 and 3. Because we were
using Euromonitor data, we did not capture foreign operations of these .
corporations that operated outside of the 52 countries covered. However, we
believe that the countries studied represent the most likely foreign market

expansion targets of international retailers.

63

Table 2: Corporations in the Sample

 

 

Corporation

Aeon

Ahold

Aldi Einkauf (Aldi)
Auchan

Axfood

Carrefour

Casino

Colruyt

Cora (Louis Delhaize)
Costco

Daiei

Dairy Farm International
Delhaize Le Lion
Foodland

Globus Handelshof
Grupo Gigante
H.E. Butt
Ito-Yokado

J Sainsbury

Lidl & Schwarz
Metcash

Metro

Modelo Continente
Pick N Pay

Rew'e

Safeway

Seiyu

Shoprite

SHV Netherlands
Tengelmann
Tesco

Uny

WaI-Mart

Whole Foods

Country of Origin
Japan

The Netherlands
Germany
France
Sweden
France
France
Belgium
France

USA

Japan
Hong Kong
Belgium
Australia
Germany
Mexico
USA

Japan

UK
Germany
South Africa
Germany
Portugal
South Africa
Germany
USA

Japan
South Africa
The Netherlands
Germany
UK

Japan

USA

USA

 

 

Table 3: Countries in the Sample

 

 

 

Countries

rgentina Mexico
Australia Morocco
Austria Netherlands
Belgium New Zealand
Brazil Norway
Bulgaria Philippines
Canada Poland
Chile Portugal
China Romania
Colombia Russia
Czech Republic Singapore
Denmark Slovakia
Egypt South Africa
Finland South Korea
France Spain
Germany Sweden
Greece Switzerland
Hong Kong Taiwan
Hungary Thailand
India Turkey
Indonesia UK
Ireland Ukraine
Italy US
Japan Venezuela
Malaysia Vietnam

 

 

We checked the distributional assumptions for both the aggregated-by-
country-and-company set and the disaggregated business segment data sets.
Appendix B contains ANCOVAs with Random Effects and relevant information
for each hypothesis. We broke down the diversification total into its two
component parts of related and unrelated diversiﬁcation.

There are two main data transformations required before analysis. The

first transformation is the logarithm and square-root transformations for the

65

independent and dependent variables that exhibit signs of severe non-normality.
The second issue relates to collinearity problems caused by including variations
of the same measure in the model. Because we hypothesize a curvilinear
relationship and have chosen a polynomial model to test the relationship, we
need to create a second order diversification term that is not severely collinear
with the first order term. We employ a mean differencing technique where we

replace our variable with a difference score. For example, if we have

Y = ,80 + ,6,x + ,62x2 + a and a sample size of n, we would let )7 = n“1 2, x, and
deﬁne z = x, —)7. When we substitute the 2’s for the x’s, the second order terms

become orthogonal to the ﬁrst order terms. As a result, we still have a moderate

and signiﬁcant correlation between the variables ( p = .765 ), as opposed to

nearly perfect correlations. But the consequences of not making these changes
can include inflated standard errors and unstable parameter estimates
(Wooldridge, 2003).

The next section examines the results of our analyses. Each hypothesis is
introduced by a summary of results followed by an interpretation of the findings.

The section ends with a discussion of our results.

Hypothesis 1

Hypothesis1: At lower levels of diversiﬁcation, FBU performance should increase.
At middle to high levels of diversiﬁcation, FBU performance will peak and then
begin to decline.

The ﬁrst hypothesis concerns the overall impact of diversiﬁcation strategy

choices on international retailers’ foreign market performance. In electing to

66

diversify operations, F BUs are leveraging slack resources and capabilities and/or
exploring opportunities to learn about new forms of retailing”. When retailers use
their existing resources more efficiently and/or innovate to meet customer
demand more effectively, we expect them to gain higher market shares. While a
certain level of diversification is presumed to boost FBU performance, we expect
to see a tapering off of performance beneﬁts from pursing diverse retail business
opportunities. Since management capabilities are not limitless, retailers may hit a
performance ceiling from juggling too many dissimilar operations and resources
get stretched too thin. We explore this notion in the analyses below, controlling
for potential corporate and country influences that will influence our FBU
parameters. In short, we ﬁnd no evidence of a curvilinear relationship between

diversiﬁcation and market share in our models. Hypothesis 1 is not supported.

Corporate Effects

In our baseline model, the first analysis answers a simple question: For
the sample of 197 FBUs, how much do FBUs vary in their mean market share?
We ran a One-Way ANOVA with Random Effects to estimate the mean for each

corporation, ﬁnd the average of these averages, and the amount of variation
between FBUs. The grand-mean corporate retail market share, 2700, is from .33%
to 1.03%. In other words, without looking at individual corporations, we can say

that the average FBU market share performance is less than 1% of the total retail

market.

 

‘2 An alternative explanation could be that retailers are having difficulties competing in their home markets.
exempliﬁed by Japanese retailers expansion into Hong Kong (Stemquist, 1997b).

67

In addition, we find that there is signiﬁcant variation of corporate means,

ﬂoj, around the grand mean, 700 (12 = 80.09929,df = 32, p < .001). The estimated

variability in these corporate means is .46457 with corporate total market share
means ranging between 0% and 2.02%. The intraclass correlation indicates that
about 18% of the variance in market share is between corporations.

Models 1 and 2 attempt to clarify the Level 1 (FBU) model. In Model 1, our
aim is to test the curvilinearity hypothesis, using a polynomial model. Both
diversiﬁcation variables are difference scores to remedy multicollinearity
problems between the two variables. The first is the mean difference of total
diversiﬁcation for the FBU, while the second is the mean difference score
squared. We ﬁnd that the directions of the variables change, as would be
expected in an inverse U-shaped model. However, the squared term is not
signiﬁcant. When we examine the correlations between the two terms, we see
that they are moderately high, but below .7. When we check for potential
curvilinear patterns in the 0-0 plot of the Level 1 residuals of the final model, the
majority of the values lie on or very close to a straight line. In sum, we reject
Hypothesis 1 that there is a curvilinear relationship between diversiﬁcation and
performance. The model results are in Table 4, while explanations are contained in

Appendix C.

68

.CGOE

05 E0: 020300 m_ 0_nm_:m> 02: a move «m EmocEQm m_ .009: 3030:: 2 80:19:00 .10220 .oﬁva Sm .505:ng .mo.v0 an 2:005:99

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m w h m m m v m m:$oEa:mm %
u nowhomo ovmwm. 3m oNromdmo oooovdmw omwmvﬁow movmdmo o {modhm av ENE. 00cm_>m0
mmmN. ovov. ammme 3 3. n? B. mmmv. o 30v. oVov. cozmt m> _m>m.._ 290900
v0.3"... 0mg”... mwomé SKY? woowé Smog. Smog. Now—2N Nb #0th _. _0>m..
35:0 :60 E005:
“mam? ~33: so“ .3250 .30...
a Bow. m gov. m mwvm. m mmwm. m vvvm. 2% .3250 :35.
2Q .3250 :8 535:2
umwmvr 8% 520059020
20:3 :85:
swim
020:3 8:85.095
m on 5. m 56. m E 5. :Am 60:28:06 =Eom
a Vehvé m non F._. m mmvmé a 2.0m.— » mowm... Emnw VN omowomé 2% €me 203859020
:0. £88552: :28
a «men. 8% .0~_w 05:00:00
5 . 00 :
m 88. m 88. m 53. ”$3. a: 8:: Em 95
5:022:25
n amend- mohmmdr m mwom. F. m mmwm. T m N35. 7 nhwmmr nvmwmwr mmmwmr Sum .o:m:m 3wa cams.
an .220 aims.
:20”. 20:00:00 0mm:m><
30th 63:“—
an.
0:0.W+ $.SO+ 03+ 9... $250+ 2.0+ ._.n+ .2005. swam

 

 

 

$0.8m. 30:01.60 5.! 3:30:20”. oo:0E:oto1-:o_ueo=_m:o>_o ”v 030...

69

Final Model
= 700 + 701*(INTDIV021. - INTDIVOZ.) + 702*(LNSIZEJ. - LNSlZE_) + Y10*D'l'02ij +
*(EXPCONTI. - EXPCONT. )*DT02ij + 720*LN'ILOUTOU. +

*(LNOUTO2I. - LNOUTO2_)*LN11.OUT0U + qu + rij

LNM KTD3U.

711
721

In the models above, we looked at the effect of the covariates on the
relationships of FBU variables (i.e., total diversification, FBU size) on market
share. Next we turn to the impact of covariates on the intercept. In model, the
intercepts represent mean corporate market shares. The final model adds the
corporation’s size, namely the total number of employees worldwide, as an
additional explanatory variable for the intercept. In this model, we assume that
corporate size has the same effect across corporations. Does corporate size

signiﬁcantly predict the mean market share? First, examining the intercept, we

see a significant relationship(;‘/‘00 = —2.3659,t = —8.405). We also observe a
positive association between international diversiﬁcation and mean market share
(7‘10 = .063623,t = 2.942). Corporate size also has a positive association with
mean market share (fqo = .375175,t = 2.524), controlling for other FBU and

corporate factors. The residual variance between the Base Model (100 = .4647)

and this model (2'00 = .25587) is lower. In other words, we explained 45% of the

true between-corporation variance in market share by adding four corporate
variables. Controlling for international diversification status, the range of

plausible values for mean market share is 0% to 1.43%. Even after controlling

7O

for the effects of international diversiﬁcation, corporations still vary signiﬁcantly in
their mean market shares (12 = 79.69298,df = 30, p < .001), meaning that other

variables can help explain variation between corporations.

The EB estimates of corporate market share means indicate that the
mean values are fairly numerically close, but that there are noticeable differences
within the corporate ranges. The only remaining issue is the slightly different way
of treating corporations with only one or two F BUs as compared to corporations
with many more FBUS. The precision of the estimates in predicting a mean with
a small number of FBUs is probably lower than the precision of the estimates for
corporations with a large number of FBUs. In this sample, the international
diversiﬁers tend to have higher market shares than most other corporations.

First, we look at the influence of international diversification on market
share means. This shows us that seven ﬁrms with the higher levels of
international diversification (e.g. Ahold, Auchan, Carrefour, Casino, lto-Yokado,
Lidl & Schwarz, and Metro) tend to have higher F BU means, on average, than
many other ﬁrms in the sample. In Figure 2, after controlling for FBU and
corporate size and within-country diversification, Carrefour has the highest mean

market share of 2.62% and Colruyt has the lowest mean market share of .008%.

71

 

>Imumm .88“. magnum
gun. z in"; ...mmsmzuom
.Em_uoon.um~ .saaoumm
.tm2-_m>>uR .>z:uo~
.oomohumm .ccmE.omco...uvm
.:>_mwnmw .>m3m.mmnmw
63mm" a éiaocwuom
.2082"? .262"? .63":
532.8..qu .8283..an
..Sm min: .9395 820"?
.mBoauﬁ 6528...:
:0qu .8685

.SoOum .9...me "x. 52950
"0 .UOOEHO .CNLOD< "V
.623. .u_o;<u~ .823

 

 

 

:9:
528.2
26..

mm on mm .0N VN

 

 

 

cougar—.295 .30sz3... 5 .90..

mm 8..

p .‘I p

r

 

 

 

3.!qu tome“ ueew

 

chasm «Stu—2 380900 cues. :o cow—3599:... .ncozquE.

«5380950
F S 3 .2 2 m o v N .
. . . . . _ _ moo
* ﬂ . 2o.
, .
I _ . .7 -
_ .. - woo.
, _ . ..
.
I - tum.
- 0%.
"N 0.52“.

72

This is also the case with the impact of the corporations with the largest
number of employees. Ahold, Auchan, Metro, Rewe, Safeway, Tengelmann,
Tesco and Wal-Mart exhibit higher mean market shares. Some of the conﬁdence
intervals between the smallest and largest corporations do not overlap. In Figure
3, after controlling for FBU size, international and within-country. diversification,
Carrefour still has the highest mean market share of 2.62%, while Whole Foods

has the lowest mean market share of .012%.

73

 

>Imumm .88“. oozing sun. 2
can; .ﬁmoazuom .ocm_uoon_umu
.szooumm .tm2-_m>>un~ Wznumm 22.28.30

.oowohumw .ccmE_o :mhuvw
.:>.omumm .>m>>£mwumm 3” mm on f8 arm W mm 8L3 $1.3 amp . ow , .m
.msomua .aiaocmuom
.208qu .osmzuﬂ .63":
5338?? .8292"? .
.25 min: .9590 826"? . ,
.

 

.mano.0uw_. 659.803 _. .10qu
.088on .Eonuum .oEmmo uh
Soootmo no .nootaum .cmco:< uv . .
._o_<um .u_oc<u~ 502$ , .

 

 

 

 

505..
$2.85. II
=me I 1

35 0.90900

 

.1 moo.

.1 2o.

woo.

I tum.

 

- coo.

eJeqs tame" ueew

 

 

9.25
~9th 222.30 cues. :o 35 382.30 ”n 0.59".

74

Does corporate retail experience signiﬁcantly predict diversification
slopes? We might expect more experienced corporations to be better at
formulating and implementing diversification strategies in their local operations.
This expertise would be expressed in positive, steeper slopes, reflecting better
results from each step towards greater diversity in their local portfolios relative to

less experienced corporations. First, we see a signiﬁcant, positive relationship
between diversification and performance(j?10 =1.245926,t= 4.385). Next we
observe that the effect from operating in the retail industry magniﬁes the
relationship between diversification and market share (in = .017352,t = 2.798).

For example, if Wal-Mart has an average diversiﬁcation score of 1.25, a one unit
change in diversiﬁcation will result in a market share increase of 3.49%. When
each additional year of experience has an impact of a .02 on the diversiﬁcation, a

one unit change in diversification results in a 3.55% increase in market share.

75

Figure 4: Corporate Experience on Diversiﬁcation-Market Share

 

 

 

- _- Relation—”imp -- - _
Level of Experience
1093‘
. " Low
-. Medium
,- High
Q 4.71 ~
S
«n
5
f
G
2
1.08
.98
.45w‘1'. T ...e...,....,
0 0.47 0.94 1.40 1.87
Diversiﬁcation

 

 

Grouping our retailers into three experience groups (low, medium, and
high), the Figure 4 above reveals an interesting difference in the relationship
between FBU diversiﬁcation strategy and market share. Essentially, the oldest
companies are able to translate their lessons gained through the years into
signiﬁcantly more robust diversiﬁcation strategies, which are associated with
greater market shares. For example, looking at the range in market shares for
Ahold, the company attains its lowest mean market share of .73% for

undiversiﬁed operations and 16.6% for its most diversiﬁed operations in Sweden,

76

holding other FBU and corporate variables constant. In contrast, Cora has a
mean market share of about .7%, holding other F BU and corporate variables
constant. This evidence supports the notion that experiential Ieaming developed
through years of retail operations creates a knowledge base that is vertically
transmitted to FBUs. This knowledge aids in FBU diversiﬁcation strategy
formulation and implementation, resulting in higher performance. The gap
between the two lesser-experienced groups and the highest experience group
suggests that there is a speciﬁc amount or range of time that must pass before
the experience coalesces into a capability.

The slopes and magnitudes are more similar between the mid— and low-
experience groups than the oldest group of companies. Why? When we go back
to the experience variable histograms, we see a drop-off in companies, which
could suggest a survival threshold or a sampling artifact. In other words, only
those chains that absorb and learn from their operational lessons survive past
the 50 year mark of operations. For example, we see a break between the
establishment of retail operations between Aldi (1948) and Lidl & Schwarz
(1930), roughly in line with the years between the Great Depression and the end
of World War II (WIMI). Supporting Delios & Beamish (2001), continued
existence after WI and exposure to extreme changes in retail conditions during
that time period prepares managers to formulate sound diversiﬁcation strategies
and equips companies to compete more effectively in local markets. Perhaps
more importantly, with more market selection experience, older ﬁrms are able to

recognize more favorable institutional environments (Henisz & Delios, 2001).

77

In this model, we assume that total outlet size has the same effect across
corporations. Does corporate outlet size signiﬁcantly affect the relationship
between chain size and market share? First, examining the mean outlet size, we
see a signiﬁcant relationship(;720 = .407786,t = 5.364). Next, in Figure 5, we turn
to the effect that operating in the retail industry magniﬁes the relationship
between the FBU's relationship between corporate outlets and market share

(f2, = —.121471,t= -3.454), controlling for other FBU and corporate factors. In

effect, the positive relationship between FBU chain size and market share

becomes weaker in larger corporations.

__ figure _5:_ Corporate Size_ on FBU Size-Market Share _Relationsl'l_ip_‘

 

 

 

 

 

 

No. of Corporate Outlets
15.029 _- Small ‘
‘ Medium
Large

4.572-
2 1
N
.c .
w o
a I

1.377 -
'E .
N
E I

.419 i

.127 J--- l t r’ j—‘T 7 f i r r f v r‘r rrrrr

2.15 4.29 6.44 8.58
Log ln-Country Outlets

78

A possible explanation is that large companies over-emphasize network
size as an indicator of efﬁciency and success. This would create a situation
where FBU managers would be encouraged to expand the number of outlets at
the expense of a more cautious growth strategy. Makro and Costco, both
warehouse clubs, have the steepest slopes, whereas a diverse group of
companies such as Rewe, Aldi, lto-Yokado, and Tengelmann have the ﬂattest
slopes. This suggests that, at least at the high end, format type may be
inﬂuencing the results. In other words, larger format retailers would naturally
have fewer outlets than smaller ones, such as convenience stores. Because the
two warehouse clubs have the steepest outlet-share slopes, it is likely that that a
simple count of outlets without controlling for square footage or retail format or
speciﬁc retailers may be distorting our ﬁndings.

In testing the time-invariance of the ﬁnal model, we re-ran the analysis
with market share for the same companies in 2004. In these tests, we continue to
use independent variables from 2002, so the results give some indication of the
lasting effects of the F BU strategies and characteristics as well as industry
characteristics on future FBU performance. This reduced the number of useable
observations to 184. Since the magnitudes and directionality of the parameter
coefﬁcients are similar, and the parameters are signiﬁcant, suggesting that F BU

portfolio strategies and industry factors do not change outcomes in the near term.

79

Country Effects
We re-tested the ﬁrst hypothesis about a curvilinear relationship between
diversiﬁcation and market share. We ran a One-Way ANOVA with Random

Effects to estimate the amount of variation between FBUs and countries. The

grand-mean retail market share, 7700: is from 52% to 1.17%, which indicates a

range in average market share levels among countries. In addition, we find that

there is signiﬁcant variation of country means, ﬂoj, around the grand mean, foo

([2 =140.50599,df = 48,p < .001). The estimated variability in these means is

.75471 with retail market share means ranging between 0% and 2.55%. The
intraclass correlation indicates that about 31% of the variance in market share is
between countries.

Models 1 through 4 clarify the Level 1 (FBU) model. In Model 1, our aim is
to re-test the curvilinearity hypothesis, using a polynomial model. We ﬁnd that the
directions of the variables change, as predicted. However, the squared term is
not signiﬁcant. As with the corporate model, we reject Hypothesis 1, indicating
that we ﬁnd no evidence of a curvilinear relationship between diversiﬁcation and

market share.

80

2 comthEoo .25.). u .m

3:85.350 .8 game. 5cm o.am...m> .5 32 2 £2.26 . oﬁva 8 E852? m. .ouoE £5.55

ova 6 E8...

:58 m. .ouoE £5.55 2 comtwano .26.). o .o ﬁva aw E3536 n .mo.va 5 EmoEcQw .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m N. o n v m m 90.0Em5n. %
u Noooﬁoo omohoé 5 o vmmwﬁmo ooohoﬁmo om wmvdoo won fmoo moNooon $95300
comm. Nova. vhoo. oven. mnom. ovnm. Nwmn. E0935
N _m>m._
.vwvoé noové 2.2.... Foo—2r romné voomé vmooé N.0 .:o=m_._m> Dmu. 555)
F _o>oi_
8:25 :50 Eon—Em.
.33.- cm.“ 5.55. mEQoEm
. E... .88. .88. .82... am .9250 :82
3%. £0.50 ..o 52:52
Com: 8% 5385295 23cm :85.
3%. .uoasow 8.89.2995
. 2F: . EB. . m 5m. . moooé . moms; .vnwmé 2% .53.). 8285995
.N. n

6. . 82853520 .98.
a tow... womomr 8% .wuw
. 55.. .23. saw 5. 8% Joins. 9.995.
a NNoNN- «mom VN- a omvo. w- a moon. w- aoNoo.- 5.. Fr chow: 8%
3%. .Emcm aims. :23. 3803..
muuutm cox?—

0 55+ 35+ 5 ..oEm... «250+ .5+ ~...n.+ .25.: «man

 

 

 

 

 

 

 

 

muootm >5:on 5.3 QEmcoaﬂom oocmELotoméozeoEm.

 

2,5 H... 2.3

81

Flﬂl' MOdOl
LNMKTD3‘) 3 YOU + TmtEMERGEj + 102¢(LNEMPLOYI ' LN:MPLOY) +
Tza'LNTLOUTDij 4‘ yzftEMERGEjnLNTLOUTO” + U0}. + To.

Yto'szij +

An inspection of the EBintercepts ﬁnds a fairly wide range in conﬁdence
intervals for market share means between countries. For clarity, we identify
countries by their membership in the smallest, average and largest size retail
markets (i.e., retail labor market size). In Figure 6, some of the larger markets
tend to have lower mean value ranges than the other groups. India, ltaly. Japan,
Norway, and the United States have low mean FBU values. It may mean that
foreign retailers have a particularly hard time winning market share for a variety
of reasons, including barriers to entry (i.e., India) or strong domestic food
retailers (i.e., Italy, US). When foreign companies do not have a strategic
competitive advantage relative to domestic competitors, intemalizetion theory
suggests that they re-consider entering the market (Dunning. 1980; Salmon 8-
Tordjman, 1989). In emerging markets, when barriers to entry and competition
are eased, we would anticipate market share means to rise, as international
retailers begin to enter the country. As for larger, more sephisticated retail
markets, there are no theoretical reasons to expect mean market shares to rise.

all things equal.

82

 

EmSwSumv .m.o=~oco>nmv
.<wDu~v .mEmbSuov .vSnmv
Swish": .Ucm__mc._.nm¢ £928.qu
.2855" 3. .885ro

.595an £23. .muwm .m_xm>o_wu\.m mm Q? 9. WV Ev mm mm 3 mm on an wN «N NN cm 3 m: 4; NF m: m w v N o
i i it _ i i ii»! _ . . _ _ _ tr _ A

r

.QoammEmnwm .m_wm:mumm
.mEmEomnX” ..mmatoaumm .ucmBanmm
.mocaaacmnFm Satcozuom .ocmEoN
soZHmN .wucmtoceoznww .oooocozunw
.8_xe_2umm .m_m>m_ms_nmm .ceamwuvm
émgumw .ocm_2_uwm .5389:qu
.mﬁEuom imaging .96! accrue
.8020": Sceczemuue .oocﬂuumF .
.Emacui seam"? .emscoouua
.dom Loewe": .mEEQooue

.955 um .mEOuw .mumcmoun e .e ,
.m_._mm_=muo .__Nm._mnm .EEQomuv . i
.mEm:<um .m__gm=<um .mcucomEuF 4 _

 

 

 

 

@084
8:622
__mEm

sees. .83 Emma co mum

 

merge—.00

 

 

 

r mood

9.95 «9.35. sees. :0 00:03:... «9:22 .393 :20”. "a 0.59".

voo.

, a . mam.

aJeus tamew ueaw

83

Finally, in Figure 7, we look at the EB confidence intervals for the
emerging market-mean market share relationship. We can see that foreign
retailers are able to capture, on average, higher markets shares than in
developed retail markets. Supporting Dunning (1988), foreign retailers enter
markets in which they possess an advantage over domestic competitors. Given
the greater sophistication in and exposure to a greater number of merchandising
management techniques in more advanced retail markets, these advantages
allow foreign retailers to garner a larger piece of the local foreign markets. Also,
the absolute sizes of these markets tend to be smaller than developed markets.
Not surprisingly, India and China have lower means than the other emerging
markets. The toughest developed market for foreign retailers is the United
States. Again, not surprising, given the number of foreign retail failures in the

highly competitive US market.

 

EmEoSumv .m.o=~oco>nwv
{manta .ocﬁiauov £3an 35560
$9.55.": .2355"? .cmﬁmhumv
.ucmtoﬁesmu—v .cooozwnov on we we 3 8 9. mm on X” 8 on em on em 8 .8 m: S 3 we .2 m e ..s _m .o
.samumm 698. .muwm degoauam . . . . . _ . . . . t _ _ . . _ still--- - . we,
.ocoawmgmuom .mﬁménmm -
.mEmEomuvm ..mmstodumm desoauwm “a ,.
.wmc_aa___ndu5 .>m2co2nom .ucmﬁmN ..e .s _
BoZHmN .mccmtmEmanN .oooocoﬁunw W l ..
a

#00.

oi

 

.ooioEumm .m_m>m_ms_nm~ .cmamwuvm . n , . e .8.
SEER .eceewmm .msocousuﬁ 4. - . , V .
.mBEHON .bmmcaIumF .mcox mcozumr R .4
.8020": $5850"? .oocmiumv . -
.225“.qu .aamum. .xascmouﬁ
.81 :83": @3528qu . - a,

@550 um .o__;0um .momcmOnn _. .
.wtmmsmuo .__N8mnm €20.33. 8
.m_=m:<nm .m__mbm:<u~ .mczcmmequ r n _

 

 

 

. - .1. .- «2.

 

 

 

 

. ‘ _. ..n - mwm.

. iii—“r

@59me . i .
coao_m>mo I- ,

 

i moo.m

Eoan_e>mo o_Eocoom_ co _o>o._

 

22m “execs. cows. :0 Sane... EeEeo_e>oo ho _o>o._ K 2:9".

aJeus temew ueew
85

Figure 8 illustrates the impact of emerging market status in the country as
an explanatory variable for the size of their local operations. In this model, we
assume that being an emerging market has the same moderating effect across
countries in that, on average, emerging markets are likely to have less developed
chains. Does a country’s level of development significantly predict the
relationship between chain size and market share? Essentially, we want to know
whether a retailer’s capital expense for expanding a store network yields similar

market share results across the world. First, examining the outlet mean, we see

a signiﬁcant positive relationship(;‘/‘20 = .4711,t = 6.787). In Figure 8, we turn to

the effect from operating in an emerging retail industry weakens the positive

association between chain size and market share (A = -.2818,t = —2.829),

controlling for other country factors. For further evidence, we examine a graph

representing all the countries in the sample.

59mg: Len/2L0! DeveEPmenLgnFEQQutlet - Market Share Relationship,

 

Level of Economic Development

 

 

 

5'87 _ Developed
,. -‘ _, ~ Emerging
1.53 . , {'1
l
‘ /
.40- .-
a , .1
.C
m
‘0’ we
E .
2 .11 ‘
.3 . ra~vw~wwvw~r. .. .. , ,, ,,
0 2.15 4.29 6.44 8.58

Log ln-Country Outlets

 

 

In every case, the slopes of the emerging markets relationships between
chain size and market share is suppressed relative to the chain size-market
share slopes of the developed countries. FBUs operating in emerging markets do
not get the same boost in market share from possessing large chains as in
developed retail markets. A possible explanation for this is in the retailing and
logistics literatures.

The efﬁciency of the technological environment is critical to success of the
retail sector (Stemquist, 2007). The technological environment relates to a
country’s information, communications, energy, and transportation
infrastructures. Of particular importance for retailers is the efﬁciency of the

logistics infrastructure because it directly impacts their operational costs. At the

87

 

firm level, logistics costs represent the costs associated with managing the ﬂow
of information, products and services within the supply chain. For retailers,
superior logistics capability is considered a sustainable competitive advantage
(Levy & Weitz, 2006).

At the national level, logistics costs represent the systemic ease at which
products, services and information flow within and between industries. Logistics
scholars have been developing estimation methods that will predict the market
potential of logistics services. Using different data inputs, early works
concentrate on estimating U.S. logistics costs as a function of transportation,
inventory carrying costs and administrative costs (Delaney & Wilson, 2002;
Heskett, Glaskowsky, & lvie, 1973). Expanding the scope of inquiry from the
United States to global logistics issues, Bowersox (1992) estimated logistics
costs as a function of GDP, government sector product, industrial sector product,
and total trade ratio. Extending this research, a number of studies using neural
networks estimation to calculate national logistics costs have come from
Bowersox and his colleagues (Bowersox & Calantone, 1998; Rodrigues,
Bowersox, & Calantone, 2005). The authors use domestic and global trade data,
transportation infrastructure, and controls for country characteristics and location
to estimate national logistics expenditures (Rodrigues et al., 2005). In support of
Rodriques and colleagues (2005), we find that logistics expenditures are
significantly different between countries and that emerging markets have much

higher logistics costs than developed markets.

88

The residual variance is reduced between the base model (200 = .75471)

and the ﬁnal model decreases(r00 = .3866). By adding all three country level

variables, we explained 48.6% of the true between-country variance in market
share, or slightly less variance than in Model 5. Controlling for retail market size
and level of development, the range of plausible values for mean market share is

0% to 1.29%. Even after controlling for the Level 2 effects, countries still vary
signiﬁcantly in their mean market share ([2 2124.91936,df = 46, p < .001).

In testing the time-invariance of the final model, we re-ran the analysis
with market share for the same companies in 2004. This reduced the number of
useable observations to 184. The magnitudes and directionality of the parameter

coefﬁcients are similar, and the parameters are signiﬁcant.

Discussion of Hypothesis 1 Findings

To summarize, we have rejected the curvilinearity hypothesis in both the
corporate and country models. Theory suggests inherent cognitive limitations of
managers with respect to their absorption and processing of information for
optimal expansion decision-making (Nayyar, 1993a). Positive diminishing
marginal returns have been associated with ﬁrms that stray too far away from
their core competencies (Markides, 1992). Our ﬁndings do not indicate
limitations to managers’ ability to successfully traverse expansion decisions. A
possible reason is that much of the previous research is based on manufacturing
samples, as opposed to service firms. For example, Palich et al. (2000) state that

focused ﬁrms cannot benefit from economies of scope. Yet for retailers, the

89

source of scope economies primarily lay in distributing large quantities of goods
through common warehousing and/or distribution facilities (Chandler, 1990).
From this perspective, even undiversified retailers, such as Lidl or CostCo,
possess economies of scope.

Yet diversiﬁcation does have its beneﬁts. Our ﬁndings suggest that an
average retailer can double its market share with one unit change in level of total
diversification. Moreover, related diversification seems to yield the highest
performance. Diversification within the most successful segment tends to
outperform between segment diversification. Undiversified ﬁrms can also survive
the jump into unknown formats. Lower coordination and cooperation costs
associated with unrelated diversiﬁers help them achieve proﬁtability (Leontiades,
1986; Varadarajan & Ramanujam, 1987). FBUs may also be willing to tolerate
lower market shares with the hopes of creating learning opportunities to support
innovation with the potential to lead to sustainable competitive advantages
(Barney, 1991; March, 1991; Nelson & Winter, 1982).

In exploring how diversification helps retailers perform, we need to look at
the directional patterns (i.e., within segment vs. between segment diversification).
In Appendix B, we provide models that decompose the diversiﬁcation measures
into related and unrelated diversification. Some researchers argue that MNCs do
not experience performance differences from different diversification strategies
(e.g., Palich et al., 2000). Our findings show that FBU outcomes vary
signiﬁcantly between related and unrelated diversification strategies. The

difference between related and unrelated diversification can be viewed as “the

90

relative distance between the knowledge need to operate in the new domain and
the degree of knowledge available in the current domain” (Kazanjian & Drazin,
1987, p. 347). Supporting the relatedness explanation, we ﬁnd that intra-segment
diversiﬁcation strategies yield relatively higher performance than inter-segment
diversification strategies. Cross-exploitation of resources within F BUs allows
ﬁrms to more efficiently utilize FBU resources. Moreover, firms benefit by
generating and sharing operational synergies via expansion into markets that
serve similar customers and/or utilize similar business models. Because this
accumulating organizational knowledge is unobservable, competitors struggle to
imitate successful systems and strategies (Barney, 1991; Dierickx & Cool, 1989;
Wemerfelt, 1995). Furthermore, in multi-business firms, economies of scale,
scope and knowledge can be distributed within the network. In summary, F BUs
are able to effectively utilize their higher-order resources to produce synergies
leading to superior outcomes.

It turns out that unrelated diversiﬁcation is good for retail FBUs, too. In
foreign markets, retailers are exposed to different customer expectations and
industry conditions that may necessitate a shift into heretofore unknown
segments for survival. Retailers may also seize the opportunity to expose
themselves to new kinds of retail businesses without having to suffer intense
scrutiny of their home markets. The learning opportunity can either be bought
through acquisitions or joint ventures, or created from scratch in a new venture.
Our models suggest an overall positive benefit from unrelated diversiﬁcation in

foreign markets, albeit a less beneficial one compared to related diversification.

91

This lends some veracity to studies suggesting that corporations can achieve
reasonable performance and can even survive entrées into unknown territories
(Leontiades, 1986; Ramanujam & Varadarajan, 1989; Ramanujam, 1989).

In addition to providing support for the association between diversiﬁcation
and performance, we also explore and control for important contextual influences
on this relationship. First, we ﬁnd that F BU market share means are highest for
retailers with the greatest levels of international diversiﬁcation and retailing
experience. Geringer et al. (1989) provide evidence that diverse information
sources from higher levels of international diversiﬁcation aids knowledge creation
and organizational flexibility to boost MNC performance. Over time retailers also
create intangible organizational capabilities (Doherty, 1999). On average, higher
levels of international diversiﬁcation and retailing experience better equips
retailers for success in foreign markets.

In our corporate model, we also ﬁnd that a corporation’s level of retail
experience magniﬁes the F BU diversiﬁcation-market share relationship. Older
ﬁrms seem to have a much easier time than other ﬁrms in translating their
diversification activities into successful performance. Long-term exposure to and
lessons learned from changes in retail practices and evolving, diverse consumer
preferences appears to be filtering down to the foreign FBUs, even after
controlling for corporate size and international diversiﬁcation experience. This
suggests that a diversification competency is developed over time as managers

gain experience in making complex resource allocation decisions, supporting

92

Dundas & Richardson (1982). This knowledge also seems to have a lasting
impact on F BU performance.

Belonging to a larger corporation has its downsides, too. We find that
large companies dampen the FBU size-performance relationship. A possible
explanation is that larger ﬁrms over-emphasize outlet growth at the expense of
more controlled growth or focus on other kinds of performance indicators.

Retail labor market size and level of development inﬂuence FBU
operations. Large retail markets adversely impact FBU mean market shares.
There are three explanations for this result. First, in many of the larger labor
markets, specific industry restrictions were made upon foreign retailers for the
protection of local retailers. For instance, large—scale retailing laws speciﬁc target
and limit expansion of foreign companies with expertise in larger formats.
However, not all large markets have such restrictions. Also, domestic competitors
may simply be stronger than anticipated. When retailers do not possess
advantages over local retailers, the “liability of foreignness” hypothesis predicts
poor performance (Hymer, 1976; Zaheer, 1995). In either case, large retail labor
markets may be a proxy for difficult retail environments. Finally, in relatively large
markets, smaller market shares may translate into greater absolute sales than
larger market shares in small markets.

Level of development also affects FBU operations. FBUs in emerging
markets have higher mean market shares than FBUs in developed countries.
But, FBUs in emerging markets tend to achieve lower outcomes per outlet than in

developed markets. This ﬁnding indicates inefﬁciencies in chain retailing in

93

emerging markets. Because the technical environment is an important
precondition of retail success (Stemquist, 2007), Rodrigues et al.’s (2005)
ﬁndings that emerging markets have signiﬁcantly higher logistics costs gives

some support for our results.

Hypothesis 2
Hypothesis 2: In foreign markets, wholly-owned subsidiaries should outperform
joint ventures, which should outperform acquisitions.

Hypothesis 2 asks the question of whether one governance mode leads to
superior performance over another. The control choices taken by retailers are
driven by retailer assumptions about the foreign market environment,
assessment of behavioral uncertainty, and need to protect its secrets, and
government policy (Anderson & Gatignon, 1986; Stemquist, 2007; Williamson,
1985). We predict that wholly owned subsidiaries should perform the best and
acquisitions the worst. Joint ventures are presupposed to perform between these
modes. To clarify, we coded governance modes as the initial mode upon
entering the country. Since a large percentage of retailers grow using multiple
modes, this was the most straightforward measure of the retailers’ desire for level
of control over local operations.

In the next series of analyses, we use a disaggregated data set. While the
first hypothesis is tested using a summary of FBU activities within a country, the
second and third hypotheses require a disaggregation of the data at the format

level. Thus, each observation reﬂects the characteristics of a particular ﬁrm’s

format-level activities within a speciﬁc country. Since retailers often operate
multiple fascia within a country and format, store brands are summarized at the

format level.

Corporate Effects

After developing the format level model, we progressively incorporate
country level explanatory variables, namely centralization status and international
diversiﬁcation. All Level 2 (i.e., country or corporate) variables are grand-mean
centered to facilitate interpretation. In summary, we are trying to explain why
some corporations may have higher mean market shares than others and why
the magnitude of the relationship between certain format characteristics and

market share may strengthen or weaken.

Final Model

LNMKT03rj = Yoo+Yo1*CENTRAl-j+Y1O*RWOSij-+720*JVU-+121*(INTDIVO2J.-lNTD|V02_)¢JVij

*HYPER’]. + 740*SUPERMKTU. + 750*LNEXP020. + uo. + r”.

+730 ]

Centralized corporations may have a positive effect on mean market share

for acquisitions and UWOS. By adding an explanatory variable to the equation at

Level 2, we change the meaning of the residual, uoj and make the variance, 2'00,

a conditional variance of average corporate format market share after controlling

for centralization and international diversification. We observe a significant

negative association between centralization (701 = —.748509,t = —2.221) and

mean market share for acquisitions and UWOS. Using corporate centralization

95

as a proxy for information sharing, this implies that the level of knowledge
sharing is too low, cross-cultural communications are poor, and/or the knowledge
is not suitably useful for the speciﬁc operation to realize synergies in the average
acquisition or UWOS. Linkages between the FBUs and the parent company are
too weak to facilitate transfer of appropriate information to help the FBU
undertake successful activities to gain market share as opposed to being focused
on meeting the administrative and reporting needs of the MNC (Gupta &
Govindarajan, 1986). Another explanation is that since the centralized parent,
exerting a stronger inﬂuence than a decentralized parent (Scott & Meyer, 1983),
may not have strong capabilities or understanding of the technical or local nature
of the FBU’s marketplace, rendering their guidance unsatisfactory. Finally, even
though a corporation is centralized in its home market, its foreign operations may
be decentralized or regionally managed.

In Figure 9, we ﬁnd a slightly positive inﬂuence on the joint venture-rnarket

share relationship (£21 = .049568,t = 2.198) via the ﬁrrn’s international

diversiﬁcation experience. In other words, some more internationally sawy ﬁrms
may be able to manage joint ventures better than less internationally experienced
ﬁrms. However, we see a fairly wide range in slopes for the experienced
international operators. These differences may reﬂect many things. First, we
notice that the joint venture deals include both related and unrelated businesses.
Noticeably, we ﬁnd no instances of a UJV in a country when there is no RJV also

present. In other words, those UJVs would not exist if the RJVs were not part of a

larger joint venture arrangement. It is unknown (but probable) that the UJVs

represent a format in which the host country partner has some experience.

Figure 9: lntemational Diversification on JV-Market Share Relationship

Level of lntemational Diversiﬁcation

 

4.62 —-- Low
' ' Medium
‘ High
, .93 i , - l
i E T 1 7 H ‘T—ﬂ"
l g :8
l m - "‘_ 7.
Q .19 *
l 2
C
l 8
s .
.04
f I ﬁ‘ ‘
'01 0.00 1.00

 

Joint Venture

 

Second, there were almost seven times more JV businesses in emerging
markets than in developed markets. To a degree, this may reﬂect speciﬁc laws or
business requiring joint venture partners. For instance, in 2002, both Ahold and
Casino have joint ventures in Argentina, which places restrictions on larger scale
formats, so joint ventures may help them maneuver the local legal and regulatory
environment. Finally, the three companies that have slightly positive slopes are
Ahold, Carrefour and Metro. These corporations operate in more than 25
countries worldwide, representing some of the most proliﬁc international retailers

in the world. But they are not necessarily the ones that have the most joint

97

ventures. This lends some credence to the notion that ﬁrms operating in many
diverse environments may possess a speciﬁc international capability, which may
be more important than a specific mode management capability (Barkema et al.,
1 996).

The general tendency for non-joint ventures to have higher market shares
is evident. Firms with more diversiﬁcation appear to exhibit higher market shares
than less diversified ﬁrms. This group of more diversiﬁed ﬁrms also exhibits a
wider range of slopes than the less diversiﬁed ﬁrms with three corporations even
having slightly positive slopes for their joint ventures. This suggests that some
experienced corporations are better than others at sharing control of their

international operations with another company.

The residual variance decreases from Model 5 (r00 = .81083) to Model 6

(2'00 = .50324) to Model 7 (r00 = .48647). Controlling for centralization and

international diversification, the range of plausible values for mean market share
is 0% to 1.59%. In other words, we explained 40% of the true between-
corporation variance in market share by adding the centralization and
international diversification variables. Even after controlling for centralization and
international diversiﬁcation effects, corporations still vary signiﬁcantly in their
mean format market shares (12 = 88.52853, df = 32, p < .001) .

In testing the time-invariance of the ﬁnal model, we re-ran the analysis
with market share for the same companies in 2004. This reduced the number of
useable observations to 334. The reliability of the corporate means to predict the

grand-mean is near zero(2 = .008), though the magnitudes and directionality of

98

the parameter coefﬁcients are similar. In the ﬁnal model, international
diversiﬁcation is not signiﬁcant, suggesting that medium-tenn format performance
does not continue to be affected by the corporation’s international diversiﬁcation
strategy. Likewise, 2004 format performance appears to be unaffected by the
state of corporate centralization. We have to revert back to the full level 1 model
with random effects to get a model mean similar and signiﬁcant parameters. This
suggests that prior corporate effects’ inﬂuence on F BU performance fade over

time.

99

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

moww. Nmom. mo 5. 9. 5. mean. m 5m. meN. no 3. 30895
N _w>w4
anrN MNNNN awe-N vmomN mN _.v.N mottN FmNnN NONQN Nb .co:m_._m> 3mm 55:5
—. _0>w.._
2:23 Eoo Eczema
smomN. a EmN. a 58. Ram oozecoexm :85.
3%
eocotoaxm $550-5
anmm. swoon. momow. “No—b. 3% dams.
:6. .Hexcmgquw
meomw. w Nmmw. aVGNm. m ommn. m vwom. Rom Ewes.
2% .eexemEcegI
m omom. a NONm . u mmNm. mNVNe. a vmmm. m ONNV. 8% due.)—
as. .8688
8:20 2.23 855m
9 :6“ 522859020
momma? .95szan
m 2.8.- .83..- w :8.- .88: .88.- .85.- .32.- 2m 58.2 2285 E2.
:Q .2385 Eon
~33..- a 3mm; 3% SemanO 85.92.00
mwwﬁv. T m 93s...- mNNN-m. _.- momma. _.- moomoé- mwmmmr mmmmkr «BNemr 83.505.
3% €8.25
20th taxi
9+ 3350+ sanctum—3+ 535+ .25.? w0>>¢+ >_.+ Boos. swam
maeetm _e>e._ 280900 53> eucmEcoten. seems-580 unecon— 6 win...

100

101

mova um =52:ch m_ Econ. 3255 2 comtmano 682. .oﬁva Hm «:85:an .mo.va “m EmoEcQw.

 

 

 

o.- m w h 0 m e m £2meth %
wNth-Nv econoﬁdww ommNmeNF omommdmww o smwwéomv canned F9. owmwws F2. vmmvdmmp 096300

 

 

 

 

 

 

 

 

 

 

o 22¢ 3:8

Country Effects

In the next section, we are looking at potential impact of country level
variables to explain the relationship that governance mechanisms may have on
format level performance. In the previous corporate model, we were able to
identify signiﬁcant corporate covariates that impact format operations in foreign
markets. In this analysis, we look at the impact of F BUs starting business in a
particular format by buying retail chains outright or acquiring more than 50% of
an existing chain.

Unlike the previous model, RWOS and JV were not significant variables.
Therefore, we test the hypothesis from the presumed least effective governance
mode — i.e., acquisitions (ACQ). In our ﬁrst series of models, we are looking at
the corporate level effects on format level characteristics and strategies, and
performance (Table 6). Models 1 through 5 clarify the Level 1 (FBU) model. In

Model 1, We establish the significance of the governance parameter
(7‘10 =.297136,t 21.658).

In models 6-7, we progressively incorporate country level explanatory
variables, namely the size of the retail labor market and level of industry
concentration. All Level 2 variables are grand-mean centered to facilitate
interpretation. In summary, we are trying to explain why formats in some
countries may have higher mean market shares than others and why the
magnitude of the relationship between certain format characteristics/strategies

and market share may strengthen or weaken.

102

Final Model
U‘MKTO3U = 700 + 701*(LNEMPL0YI. - LNEMPLOY.) + 710*DOM020. + 720*RELT002U +
730*ACQMOREU + 740*LNEXPOZU 4' ysotLNOUTOZU. +

751*(SQRTCONCJ. - SQRTCONC.)*LNOUT02U + uoj + r0.

We predict that the size of the retail labor market may have an effect on a

format’s mean market share. By adding an explanatory variable to the equation
at Level 2, we change the meaning of the residual, um and make the variance,
2'00, a conditional variance of average corporate market share after controlling for

labor market size and industry concentration. We observe a signiﬁcant negative
association between labor market size (f0, = —.277542,t = —4.054) and mean

market share.

103

 

 

Emc§>umv .m_o:~oco>uwv

(wants. .ocﬁisuow .xaumv
$9.5": .uc..___.ﬁu$ SEE"?
.ocmtou_3mu 3.. .couoawuov
.c_mawumm 690x .mnmm .m_v_m>o_wn~.m
.anmmgwuwm £331.48
.mEmEomuvm ..mmatoaumm .ucsoaumm
.moc_aa___cn_u_.m SmEoZuom .u:m_moN
BoZumN .wucmEEmZHmN .oooouozunm
.oo_xo_2no~ .m_w>m_ms_umw .cmamauvm
.>_s_umm .ncgewmm .gmwcousua
.mﬁsuom 585..."? 6:3. acorn?
.835": ScmccoOHoe deem-Eu?
.ucmEEnvF .ESmumF .EmEcoouNF
.8”. :88": .3958qu

«550 no 6.20% .momcmOuh
dragsmuo ..ﬁmﬁum .E:_m_omnv
.m_=m:<um .m=§ms<um .mcacomZnF

 

 

onus;
E3622 ll
=mEm ill

.93: .83 =93. so 8.5

35500

ommvmvjwvwvmmwm3NmomwNwNVNNNomm

L l , Farlpllpllellrl

..._

 

 

Fmrwrwrowm .w W N o

6

Foo.

moo.

omo.

3...

vmo.

8.]qu tax-law ueew

 

 

m0; ocu m>_. .m:o=_m_:ao< brows-:02 .8 22m aeros— cmos. :o 35 «9:22 Loan.— 6.. 2:9".

104

Perhaps not surprisingly, the biggest labor markets India, China and the
United States represent the lowest data ranges with FBU mean market shares
for entry modes other than majority acquisitions of less than .01%. The highest
mean market shares (around .06%) for the small countries are in diverse
European markets, such as Bulgaria and Finland.

We also ﬁnd a slightly positive inﬂuence on the chain size-market share

relationship ()3, = .054136,t = 1.871) due to industry concentration. In other words,

more concentrated markets tend to magnify the relationship between chain size

and market share. The graph below shows country level detail on this

relationship.

Figure 11: Industry Concentration on Chain Size-Market Share Relationship

 

 

 

 

Industry Concentration
3.034~ _ . Low
‘ — Medium
High
1.492“
2
ﬂ
3
m
'5 .733 -
x
h
N
E
.361 .
176 ./,. as... - r . . r r . 7 v - T . ﬂ
0 2.15 4.29 6.44 8.58
Log ln-Country Outlets

105

 

Industry structure effects business performance, which supports the
literature (e.g., Rumelt, 1991; Schmalensee, 1985). The graph above illustrates
that formats in more concentrated markets tend to have a stronger market share
impact from having more outlets relative to FBUs operating formats with smaller
chains. Examples of the least concentrated markets include China and India,
while highly concentrated markets include Northern European countries. While
highly concentrated markets can potentially represent a barrier to market entry,
these can also be the markets in which retailers can use premium pricing policies
and achieve higher proﬁts (Cotterill & Haller, 1992). We are looking at
corporations that already made the decision to enter despite the visible
consolidation of market power amongst the four largest incumbents. In some
cases, international retailers are the “incumbents”. For example, in the Czech
Republic, Ahold, Rewe, Kaufland, and Tesco are the largest chains. One
explanation is that corporations choose to enter highly concentrated markets only
when they have a distinctive advantage relative to the top four competitors —
domestic or foreign. In this pricing environment, unique retailing capabilities
should translate into relatively higher margins and higher proﬁts (Levy & Weitz,
2007). Because we controlled for governance mode, membership in the
dominant segment, and in-country experience, we can say that these cross-level

effects are specific to network size.

The residual variance decreases from Model 5 (100 = 1 .08165) to Model 6

(2'00 2 .63243) to Model 7(r00 = .62218). Controlling for industry concentration

and market size, the range of plausible values for mean market share is 0% to

106

1.60%. In other words, we explained 42.5% of the true between-country
variance in market share by adding the Level 2 covariates. Even after controlling
for centralization and lntemational diversiﬁcation effects, corporations still vary

signiﬁcantly in their mean format market shares
(12 =188.87164,df = 47,p < .001). In other words, average market shares for

foreign chains still show considerable variation after we control for size and
international experience.

In testing the time-invariance of the ﬁnal model, we re-ran the analysis
with market share for the same companies in 2004. This reduced the number of
useable observations to 334. The magnitudes and directionality of the parameter
coefficients are similar and signiﬁcant with the exception of the industry
concentration effect. In the ﬁnal model, 2002 industry concentration is not
significant, suggesting that medium-term format performance does not continue
to be affected by previous industry concentrations. By entering, foreign retailers
offer local consumers one more choice. Some formats are so different from any
other existing formats, they can create consumer demand by offering different or
less expensive products (i.e., make the pie larger), particularly in growing
economies. In most cases, in the act of entering, the foreign retailer is diluting
the market power of retail leaders. Before making the entry decision, retailers
need to assess whether they have the distinctiveness or size to change the
Industry in some fundamental ways. For instance, we have seen large
international retailers, such as Carrefour or Wal-Mart, enter large emerging

markets and have the buying power to change competitive dynamics. This

107

suggests that some prior national effects’ inﬂuence on FBU performance fade
overtime, while other structural factors may have a longer lasting impact (or do
not change very quickly). Retailers need to assess whether their strategic
competitive advantages can help them re-write the local rules of competition to

nullify the impact of certain industry structures.

108

 

:Q

605:0qu 9:30.“:

 

 

 

 

 

 

 

 

 

 

 

 

a NNoh. «002.. a amok. a hath. a E K. Rom dam—2 nova—mm
:N 68595 853...
m mwmoé .. Nova... a vmwo... a mom _.._. a m2. F... mmhoo. . 8% dam—2 EmEEoD
3%. .mmosmam EmEEoo
«new. .83. .85 .88. .98. .88. ..Sm. 2.. .585. 85.58..
3%. .m:o_=m_:co< Ego
. nth- . 3%.- ..m .EoeaaEm .551
a Vommd- a mowod- a omood- mmmho. .3 a mmmh. T a mvmm. _... a noun... ...owt..- 8% duo—2
3%. €8.25
manor—m Dean—

ucoo... ass—533+ «0330... 39m... 6380!... SOD... GO<+ .0622 «mam

 

 

 

 

 

 

 

 

 

 

Eootm _o>o._ auczoo 5:5 oocnzzotom oocoEo>oG «astou— K min..-

109

nova so Emo_.._co_m 2 600E 9632.. 9 52.82.00 .0602. .otva Hm EmoEcQw. 693 um EmoEcQw.

 

 

 

 

 

 

 

 

 

 

 

 

or a m h o m V m ﬂoumEmcwa %
ohmode-P o nnmwdv: 0 man». Pomp o woun- KNF u m2. 3. Four 0 mochﬁmmv ohmhodumw :- 36wa 8.8300
NNNO. VNmo. no 5. .- mmNa. avg. mug. NNON. or K. Eco-.35
N .904
03m... 83... 06:4. 2N»... oommé 3mm. P. :NFN Nun—um «.0 6033.9 3m". 5255
F .90..
3:33:60 52:51

7.
o 3.3. .cozgcoocoo bum—Sm.
. ER. .mmam. . o .2... a... 2250 5.22
am .2250 .28
.38. .mowm. .Nam. .38. 2.. 38$me 522

 

 

 

 

 

 

 

 

 

 

h gnu-P Pup-00

110

Discussion of Hypothesis 2 Findings

The corporate and country models testing Hypothesis 2 reveal mixed
results. In the corporate models, we see the predicted ordering of performance is
supported — wholly-owned subsidiaries have the highest performance,
acquisitions the worst, and joint ventures in between the two. In contrast, in the
country model, acquisitions have a higher performance than the intercept, which
presumably captures the RWOS and JV modes. This means that after
controlling for membership in the dominant segment, size, and local age, we are
left with a better result for acquisitions than the other modes. This contradicts
theoretical predictions that in ex post expansion WOS should outperform other
entry modes due to relative ease of cross-market coordination and collaboration.
Fluid knowledge flows within the MNC should enable FBUs to realize synergies
between strategic resources for higher performance, supporting Rumelt (1974).
Differing covariates may partially explain the varying results. By definition the
dominant and related covariates captured the most successful formats, which
may also be correlated with some of the governance terms and the dependent

variable. Format dominance is signiﬁcantly correlated with RWOS (,0 = .318) and
market share 2003 (p = .213). This is consistent with the prediction that retailers

internalize their transactions to protect their strategic assets (Dunning, 1980).
We also see that the contextual variables have varying influences on retail
performance over time. Some strategic and structural factors have long lasting
implications, while others do not. In summary, our two sets of analyses provide
seemingly contradictory results.

111

In the corporate model, joint venture performance surpasses UWOS and
acquisition entry modes, but exhibits smaller outcomes than RWOS. We also ﬁnd
some evidence that international diversiﬁcation experience strengthens the
relationship between JV and performance. In starting new joint ventures, both
partners face information asymmetries and can have similar responses to
opportunistic behavior of the other partner (Axelrod, 1984; Woodcock et al.,
1994)

Some caution should be taken with the results. The variation in
performance amongst the companies using JVs could either point to a
governance competency or simply reﬂect imposed partnerships by host country
governments. Some companies may possess a speciﬁc competency in
managing joint ventures, while other companies may just be forced to accept
joint venture partners in certain restricted countries. China and Russia, for much
of the 1990’s, required foreign companies to engage with domestic partners in
order to enter the market. Our analysis does not allow us to differentiate between
these two cases.

Mixed results for the acquisition mode muddy the interpretability of the
ﬁndings. In the corporate model, ACQ underperforms all others. Moreover, this
result was exacerbated when the parent was centralized. In contrast, the country
model shows ACQ outperforming the other modes combined. This ﬁnding seems
more reliable given covariates for labor market size and membership in the
dominant segment. In other words, holding market size and most important

segment constant, acquisitions perform better than other modes.

112

To understand why the analyses might reveal these results, we go back to
the literature. Traditionally, the assumption is that mode decisions are difﬁcult
and costly to change (Williamson, 1985). Though this may be true, it may not be
as relevant in the context of multi-business retailing. Wal-Mart’s purchase of a
stake in Seiyu in Japan illustrates the point. In 2002, Wal—Mart initially purchased
a 6.1% stake in the Japanese company and increased its stake to 53.3% by
2005 after Seiyu met performance benchmarks. Moreover, because the first
move into a format may have been one mode, this does not preclude expansion
via other modes. For instance, Ahold practices “strategic inﬁlling” to build up its
store networks when necessary for efﬁciency gains (Elshof, 2005). Firms also
make shifts over time in their preference of entry modes (Lu, 2002). In other
words, mode preference may be less stable - even within the same corporation
or country — than is presumed. Though some retailers, such as hard discounters
Aldi and Lidl do tend to stick to one mode (e.g., WOS), others do not. In short,
ﬁrms do not necessarily utilize the same mode consistently, supporting Lamont &
Anderson (1985).

Finally, market structure variables influence our results. Larger labor
markets are associated with smaller format mean market shares, after controlling
for dominant segments, in-country experience and chain size. In addition, retail
market concentration magnifies the chain size - market share relationship. In
support of Cotterill and Haller (1992), retailers choose to enter concentrated

markets to attain premium prices and profits from their competitive resources.

113

It may be instructive to challenge whether we should be looking at
governance mode and direction separately for multi-business operations, such as
retailing. As suggested by Ramanujam & Varadarajan (1989), it appears to be
more instructive to look at the joint decisions that make up diversiﬁcation
strategy, i.e. scope and position. Though they may not be made by the same
people, these decisions are made simultaneously in the expansion process.
When only one element is considered, we may be misrepresenting diversiﬁcation
strategy. In addition, because we know that F BUs are adjusting their strategies
on a format-by-format basis, we have an opportunity to dig deeper into the
relationship between FBU diversiﬁcation and its associated performance

Outcomes.

Hypothesis 3

Hypothesis 3: Related diversiﬁers that expand via organic growth should
outperform all other strategy mixes. Conversely, we expect that unrelated
diversiﬁers that expand via partnership or aoqirisitions to be the lowest
performers.

The ﬁnal hypothesis relates to the relative efﬁcacy of six key diversification
strategies (i.e, related/unrelated and 3 govemance modes) on format
performance. We predict that related diversifiers of all types should outperform
unrelated diversifiers. In Table 8, we observe two things. First, in our sample,
retailers use related diversification strategies more than unrelated strategies.

Second, we are not seeing the expected frequency of pairs that one might

expect. For example, retailers use UACQ almost as often as RJVs. We conduct

114

several analyses to test the impact of these strategies on performance and
discuss the ﬁndings below.

Table 8: Strategy Frequency

 

 

 

RWOS RACQ RJV UWOS UACQ UJV

 

 

 

Corporate Effects

In the next series of models, we attempt to clarify the variability of the
regression equations across corporations. Our Level-1 model remains the same
as model 5. In models 6-7, we progressively incorporate corporate level
explanatory variables, namely centralization status and corporate size (i.e., total

selling area).

115

Final Model

LNMKTOBU. = 700 + 701tCENTRALj + 702::(SELLAREAJ. - SELLAREA. ) + 710: RWOS”. + 720: RJV”.
:RACQU + 740: UNWOSij + 750' UNJVI.j + YGO’HYPERij + 770::SUPERMKTI.j

aunCONVENIEI.j + ygosDISCOUNTij tLNOUTOZI.j + u

+730

+Y1m +rij

+ 730 0i
We test whether being from centralized corporations has an effect on

format’s mean market share. By adding an explanatory variable to the equation

at Level 2, we change the meaning of the residual, ac] and make the variance,

Too, a conditional variance of average corporate market share after controlling for

centralization. Similarly to previous models, we observe a signiﬁcant negative

association between centralization (970, = —.870249,t = -2.280) and mean market

share in Figure 12. The residual variance between corporations is 2'00 2 .78834.

A range of plausible values for corporate means is between 0% and 1.82%.
Further we explain 18.48% of the true between-corporation variance in market
share by adding the centralization covariate. Again the Daiei formats are
behaving signiﬁcantly differently from the rest of the formats. The company’s
operations are atypically small and generate too little market share, as compared
to other ﬁrms’ foreign retailing operations. In general, decentralized companies

have higher mean market shares than centralized companies.

116

 

 

>Imuvm .88“. going sun. 2
onuNm .5286qu .ocmﬁoouuom
.saaonmm .tm2-_m>>um~ .>z:u\.N

.oomohuom .ccmE_omco._.um~
.:>_omuvm .>m2,o._mwnmw .oBomnNN
.siaocmiu .oseozuom
.965.an .63"? 53269.":
.ocmxo>-o:u9 .53 wine
.2590 oaawui £320"?
.mN_m£on_uNF .75": .580qu
.oogmoonm .EoOuw .oEmmo nu
Sohotmo um .oootaum .cmco:< nv
._o_<um .o_on<uw .coo<ur

 

 

mm;

02 -l

Eaton. omuﬁbcoo

23:22:00
omwmwmnmwNmeNmuwwmwwvLime?m

 

o v

_ 2

 

Foo.

moo.

vmo.

wow.

oovn Jar mus 19mm ueew

 

 

2.25 3qu 095 coo—z :o «022:. cozauzmbcoo "NF 0.59%

117

In the ﬁnal model, we add a corporate size covariate (e.g., total selling
area). We observe a signiﬁcant negative association between centralization
(701 = —.573722,t = —2.066) and corporate size (fez = —.000016,t = -4.1 10) and

mean market share. The residual variance between countries is r00 = .29153. A

range of plausible values for corporate market share means is between 0% and
1.15%. The model explains 63% of the true between-corporation variance in
market share.

By the last two models, the relative magnitudes of the governance

variables stabilized to the following ordering:
RWOS>RACQ>RJV>UWOS>UJV>UACQ. Although the ordering reversed
between RWOS and RACQ by the ﬁnal model, the surprise is that RJV does not
outperform an RACQ in any of the models. In fact, in the ﬁnal two models with
Level 2 covariates, UWOS and RJV parameters have similar magnitudes and
directions. Although the UWOS is not significant, even if the true value is zero,
this suggests that some companies may find it almost as difﬁcult to have a
partner in' a known business as to start a greenﬁeld investment in an unrelated
format.

In testing the time-invariance of the ﬁnal model, we re-ran the analysis
with market share for the same companies in 2004. This reduced the number of
useable observations to 327. The magnitudes and directionality of the parameter
coefﬁcients are similar. In the ﬁnal model, selling area is not signiﬁcant,
suggesting that medium-term format performance does not continue to be
affected by the corporation’s global size. We have to revert back to Model 6, with
only the centralization effect with similarly signiﬁcant parameters. This suggests

that prior corporate size effects cease to inﬂuence F BU performance over time.

118

 

mmvno. T

:59-

nova—MT

a ONMN. T

KVNN. T

«OQON.T

avnom. T

”comm—x..-

2% 582 >3

 

R§

0.3:0> E6... 020.0093

 

owhm.

VONN.

mwoo.

mvmo.

mmmo.

Nmoo.

vomo.

mnwoo.

2% .52). 09>:

 

:0 522830 8:20
2.22, 85.25

 

a mvom.

a mwmn.

m vomm.

amwvm.

a mnmm.

a mama.

«comm.

meo. _.

as .505. 09».

 

2%

.:o=_m_:co< “emu—0.0.x

 

05mm.

omom .

owmo.

wroo.

mnmo.

vmmo.

oomo.

#5....

as .522 >2

 

3Q

.0._2c0> E2. U0Hm_0¢m

 

a momm.

wowmoe

a momm.

evvvm.

m womm.

92.5.

avomo. —.

2% .502 002E

 

:m 52233
8:30 >__o§> 85.0”.

 

morocco:

8% .0~_w 0205900

 

ammmmr

a monwr

SR
.30an0 UON=SEOO

 

a CROWN-

m mmomN-

amnovN-

«mm—.mé-

sky—Le-

awmnw. T

ammmo. T

«ammo. T

8x
.36 “03:05. 00002

 

3%. .3020:

 

30th U0!"—

 

 

36+

 

_.w.=:00+

 

$330+

 

005+

 

c0>coO+

 

L0osw+

 

c0d>z+

 

.0305.
008

 

 

 

muo0tm 083960 5.3 325 cos—859.020 _0>0._-00E._ou_ "m 030...

119

oﬁvq E 2:005:90 m_ _0uoE

£5305 2 :85an0 .0622 u .moVa 6 2:85:90 3 _0ooE 030305 2 50:00:80 .085. o .oﬁva “a E85369 .mo.va “0 E8556

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

9. 3. Mr N_. E. or m m 0.90:5qu %
n8 .5 on um um cm ..0 NV 8558

www.mm: mnoko: VNMNNN.‘ NNNQONr mr VEN—t omothv mmo.NmN_. nmmev .
mm 5N. vmmmm. monom. “wwwn. tum Eu. owwa. vmovo. omvvo. 30825
N _0>0._
b
Donn _‘ mach _. wrov _‘ mooo N meo N mmor N 2%.. N NNON N .:o=mtm> Dmn— EES>
_. _0>0._
35:2. :30 Eon—Em
a mnvm. a Row. a mmow. act“ 6005. 36 HEB".
3mm. .0N_w HEB".
50v. 3mm. Noem. n.mmow. 8% 6005. 05.5805
an £025,805 com
a wmmo. T a no Po. T vaor amomor mommmr 8% 0o:0_c0>coo
gm 99m 0oc0E0>co0
w 3N. EON. a mnmm. a mNmm. amva. a onm. 2% 600—2 uwvtmctwasw
N..momo....u.._zzc0n_q.w
8k

a 0N8. a 0 row. a vn E. a Numm. amONv. a 00mm. nxwmm.

600.2 “00:05.00?

 

 

 

 

 

 

 

 

 

 

3% #35859”:

 

 

m 032. 63:00

120

Country Effects

The next series of analyses models country-level effects on F BU format
operations. The Base Model to Model 5 clarify the Level 1 (FBU) model. Our

Base Model includes ﬁve of the six governance pairs. We ﬁnd that parameters

for RWOS (2710 = .701736,t = 2.87), RJV (72,, = .541675,t 21.881) RACQ
(7'30 =1.154387,t = 4.690) and UJV 0750 = —.774334,t = —1 .702) are signiﬁcant,

while the parameter UWOS is not significant. The directions for each of the
related/unrelated variables are as expected, but the magnitudes are not. In this
model, estimated mean market shares are RACQ>RWOS>RJV and the superior
performance from being an RACQ does not change in our country analyses.

In the next series of models, we attempt to clarify the variability of the
regression equations across countries. Our Level-1 model remains the same as
model 5. In models 6-7, we progressively incorporate country level explanatory
variables, namely retail labor market size and retail market size (i.e., total selling

area).

Final Model

LNMKT03U. = 700 + 701*(LNEMPLOYJ. - LNEMPLOY. ) + 710*RWOSU + yzotRJVij + 730*RACQU.
+ 7401.:UNWOSI.j + 750*UNJVU. + Y60*HYPERU. + 770*SUPERMKTU +
780*C0NVENIEU. + ygoauDlSCOUNTU. + momLNOUTOZij +

*(LNRETAIL. - LNRETAIL )tLNOUT 02.. + u . + r..
J ' U 0} U

 

Y101

121

In Figure 13, the size of the retail market labor force may have an effect
on format’s mean market share. Similarly to previous models, we observe a

significant negative association between retail labor market size

(270, = —.270479,t = —4.226) and mean market share. The residual variance

between countries is too = .61858 . A range of plausible values for format means

is between 0% and 1.59%. Further we explain about 41% of the true between-
country variance in market share by adding the labor market size covariate.

In the ﬁnal model, we add a corporate size covariate (e.g., total selling
area). We observe that corporate size (fog = —.072936,t = —3.042) has a

dampening effect on the relationship between a format’s chain size and retail

market share.

122

 

82:6 958.... 03

 

 

mm... v20 aw... mew o .
. . . . . . . - ; . . . . . . . . ...\ m8
- 5o.
.1 3m.
2
:33
090.. -
5:622 w .,
=me l “8:

2:22 :29... .6 020

BJqu tame“

 

 

3:05:20”. 0:25 “9.3.2.35 SE3".
:o «009:. $5 «9.3.2 :33. 6.. 0.59“.

123

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.85. .88. .32. .923. .03... .33. ...mm. s“. 582.02.552.21
3% ..0..0E.0a2I
38.- 08m.- ..08. .23.- ..mmm. .25.- 3...- .25.- ... .cm22>3
2m
0.5.:0> 2.22. 00:20.25
.0... A3.... 80.. £5.- 88. 0.8.- 00R. 0.8.- 32.282002:
.2... 5222.30
855 26:22 0222::
.mwmm. .88. .28. :28. .030. .88.. .88.. .32.. 2.250200%
2% 622.2300... 00.0.0m
.008. .50.. .m.~0. 8.... 00m... .30... .NN... ...vm. 2.. .5822...
3.0. .05.:0> .50., 00.0.0m
.080. .NE. .38. .008. .58. .28. 8.6. .50.. 2.. 5.0200222
:2 522830
8:26 2.92., 022$.
LEN.- .mEN. 52.020.352.83
.880- .Nm.o.m- .823. .82...- .88..- .83.- .83.... .28... 2. 020.2202 822:.
3%. 300.02..
202.0 02....

0290.5? 8.5.5.? .0.SO+ 005+ 250+ 005+ .0027? WWW“.

 

 

 

 

 

 

 

 

 

 

0.00.3 22560 5.3 30.25 co_.00:_0.0>_o .0>0..-.0E.ou. 5. 0.00..-

124

9 50:09:00 .0005. . .mo.vq .0 E00...

E90 0. .0uoE 0.6302. 0. 50:00.28 .0022. .o...va .0 E00.

o..va .0 2:005:90 0. .0coE 030305.

590.. .83 5 582520.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m. 3. m.- N—. E. o. m w 0.0.05.0..01 n

o m u . a m u N . u m 0.. mm 00c0_>0o
ormNmor ovo.VNo.- Nva. E. mm®.®VN—. 8 ..meN. VNnémN. mvamN. meoN. .

mva. 09.0. mmvo... ..vvn. Numn. ..an. omon. Shh. 3020......

N .020.

mmVNe 9.3.... 9.5.... 30:... mwmn... awn-K... mrrmé Nmmm... "-0 .co..0..0> Dmu. GEES

. 55:

3:050:80 Eoucam

.85.. a... 0.05 202.05.

. mm .m. . .mom. . mt... a... 600.). 005 202:8“.

5%. .0N_m HES”.

0.0.. 320.. N09. .38... a... 6005. .02-580.0

5%. 2025805 tow

. vovv. .- . KS. .. . mmvv. .- . mvmmr . 3.8.- a... 00c0E0>coo

9% 0.05 0825200

mm .o. mmmor mmvo. . 05m. . 3N... . omvm. 2.. 2005. 02.0.2003

...m.0x.0E.0a:m

 

av 0.90... 63:00

 

125

We observe that in markets with a greater number of retail outlets the

association between format size and market share weakens. The residual

variance between countries is rm 2 .29153. A range of plausible values for

corporate means is between 0 and 1.11%. Compared to Model 5, the ﬁnal model
with two country size variables explains approximately 72% of the true between-
country variance in market share. By the last two models, the relative
magnitudes of the governance variables stabilized to the following ordering:
RACQ> RJV>RWOS> UWOS>UJV>UACQ. This contradicts predictions in the
literature that would have the RWOS strategy dominating others.

In testing the time-invariance of the ﬁnal model, we re-ran the analysis
with market share for the same companies in 2004. This reduced the number of
useable observations to 327. The magnitudes and directionality of the parameter
coefﬁcients are similar and signiﬁcant. However, for the related pairs mean
performance ordering changed again (RACQ>RWOS>RJV). That provides a
second piece of evidence that RACQ outperforms other related governance
modes. In addition, the structural industry variables (i.e., retail labor force size
and total retail outlets) remain an inﬂuence on the intercept and the chain size-

performance slope.

Discussion of Hypothesis 3 Findings

If we compare the final Level-1 models and the final contextual variables
with country and corporate affects, we see that in three out of four models,
RACQ is the most successful diversification strategy. The reliable 2004

corporate models switch the order with the RWOS garnering more market share

126

than the RACQ. whereas the country model has RACQ>RWOS>RJV. Extant
research suggests that for related diversiﬁcation, RWOS should outperform other
strategies (Busija et al., 1997; Rumelt, 1982). Our ﬁndings do not support this
hypothesis. We find that for most companies and in most countries, the RACQ
strategy outperforms a RWOS strategy.

We must go back to the underlying theoretical assumptions about causal
ambiguity and absorptive capacity to look for clues. Our results suggest that, for
average F BU operations, both causal ambiguity and absorptive capacity are high
enough to make buying knowledge from an outside firm less costly than
developing it in-house. Some evidence implies that retailing operational know-
how is more codifiable and transferable than other kinds of complex
manufacturing know-how, such as R&D (Morgan et al., 2003; Szulanski, 1996).
High causal ambiguity results from imperfect mobility of ﬁrm resources, leading to
heterogeneous firm performance (Barney, 1991). This is perplexing since
retailers who operate within the same segment should understand retailing
fundamentals for that type of business, particularly if most of the sticky
knowledge is embedded in back office operations (Coe, 2004). Our findings
suggest the opposite — success in local market operations resides in the store-
front or the visible parts of the 4P’s (Kotler, 1997)”.

Our findings are more in-Iine with Vargo & Lusch (2004). In a wider
marketing context, they deﬁne services as “the application of specialized

competencies (knowledge and skills) through deeds, processes, and

 

‘3 However, visibility does not mean reproducibility. Firms can possess visible advantages, but
how they arrived at them may be difﬁcult to assess (i.e., causal ambiguity).

127

performances for the beneﬁt of another entity. Knowledge competencies are
considered the firm’s most important resources and sources of strategic
competitive advantages (Day, 1994). Building on the resource-based view of
resources, resources can be operand or operant. Operand resources are
physical, inert resources that must be acted upon in order to produce an effect,
whereas operant resources are the assets that allow the ﬁrm to act upon the
operand or other resources (Constantin and Lusch 1991). For example,
exemplary supply chain management competencies function as operant
resources producing strategic competitive advantages (Srivastava, Shervani and
Fahey 1999). We can view ﬁrm knowledge competencies as “operant resources,
consisting of prepositional (abstract) and prescriptive (technical) knowledge,
which allow ﬁrms to apply their knowledge to create better value propositions for
their customers relative to their competitors” (Vargo & Lusch, 2004a, p. 5).
Operant resources are also less transparent to competitors. From a service-
centered view, customers are “co-producers” of the value proposition by
interacting with the ﬁrm to infuse the product/service with value and by actively
participating in the exchange. VIchin this context, operational retail knowledge is
technical, whereas the customer-interface in the stores is more abstract. All
things equal, when an international retailer buys a local ﬁrm, it is tapping into the
ﬁrm’s prepositional resources. Used to dealing with and adapting to complex and
changing consumer expectations, retailers may possess collaborative
competences, a higher order resource that helps ﬁrms “continually renew, create,

integrate and transform their service innovations” (Lusch, Vargo, 8. O'Brien,

128

2007, p. 9). If we view retailing as fundamentally a local phenomenon, then cost
and time to develop these resources in-house would be too high and too long.
Furthermore, Tanriverdi 8. Venkatraman (2005) address potential knowledge
advantages for multi-business ﬁrms, including customer relationship
management and alliance management. Thus, despite Ansoffs (1988) assertion
that strategy should not overly emphasize customers (relative to business
mission), the strongest retailer advantages seem to come from prioritizing
customer needs.

Less surprising perhaps is the fact that related strategies outperform
unrelated strategies. Delving into new, unfamiliar terrains is risky. Moreover,
MNCs should possess unique advantages relative to competitors to be able to
compete and win in foreign markets (Buckley & Casson, 1976; Dunning, 1988).
The costs of integrating organizational cultures and managing completely new
businesses make it nearly impossible for unrelated acquisitions to succeed
relative to other diversiﬁcation strategies, providing additional support to Porter
(1987). Yet, theory suggests that unrelated acquisitions should be the most
successful of the unrelated strategies since the costs of internal knowledge
development should be much higher than the purchase cost of that same
knowledge. In fact, the UWOS strategy is most successful in our sample.
Because the retailer would lack both the technical and the abstract knowledge to
operate the business, casual ambiguity would be particularly heightened.
Selection and management of the appropriate partner would also be particularly

costly and difficult. Instead if the retailer wants to experiment in a new area for

129

that market, at least the organizational culture is constant, facilitating better
communication, which should ease Ieaming and knowledge transfer. An
alternative explanation is that, formats using a UACQ strategy are part of larger
portfolio of acquisitions, whereas formats using UWOS represent proactive
attempts by the international retailer to learn a new business.

In the retailing industry, business models can shift dramatically between
segments (Bhatnagar & Ratchford, 2004). In terms of the value/price continuum,
hard discounters primarily make a proﬁt by delivering on a pricing promise to
customers and depending upon higher sales volumes and modest customer
service levels. For instance, Aldi and Lidl offer their customers rock bottom
prices, wide product variety and shallow assortments, and reasonable quality
private label products to inﬂate otherwise thin margins. While this fundamental
business model does not drastically change within a discount segment (e.g.,
hyperrnarkets, supercenters, discounters, etc), another segment (e.g.,
supermarkets, convenience stores) earns proﬁts in a radically different way that
emphasizes the product quality and customer service promised to justify higher
markups. Whole Foods is a high-end supermarket known for its organic and
specialty foods, deep assortment of wines and superior customer service.
Business models between segments are driven by fundamentally different retail
strategies to satisfy vastly different consumer expectations. Consequently,
decisions to expand within the segment, irrespective of mode, are consistent with

retail marketing realities.

130

These ﬁndings also support the excess capacity and synergy creation
hypotheses. International retailers are able to use their slack resources
effectively in foreign markets. Operating internationally, regardless of their
governance choice, provides retailers with opportunities to discover new
operational applications or synergies and to spread the costs between
businesses (Markides & VWliamson, 1996). Ansoff (1969) also argues that when
managers try to expand at a faster rate than can be sustained internally,
expansion via acquisition is the most used alternative.

We turn next to the ordering within the three pairs of related diversifiers.
Interestingly, our findings indicate the RACQ strategy is surprisingly effective for
international retailers. Foreign ﬁrms seem to be co-opting the competition by
joining forces with them. This supports the ﬁndings of Trautwein (1990).
Logically, a critical rationale for retailer acquisitions is access to suitable real
estate for the format. Conversely, poor network real estate is a deal-breaker in
retailing, and as such, those chains with poor real estate would be less likely to
be acquired. Real estate is probably even a more important factor than the retail
brand itself. Retailers have been known to absorb other chains and phase out
local brands. For example, Tesco bought Kmart stores in the Czech Republic
and converted them to their hypermarket brand. Even their own brands are
replaceable. Recently, Carrefour Champion in Spain could see that they were
losing the supermarket wars, so they re-branded Champion stores to their
hypermarket format. Strong retail brands are an addedbonus to a strong

property portfolio. When consumers form emotional attachments to local brands,

131

foreign retailers can instantly acquire name recognition and customer loyalty.
Retailers also obtain access to local managerial talent and organizational
knowledge of how to operate in that marketplace. Because retailers have very
high fixed costs and low margins relative to manufacturers (Cortsjens &
Cortsjens, 1995), pressure to deliver revenues to offset high entry and startup
costs is intense. Buying into a knowledgeable managerial pool can accelerate the
process of making money. Moreover, because the acquirer understands the
basic rules of the segment, it is in a better position to evaluate and select
companies that have a reasonable level of existing expertise (Eisenhardt, 1989).
Thus, potential access to prime real estate, known brands, and local retail
expertise make related acquisitions an attractive diversiﬁcation strategy.

The evidence on related joint ventures is somewhat equivocal. Given our
previous analyses in Hypothesis 2, the ability of the parent company to manage
foreign joint ventures varies signiﬁcantly, even for experienced lntemational
retailers. This suggests a particular capability may exist to manage international
joint ventures (lJVs). However, Barkema et al. (1997) suggest that it is not the
corporation’s international diversification experience, but rather its experience in
managing domestic joint venture partners and international WOS that aids IJV
survival. Wnile we cannot refute this study, our data suggests that international
diversiﬁcation experience inﬂuences how well retailers can translate an RJV
strategy into performance.

While there may be learning benefits and other advantages to fully owning

and operating a foreign subsidiary, it is rather unclear whether it pays to do so. In

132

traditional TCA/intemalization literatures, a firm chooses governance forms to
minimize transaction costs and protect proprietary information and products.
Further, Woodcock et al. (1994) argue that ex post costs of marginal
procurement and ownership and control costs should make new ventures less
costly. We see that the RWOS strategy outperforms the unrelated Counterpart,
suggesting some kind of segment expertise and potentially other capabilities
brought from the home country boost RWOS performance. Since Hypothesis 3
benchmarked all the other diversification strategies against the expected best
performance of RWOS, we can say that this ordering hypothesis was not
supported. Although our data shows this strategy being the most frequently used,
it is not necessarily the best strategy option.

To a certain extent, the differing covariates may also explain some of the
divergent results. We see that the contextual variables have unstable influences
on retail performance over time. Logically, we would expect that corporate effects
change more quickly than industry effects. Firms have the ability to make
significant strategy shifts whereas major industry changes occur less frequently
and with lesser magnitude. Our study supports Rumelt (1991), who ﬁnds that
industry structure has a small, but lasting impact on ﬁrms. From a public policy
perspective, the predictability of the industry change is an important factor in
attracting (and keeping) foreign investment.

In the next section, we will attempt to draw conclusions from the multiple
pieces of evidence we have collected. Collectively, these analyses help us tell a

story about the value of diversiﬁcation to retail performance in foreign markets.

133

CHAPTER V
CONCLUSIONS

The objective of this study is to understand the diversiﬁcation-performance
relationship in the context of international retailing. We treat MNCs as economies
with vertical and horizontal knowledge ﬂows that interact with and between each
level of operations (Madhok & Liu, 2006). Operating in diverse local
environments and possessing heterogeneous resource advantages, F BUs
develop diversiﬁcation strategies intended to elicit successful outcomes.
Knowledge and other competencies gained in foreign markets are not only
critical for F BU success, but also crucial for MNC development of strategic
competitive advantages (Teece et al., 1997). Being able to exploit and expand on
these strategic competitive advantages is the main challenge of diversiﬁcation
strategy.

Using a rich literature base, our research challenges the relevancy of
extant theory at the foreign market operations level. Table 11 contains a

summary of our ﬁndings.

134

 

002.00.00.00
.050 000000.30
0.3000 0000.0.0030

 

0000000000 am“. 00 0.0000 00.00. .30 ...000 000 0.3.02.0 b.0300. . 0.000 .0000. . 00.56-2003
0200.020 00.0.00: 00.... 00000 000.000 0200.020 00.0.00 . 00.00 200.000 00302.. 0200000
002.00.00.00 200.020 .050 00000930 002205000 00.0.00 . 02.00.20.020 0 . 0.050.020 00.0.00 0
0000000000

000 0000. >3 0003.00 0.0000220. 00. 00500020 000 0.00.0020
000200020. 000000000 .000.0>0 00 0003.00 0. =0. 00.3.0? .06.. o 0.000 .0000. . 00222300000030?
0.000 .0000. 000.00 .0000... 00. 000 002023000 .0020 20.0300 0... 0. o 0000. .0.o..A0000_0_0n.30
00.2000 00 .00020 000.0 00.000.000.30 mO>>m .0090 0.0000000 00. 0. . 0000E0>o0 . 000.502.0022 N

 

02.0.8000 0000000000
r 00.0 .0330 0... 000.000 0.0000. 00.0.0E0 000.000 0. 0.0000.

0000.000 00.... 00.000 .9000. 000.0 .026. 0>00 0.0.00.0 00.0.0Em. ..
0.000

.02-.0E 00000>0 .000 5.2, 00.0.8000 0.0 0.0000. 0000. 20.0. 00.0.. .
0.0000020. 0000000000 r 00.0 .0030 .0000. 00.0.0. 0...

0000.000 003.00 .0330 0000. 0 0.05 02.080000 E0000 0 E0... 00.0.00 .
0.000 .0000. 000.0

.000... 0. 00.0.0. 2.022.000 0. 00000090 02.00.20.020 002.0520. .
0000000000 000 02.00.20.030

0003.00 0.0000220. 00. 000500000 00000090 20.0.. .o 0.000.. .

 

 

 

0.050.020 00.0.0003 002000.30 0.050.020 00.0.00 . 0000800000
0000002000 0. 0.0000020.
000 02.00.20.020 .0000. 0003.00 0.0000020. 0000.. 02:08 < . 0.000 .0000. . 000000 3 000020.
00000030 .00 300022230 . 02.00.20.020 . 00 000 02.00.00.020 .
000.00.". 20v. 00.0000> 20v. 0.0002200 I

 

 

 

0:300”. 00 200.0an N... 0.00...

135

 

For bricks-and-mortar retailers, diversification is generally a sound
strategy. In our sample, the average foreign ﬁrm garners between .07 and .09%
retail market share. With a one unit change in diversiﬁcation, an average ﬁrm can
double its market share, holding size of operations constant. We ﬁnd
approximately the same result in both our country and corporate models.
Furthermore, our sample reveals no evidence to suggest that retailers only
beneﬁt from a certain degree of diversiﬁcation. Our data strongly suggest a linear
relationship between format diversiﬁcation strategy and performance.

Our ﬁndings also address how F BUs diversify. While FBUs beneﬁted
more from sticking within their retail segment, between segment expansions also
facilitate market share boosts. This is conﬁrmed when controlling for corporate
and country factors. These results support the complementary relatedness
hypotheses and provide further evidence that related diversifiers should
outperform unrelated diversiﬁers. However, even unrelated diversification is
associated with market share gains. Experimenting with new formats may also
possibly aid long-term survival prospects by supporting Ieaming and innovation
objectives.

Our ﬁndings also shed some light on contextual influences. Age matters.
More experienced ﬁrms are able to translate their organizational knowledge into
stronger gains from their diversification strategies. This implies that knowledge
sharing is taking place between the parent firm and its foreign operations.
Conversely, size doesn’t always matter. Larger corporations do not see the

same results as smaller corporations from large retail store networks. In addition,

136

corporate size can explain mean market shares better than lntemational
diversiﬁcation strategy, or the number of countries in which a firm operates. This
suggests limitations to worldwide expansion strategies. In summary, age, size,
and international diversiﬁcation experience help us explain about 45% of the
variations between corporations.

Our country model findings also offer some interesting insights into FBU
operations. First, level of development and market size are important variables in
explaining between country variance. Opportunities to capture higher market
shares in emerging markets help explain differences between mean performance
between emerging and developed markets. Our results also suggest that
conventional wisdom that larger F BU chains outperform smaller ones may not
hold in emerging markets. Higher logistics costs in emerging markets may
account for larger chains’ inability to capture market share. In other words, the
coordination costs associated with maintaining larger chains in more challenging
retail environments may limit merchandising strategies required to capture
customer loyalty and repeat purchases. This is an avenue of research worth
exploring.

Our last two hypotheses explore the nuances of diversification strategy.
First, we attempt to explain how F BUs govern their businesses. In our corporate
model, we ﬁnd that mean RWOS market share is four times larger than JV
market share and approximately seven times larger than ACQ and UWOS. This
result needs to be taken with caution because our country model shows that

ACQ outperforms other combined governance types. The models are also not

137

directly comparable since they control for membership in the dominant business
segment, experience, and size in different ways.

Level 2 covariates are also tested. In the corporate model, we
demonstrate that the ﬁrm’s level of lntemational diversiﬁcation inﬂuences the
average FBU’s ability to manage its joint ventures. We also provide some
evidence that having a centralized parent explains some of the variance in the
intercept term. In the country model, the size of the retail labor market and the
level of industry concentration inﬂuence our FBU parameters. Countries with
large retail labor markets are more attractive for retailers. Men our intercept
represents non-chain acquisitions, retail employment explained a signiﬁcant
amount of the variance in mean performance. Taken in conjunction with the
beneﬁts from concentrated markets, we explain about 43% of the differences
between countries.

Our last hypothesis provides evidence on the relative efﬁcacy of six
diversiﬁcation strategies. By jointly considering mode and direction and
controlling for many key formats, we reveal a clearer picture of mean market
outcomes. Our ﬁndings suggest that RACQ strategies have greater mean
performance outcomes than suggested in the extant literature, while UACQ
strategies exemplify expected poor outcomes. Simultaneously considering key
elements of diversiﬁcation strategy supports beneﬁts of acquisitions theorized in
the literature. Similarly, joint ventures and WOS strategies did not have the
expected average performance relative to RACQ strategy. In our two models,

we test several effects including corporate centralization, labor market size and

138

retail market size. We are able to explain 63% of the between corporation
variance and 72% of the between country variance, which would significantly

impact regression results if not controlled.

LIMITA'I10NS AND FUTURE RESEARCH

Some caution is warranted in interpretation and generalization. These
models do not prove causality, merely demonstrate the association between
variables. Also, since our data structure is unbalanced, the parameter estimates
are likely better representations of the F BUs and countries with more
observations than those with fewer observations. However, given the enormous
costs and data constraints in FBU data across nearly 49 countries, we believe
that these data represent a realistic starting point for international retailing
analyses. We also use an entropy measure of diversiﬁcation. Critics argue that
entropy measures do not measure actual, shared resources and capabilities and
their market cost for undiversified companies (Markides & Williamson, 1996).
Finally, market share is not the ideal outcome variable, but it is the only reliable
variable at this level of analysis across corporations and countries. In contrast,
most diversification studies use other ﬁnancial and accounting performance
measures. We believe that market share is a consistent and reliable performance
measure widely used in retailing practice. Therefore, our results are relevant for
retail practitioners.

Another issue concerns the limitations of the market share measure.

Although common in the literature, there are speciﬁc challenges to using a

139

measure that is bounded between zero and one. Schwartz (1991) discusses the
inherent problems with using this type of variable in regression-based
techniques, including breaking of the unboundedness and connecting the
variance of endogenous variable to its value assumptions. The consequence of
using a bounded dependent variable is that the relationship between the
dependent and independent variables can change at different points along the
line. For example, if ten years difference in in-country experience leads to a 5%
increase in market share, then that relationship should continue at other points
along the predicted line. Thus, if the predicted market share is 10% for a
company with 30 years experience, then it would be 15% for a company with 40
years experience. Although violations of regression assumptions do not
necessarily negate the value of the results, statistical inferences are considerably
strengthened by ﬁnding alternative methods that meet statistical assumptions.

Solutions to this limitation will be pursued in future research. One possible
resolution is to create an arcsine transformation on the market share variable
(High, 2005). Another solution could be to use another kind of HLM. The
Hierarchical Generalized Linear Model for Binary Outcomes can be used. The
predicted log-odds could be translated as the probability that an F BU within a
given corporation or country would be chosen by domestic customers.

The intra-industry context limits the range of diversification choices
examined in this analysis. Though most of our food retailers were not vertically
integrated or operating in other service industries, this limitation may be

responsible for the lack of curvilinearity seen in other diversiﬁcation studies.

140

This study represents a ﬁrst step in examining international retailing
diversification strategies. Taken collectively, our analyses provide some
compelling evidence that foreign market operations are an important and
appropriate subject of study. While our research clearly supports some aspects
of extant theory, we also raise some new questions. These questions provide
the basis for future research.

An immediate extension of this research includes testing the stability of
the parameter estimates and finding more contextual variables to explain country
and corporate differences. RBV and diversification theorists suggest that
strategic competitive advantages possess ﬁnite life spans. Either competitors will
eventually ﬁgure out how to reproduce the advantage or will innovate better
solutions to similar problems (Dierickx & Cool, 1989; Grant, 1996). This requires
time series data and analyses, which would bolster our investigation in expansion
strategies, and answer questions about the temporal affects of our contextual
variables. Vlﬁth this same kind of data, we are also in a position to explore
growth models, an area unexplored in the diversification literature. Individual
strategies also deserve attention. For example, extending the Barkema et al.
(1997) study assumptions about international diversiﬁcation experience is worth
exploring. In addition, the data itself suggests survival issues for lntemational
retailers. Researchers have suggested that foreign subsidiaries that diversify into
unrelated products from the MNC have lower survival rates (Li, 1995). It would
be interesting to see if this hypothesis holds in an international retailing context.

Additionally, the measurement limitation pointed out by Markides & Williamson

141

(1996) points to an opportunity to examine which resources are shared amongst
F BUs and how they relate to FBU performance. The relative recency of retail
internationalization may allow us to collect data at the FBU level to test survival
hypotheses. Finally, it would be interesting to compare our diversification-
performance results with retailers who have not internationalized. These pursuits
will signiﬁcantly contribute to the international retailing literature.

Since lntemational retail chains are a fairly recent phenomenon, more
research in this area is an important contribution to both academics and
practitioners. This study is intended to aid retailers in analyzing the impact of
strategic expansion decisions in foreign markets. We provide evidence that
refutes conventional wisdom on the relative benefits of different diversiﬁcation
strategies. We also identify how inter- and intra-segment strategies affect
performance. Deepening our knowledge about the impact of these strategies

enables managers to better evaluate market entry and expansion choices.

142

REFERENCES

Abell, D. F. (1980). Deﬁning the Business: The Starting Point of Strategic
Planning. Englewood Cliffs; Londonzz Prentice-Hall.

Agarwal, S., & Ramaswami, S. N. (1992). Choice of Foreign Market Entry Mode:
Impact of Ownership, Location and lntemalization Factors. Journal of
lntemational Business Studies, 23(1), 1-27.

Aharoni, Y. (1966). The Foreign Investment Decision Process. Boston, MA:
Harvard University.

Alexander, M, & Quinn, B. (2002). lntemational Retail Divestment. International
Journal of Retail & Distribution Management, 30(2/3), 112-125.

Amit, R., & Livnat, J. (1988). DiversiﬁcatiOn Strategies, Business Cycles and
Economic Performance. Strategic Management Journal, 9(2), 99-110.

Amit, R., & Schoemaker, P. J. H. (1993). Strategic assets and organizational
rent. Strategic Management Journal, 14(1), 33-46.

Anderson, E., & Gatignon, H. (1986). Modes of Foreign Entry: A Transaction
Cost Analysis and Propositions. Journal of International Business Studies,
17(3), 1-26.

Ansoff, H. l. (1969). Business Strategy: Selected Readings: Penguin.
Axelrod, R. (1984). The Evolution of Cooperation. New York: Basic Books.

Bane, W. T., & Neubauer, F.-F. (1981). Diversification and the Failure of New
Foreign Activities. Strategic Management Journal, 2(3), 219-233.

Barkema, H. G., Bell, J. H. J., & Pennings, J. M. (1996). Foreign Entry, Cultural
Barriers, and Learning. Strategic Management Journal, 17(2), 151-166.

Barkema, H. G., Shenkar, O., Vermeulen, F., & Bell, J. H. J. (1997). Working
Abroad, Working with Others: How Firms Learn to Operate lntemational
Joint Ventures. Academy of Management Journal, 40(2, Special Research
Forum on Alliances and Networks), 426-442.

Barney, J. (1991). Firm Resources and Sustained Competitive Advantage.
Journal of Management, 17(1), 99-120.

143

Barney, J. (2001). Is the resource-based "view" a useful perspective for strategic
management research? Yes. Academy of Management. The Academy of
Management Review, 26(1), 41.

Bartlett, C., & Ghoshal, S. (1989). Managing Across Borders. Cambridge, MA:
Harvard University Press.

Bass, F. M., Cattin, P. J., & Wrttink, D. R. (1977). Market Structure and Industry
Influence on Profitability. In H. Thorelli (Ed.),
Strategy+Structure=Performance (pp. 181-197). Bloomington, IN: Indiana
University Press.

Baum, J. A. C., Calabrese, T., & Silverman, B. S. (2000). Don't go it alone:
Alliance network composition and startups' performance in Canadian
biotechnology. Strategic Management Journal, 21(3; Special Issue:
Strategic Networks), 267-294.

Beamish, P. W., & Banks, J. C. (1987). Equity Joint Ventures and the Theory of
the Multinational Enterprise. Journal of International Business Studies,
18(2), 1-16.

Bergh, D. D. (1995). Problems with Repeated Measures Analysis: Demonstration
(with a Study of the Diversiﬁcation and Performance Relationship.
Academy of Management Journal, 38(6), 1692-1708.

Bergh, D. D., & Lawless, M. W. (1998). Portfolio restructuring and limits to
hierarchical governance: The effects of environmental uncertainty and
diversiﬁcation strategy. Organization Science, 9(1), 87-102.

Bettis, R. A., & Hall, W. K. (1982). Diversiﬁcation Strategy, Accounting
Determined Risk, and Accounting Determined Return. Academy of
Management Journal, 25(2), 254-264.

Bettis, R. A., & Prahalad, C. K (1995). The dominant logic: Retrospective and
extension. Strategic Management Journal, 16(1), 5-14.

Bhatnagar, A., & Ratchford, B. T. (2004). A model of retail format competition for
non-durable goods. lntemational Journal of Research in Marketing, 21 (1 ),
39-59.

Bianchi, C., & Arnold, S. (2004). An Institutional Perspective on Retail
Internationalization Success: Home Depot in Chile. The International
Review of Retail, Distribution and Consumer Research, 14(2), 149-169.

Bourgeois, L. J. (1980). Strategy and Environment: A Conceptual Integration.
Academy of Management Review, 5, 25-39.

144

Bowen, H. P., & Wrersema, M. F. (2005). Foreign-based competition and
corporate diversiﬁcation strategy. Strategic Management Journal, 26(12),
1153-1171.

Bowersox, D. J. (1992). Framing Global Logistics Requirements. Paper
presented at the Proceedings of the Council of Logistics Management.

Bowersox, D. J., & Calantone, R. J. (1998). Executive Insights: Global Logistics.
Journal of lntemational Marketing, 6(4), 83-93.

Brouthers, K. D. (2002). Institutional, cultural and transaction cost influences on
entry mode choice and performance. Journal of International Business
Studies, 33(2), 203-221.

Buckley, P. J., & Casson, M. (1976). The Future of the Multinational Enterprise.
London: MacMiIIan.

Burgelman, R. A., & Doz, Y. L. (2001). The Power of Strategic Integration. MIT
Sloan Management Review, 42(3), 28.

Burt, S., Dawson, J., & Sparks, L. (2003). Failure in lntemational Retailing:
Research Propositions. lntemational Review of Retail, Distribution and
Consumer Research, 13(4), 355-573.

Busija, E. C., ONeill, H. M., & Zeithaml, C. P. (1997). Diversiﬁcation strategy,
entry mode, and performance: Evidence of choice and constraints.
Strategic Management Journal, 18(4), 321 -327.

Carroll, G. R., & Swaminathan, A. (1992). The Organizational Ecology of
Strategic Groups in the American Brewing Industry from 1975 to 1990.
Industrial and Corporate Change, 1 (1 ), 65-97.

Caves, R. E. (1989). Mergers, Takeovers, and Economic Efﬁciency. lntemational
Journal of Industrial Organization, 7(1), 151.

Cavusgil, S. T. (1980). On the internationalization process of ﬁrms. European
Research, 8(November), 273-281.

Cavusgil, S. T., Ghauri, P. N., & Agarwal, M. R. (2002). Doing Business in
Emerging Markets. Thousand Oaks: Sage.

Chandler, J., Alfred D. (1990). Strategy and structure [electronic resource] :
chapters in the history of the industrial enterprise Retrieved May 10, 2007,

 

 

from
http://www.netlmrvcomlurlagi.asp?action=summary&v=18tgaokidj1 139
9.

145

Chatterjee, S. (1990). Excess Resources, Utilization Costs, and Mode of Entry.
Academy of Management Journal, 33(4), 780-800.

Chatterjee, S., & Blocher, J. D. (1992). Measurement of Firm Diversiﬁcation: Is It
Robust? Academy of Management Journal, 35(4), 874-888.

Child, P. N. (2006). Lessons From A Global Retailer: An Interview with the
President of Carrefour China. McKinsey Quarterly(Special Edition: Serving
the New Chinese Consumer).

Christensen, H. K., & Montgomery, C. A. (1981). Corporate Economic
Performance: Diversiﬁcation Strategy Versus Market Stmcture. Strategic
Management Journal, 2(4), 327.

Coe, N. M. (2004). The lnternationalisation/Globalisation of Retailing: Towards
an Economic - Geographical Research Agenda. Environment and
Planning A(36), 1571 -1 594.

Cohen, W. M., & Levinthal, D. A. (1990). Absorptive Capacity: A New
Perspective on Learning and Innovation. Administrative Science Quarterly,
35(1), 128-152.

Colla, E. (2004). The Outlook for European Grocery Retailing: Competition and
Format Development. The lntemational Review of Retail, Distribution and
Consumer Research, 14(1), 47-69.

Cortsjens, J., & Cortsjens, M. (1995). Store Wars: The Battle for Mindspace and
Shelfspace. New York: John VWey & Sons.

Cortstjens, J., Corstjens, M., & Lal, R. (1995). Retail Competition in the Fast-
Moving Consumer Goods Industry: The Case of France and the UK.
European Management Journal, 13(4), 363-373.

Cotterill, R. W., & Haller, L. E. (1992). Barrier and Queue Effects: A Study of
Leading Us Supermarket Chain Entry Patterns. The Journal of Industrial
Economics, 40(4), 427-440.

Crossan, M. M., & Berdrow, I. (2003). Organizational Learning and Strategic
Renewal. Strategic Management Journal, 24(11), 1087-1105.

Daft, R. L. (1982). Bureaucratic versus Nonbureaucratic Structure and the
Process of Innovation and Change. In S. B. Bacharach (Ed), Research in
the Sociology of Organizations: Theory and Research. Greenwich, CT: JAI
Press.

146

Datta, D. K., Rajagopalan, N., & Rasheed, A. M. A. (1991). Diversification and
Performance: Critical Review and Future Directions. Journal of
Management Studies, 28(5), 529-558.

Davis, P. 6., Robinson, J., Richard B., Pearce II, J. A., & Park, S. H. (1992).
Business Unit Relatedness and Performance: A Look at the Pulp and
Paper Industry. Strategic Management Journal, 13(5), 349-361.

Day, G. S. (1994). The capabilities of market-driven organizations. Journal of
Marketing, 58(4), 37-52.

Delaney, R. V., & Wilson, R. (2002). 14th Annual State of Logistics Report: The
Case for Reconﬁguration. Washington, DC: Cass Information Systems
and ProLogis.

Delios, A., & Beamish, P. W. (2001). Survival profitability: The roles of
experience and intangible assets in foreign subsidiary performance.
Academy of Management Journal, 44(5), 1028.

Dierickx, l., & Cool, K. (1989). Asset Stock Accumulation And Sustainability Of
Competitive. Management Science, 35(12), 1504.

Doherty, A. M. (1999). Explaining lntemational Retailers' Market Entry Mode
Strategy: lntemalization Theory, Agency Theory, and the Importance of
Information Asymmetry. lntemational Review of Retail, Distribution and
Consumer Research, 9(4), 379-402.

' Dranova, D., Peterof, M., & Shanley, M. (1998). Do Strategic Groups Exist? An
Economic Framework For Analysis. Strategic Management Journal, 19,
1029-1044.

Duncan, C., Jones, K., & Moon, G. (1998). Context, Composition and
Heterogeneity: Using Multilevel Models in Health Research. Social
Science Medicine, 46(1), 97-117.

Dundas, K. N. M., & Richardson, P. R. (1982). Implementing the Unrelated
Product Strategy. Strategic Management Journal, 3(4), 287-301.

Dunning, J. H. (1980). Toward an Eclectic Theory of International Production:
Some Empirical Tests. Journal of International Business Studies, 11 (1 ), 9-
31.

Dunning, J. H. (1988). The Eclectic Paradigm of International Production: A

Restatement and Some Possible Extensions. Journal of lntemational
Business Studies, 19(1), 1-31.

147

Dunning, J. H. (1995). Reappraising the eclectic paradigm in an age of alliance
capitalism. Journal of International Business Studies, 26(3), 461 -491.

Dussauge, P., Garrette, B., & Mitchell, W. (2000). Learning from competing
partners: Outcomes and durations of scale and link alliances in Europe,
North America and Asia. Strategic Management Journal, 21(2), 99-126.

Eisenhardt, K. M. (1989). Agency Theory: An Assessment And Review. Academy
of Management. The Academy of Management Review, 14(1), 57.

Eisenhardt, K. M., & Martin, J. A. (2000). Dynamic capabilities: What are they?
Strategic Management Journal, 21(10/11), 1105-1121.

Ekeledo, l., & Sivakumar, K. (1998). Foreign market entry mode choice of service
firms: A contingency perspective. Academy of Marketing Science. Journal,
26(4), 274-292.

Elshof, P. (2005, February 2005). Ahold Company Proﬁle.

Erramilli, M. K., & Rao, C. P. (1993). Service ﬁrms' international entry-mode
choice: A modiﬁed transaction-cost analysis approach. Journal of
Marketing, 57(3), 19-38.

Farjoun, M. (1994). Beyond industry boundaries: Human expertise, diversiﬁcation
and resource-related industry groups. Organization Science, 5(2), 185.

Farjoun, M. (2002). Towards an organic perspective on strategy. Strategic
Management Journal, 23(7), 561 -594.

Fowler, K. L., 8. Schmidt, D. R. (1989). Determinants Of Tender Offer Post-
Acquisition Financial Per. Strategic Management Journal, 10(4), 339.

Fuentelsaz, L., & Gomez, J. (2006). Multipoint Competition, Strategic Similarity
and Entry Into Geographic Markets. Strategic Management Journal, 27,
477-499.

Gary, M. S. (2005). Implementation strategy and performance outcomes in
related diversification. Strategic Management Journal, 26(7), 643.

Geringer, J. M., Beamish, P. W., & daCosta, R. C. (1989). Diversification
Strategy and Internationalization: Implications for MNE Performance.
Strategic Management Journal, 10(2), 109-119.

Ghoshal, S. (1987). Global Strategy: An Organizing Framework. Strategic
Management Journal, 8(5), 425-440.

148

Ginsberg, A., & Venkatraman, N. (1985). Contingency Perspectives of
Organizational Strategy: A Critical Review of the Empirical Research.
Academy of Management Review, 10(3), 421 -434.

Granovetter, M. (1985). Economic Action and Social Structure: The Problem of
Embeddedness. American Journal of Sociology, 91 (November), 481 -51 0.

Grant, R. M. (1996). Toward a knowledge-based theory of the ﬁrm. Strategic
Management Journal, 17, 109.

Greve, H. R. (1996). Patterns of competition: The diffusion of a market position in
radio broadcasting. Administrative Science Quarterly, 41(1), 2960.

Gupta, A. K., & Govindarajan, V. (1986). Resource Sharing among SBUs:
Strategic Antecedents and Administrative Implications. Academy of
Management Journal, 29(4), 695-714.

Hahn, E. D., & Doh, J. P. (2006). Using Bayesian Methods in Strategy Research:
An Extension of Hansen et al. Strategic Management Journal, 27, 783-
798. ‘

Hamel, G., & Prahalad, C. K. (1985). Do You Really Have a Global Strategy?
Harvard Business Review, 63(4), 133-145.

Hamel, G., & Prahalad, C. K. (1990). The Core Competencies of the Corporation.
Harvard Business Review, 68(3), 79-91.

Hannan, M., & Freeman, J. (1987). The Ecology of Organizational Founding:
American Labor Unions, 1836-1985. American Journal of Sociology,
92(4), 910-943.

Henisz, W. J., & Delios, A. (2001). Uncertainty, imitation, and plant location:
Japanese multinational corporations, 1990-1996. Administrative Science
Quarteriy, 46(3), 443.

Hennart, J.-F. (1982). A Theory of the Multinational Enterprise. Ann Arbor, MI:
University of Michigan Press.

Hennart, J.-F. (1988). A Transaction Costs Theory of Equity Joint Ventures.
Strategic Management Journal, 9(4), 361 -374.

Herman, E., & Lowenstein, L. (1988). The Efficiency Effects of Hostile

Takeovers. In J. C. J. L. a. Ackerman (Ed.), Knights, Raiders and targets:
The Impact of The Hostile Takeover. New York: Oxford University Press.

149

Heskett, J., Glaskowsky, N., & lvie, R. (1973). Business Logistics. New York: The
Ronald Press Company.

High, R. (2005). Data Analysis Issues: Binary Data and the Arcsin
Transformation. Retrieved July 7, 2007

Hitt, M. A., Hoskisson, R. E., & Kim, H. (1997). lntemational Diversiﬁcation:
Effects on Innovation and Firm Performance in Product-Diversiﬁed Firms.
Academy of Management Journal, 40(4), 767-798.

Holm, U., & Sharma, D. D. (2006). Subsidiary marketing knowledge and strategic
development of the multinational corporation. Journal of International
Management, 12(1), 47-66.

Hoskisson, R. E., & Hitt, M. (1990). Antecedents and Performance Outcomes of
Diversiﬁcation: A Review and Critique of Theoretical Perspectives. Journal
of Management, 16(461-509). '

Hult, G. T., David J. Ketchen, J., Grifﬁth, D. A., Finnegan, C. A., Gonzalez-
Padron, T. L., Harmancioglu, F. N., et al. (2007). An Assessment of Cross-
Cultural Data Equivalence in lntemational Business (2nd Review). Journal
of International Business Studies.

Hymer, S. (1976). The international operations of national ﬁrms: a study of direct
foreign investment. Cambridge, MA: MIT Press.

Jacobson, R., & Aacker, D. A. (1985). Is Market Share All that It's Cracked Up to
Be? Journal of Marketing, 49(Fall), 11-22.

Jacquemin, A. P., & Berry, C. H. (1979). Entropy Measure of Diversiﬁcation and
Corporate Growth. The Journal of Industrial Economics, 27(4), 359.

Jahera, J. 8., Lloyd, W. P., & Page, D. E. (1987). Firm Diversification and
Financial Performance. Quarterly Review of Economices and Business,
27(1), 51-62.

Jensen, & Meckling. (1976). Theory of the Firm: Managerial Behavior, Agency
Costs and Ownership Structure. Journal of Financial Economics,
3(October), 305-360.

Johanson, J. A. N., & Vahlne, J.-E. (1977). The Internationalization Process of
the Firm - a Model of Knowledge Development and Increasing Foreign
Market Commitments. Journal of International Business Studies, 8(1), 23-
32.

150

Johansson, J. A. N., & Vahlne, J.-E. (1977). The lntemationalization Process of
the Firm - a Model of Knowledge Development and Increasing Foreign
Market Commitments. Journal of International Business Studies, 8(1), 23-
32.

Jones, G. R., & Hill, C. W. L. (1988). Transaction Cost Analysis of Strategy
Structure Choice Strategic Management Journal, 9(2), 159.

Kay, N. (1982). The Evolving Firm: Strategy and Structure in Industrial
Organisation. London: Macmillan.

Kazanjian, R. K., & Drazin, R. (1987). Implementing Internal Diversification:
Contingency Factors for Organization Design Choices. Academy of
Management. The Academy of Management Review, 12(2), 342.

Kim, W. C., & Hwang, P. (1992). Global Strategy and Multinationals' Entry Mode
Choice. Journal of International Business Studies, 23(1), 29-53.

Kim, W. C., Hwang, P., & Burgers, W. P. (1993). Multinationals' diversiﬁcation
and the risk-return trade-off. Strategic Management Journal, 14(4), 275-
286.

Klein, K. J., Dansereau, F., & Hall, R. J. (1994). Levels issues in theory
development, data collection, and analysis. Academy of Management
Review, 19(2), 195-229.

Kobrin, S. (1982). Managing Political Risk Assessment. Berkeley, CA: University
of California Press. '

Kogut, B. (1983). Foreign Direct Investment as a Sequential Process. In C. P. K.
a. D. B. Audretsch (Ed), The Multinational Corporation in the 1980s (pp.
38-56). Cambridge, MA: MIT Press.

Kogut, B. (1985). Designing Global Strategies: Proﬁting from Operational
Flexibility. Sloan Management Review, 27, 27-38.

Kogut, B., & Singh, H. (1988). The Effect of National Culture on the Choice of
Entry Mode. Journal of International Business Studies, 19(3), 411-432.

Kotler, P. (1997). Marketing Management: Analysis, Planning, Implementation
and Control (9th ed.). Upper Saddle River, NJ: Prentice Hall.

Lamont, B. T., & Anderson, C. R. (1985). Mode of Corporate Diversification and

Economic Performance. Academy of Management Journal, 28(4), 926-
934.

151

Lee, V., & Bryk, A. S. (1989). A Multilevel Model of the Social Distribution of
Educational Achievement. Sociology of Education, 62, 172-192.

Lenway, S. A., & Murtha, T. P. (1994). The state as strategist in international
business research. Journal of International Business Studies, 25(3), 513-
535.

Leontiades, M. (1986). The Rewards of Diversifying into Unrelated Businesses.
The Journal of Business Strategy, 6(4), 81.

Levitas, E., & Chi, T. (2002). Rethinking Rouse and Daellenbach's rethinking:
isolating vs. testing for sources of sustainable competitive advantage.
Strategic Management Journal, 23(10), 957-962.

Levy, M., & Weitz, B. (2007). Retailing Management (6th ed.). New York:
McGraw-HiIl/lrwin.

Li, J. (1995). Foreign entry and survival: Effects of strategic choices on
performance in lntemational markets. Strategic Management Journal,
16(5), 333-351.

Li, J., & Guisinger, S. (1991). Comparative Business Failures of Foreign-
Controlled Firms in the United States. Journal of International Business
Studies, 22(2), 209.

Li, S. X., & Greenwood, R. (2004). The Effect of Within-Industry Diversiﬁcation on
Firm Performance: Synergy Creation, Multi-Market Contact and Market
Structuration. Strategic Management Journal, 25(12), 1131.

Lord, M. D., & Ranft, A. L. (2000). Organizational Learning about New
lntemational Markets: Exploring the Internal Transfer of Local Market
Knowledge. Journal of International Business Studies, 31(4), 573-589.

Lu, J. W. (2002). Intra- and inter-organizational imitative behavior: Institutional
inﬂuences on Japanese ﬁrms' entry mode choice. Journal of lntemational
Business Studies, 33(1), 19-37.

Lubatkin, M., & O'Neill, H. M. (1987). Merger Strategies and Capital Market Risk.
Academy of Management Journal, 30(4), 665-684.

Luke, D. A. (2004). Multilevel Modeling (Vol. 143). Thousand Oaks,CA: Sage.

Lusch, R. F., Vargo, S. L., & O'Brien, M. (2007). Competing through service:
Insights from service-dominant logic. Journal of Retailing, 83(1), 5.

152

Madhok, A. (1997). Cost, value and foreign market entry mode: The transaction
and the ﬁrm. Strategic Management Journal, 18(1), 39-61.

Madhok, A., & Liu, C. (2006). A coevolutionary theory of the multinational ﬁrm.
Journal of International Management, 12(1), 1-21.

Makadok, R. (2001). Toward a synthesis of the resource-based and dynamic-
capability views of rent creation. Strategic Management Journal, 22(5),
387-401.

Makino, S., Isobe, T., & Chan, C. M. (2004). Does Country Matter? Strategic
Management Journal, 25(10), 1027-1043.

March, J. G. (1991). Exploration and Exploitation in Organizational Learning.
Organization Science, 2(1), 71 -87.

Markides, C. C. (1992). Consequences of Corporate Refocusing: Ex Ante
Evidence. Academy of Management Journal, 35(2), 398-412.

Markides, C. C., & Williamson, P. J. (1994). Related diversification, core
competences and corporate performance. Strategic Management Journal,
15(Special Issue), 149-165.

Markides, C. C., & Williamson, P. J. (1996). Corporate Diversification and
Organizational Structure: A Resource-Based View. Academy of
Management Journal, 39(2), 340-367.

Mascarenhas, ‘B. (1992). Order of Entry and Performance in lntemational
Markets. Strategic Management Journal, 13(7), 499-510.

McComb, S. A., Barringer, M. W., & Boume, K. A. (2004). Understanding Part-
Time Work Arrangements: A Multi-Level Perspective. In F. J. Yammarino
& F. Dausereau (Eds), Multi-Level Issues in Organizational Behavior and
Processes (Vol. 3, pp. 3-55). New York: Elsevier.

McGahan, A. M., & Porter, M. E. (1997). How much does industry matter, really?
Strategic Management Journal, 18(Special Issue), 15-30.

Milgrom, P., & Roberts, J. (1995). Complementarities and fit: Strategy, structure,
and organizational change in manufacturing. Journal of Accounting &
Economics, 19(2,3), 179.

Miller, D. J. (2004). Firms' technological resources and the perforrnanoe effects

of diversiﬁcation: a longitudinal study. Strategic Management Journal,
25(11),1097-1119.

153

Miller, D. J. (2006). Technological diversity, related diversiﬁcation and ﬁrm
performance. Strategic Management Journal, 27(7), 601.

Mintzberg, H., Ahlstrand, B., & Lampel, L. (1998). Strategy Safari: A Guided Tour
on the Wilds of Strategic Management. New York: Free Press.

Misangyi, V. F., Elms, H., Greckhamer, T., & Lepine, J. A. (2006). A new
perspective on a fundamental debate: a multilevel approach to industry,
corporate, and business unit effects. Strategic Management Journal,
27(6), 571-590.

Mishra, D. P., Heide, J. B., & Cort, S. G. (1998). Information asymmetry and
levels of agency relationships. JMR, Jouma/ of Marketing Research,
35(3), 277.

Montgomery, C. A., & Wemerfelt, B. (1988). Diversiﬁcation, Ricardian Rents, and
Tobin's q. The Rand Journal of Economics, 19(4), 623.

Morck, R., Schleifer, A., & Vishny, R. (1990). Do Managerial Motives Drive Bad
Acquistions? Journal of Finance, XL V(1 ), 31 -48.

Morgan, N. A., Zou, S., Vorhies, D. W., & Katsikeas, C. S. (2003). Experiential
and Informational Knowledge, Architectural Marketing Capabilities, and
the Adaptive Performance of Export Ventures: A Cross-National Study.
Decision Sciences, 34(2), 287-321.

Mowery, D. C., Oxley, J. E., & Silverman, B. S. (1996). Strategic alliances and
interﬁnn knowledge transfer. Strategic Management Journal, 77-91.

Nayyar, P. R. (1992). On the Measurement of Corporate Diversiﬁcation Strategy:
Evidence from Large US. Service Firms. Strategic Management Journal,
13(3), 219-235.

Nayyar, P. R. (1993a). Performance Effects of Infom'iation Asymmetry and
Economies of Scope in Diversiﬁed Service Firms. Academy of
Management Journal, 36(1), 28-57.

Nayyar, P. R. (1993b). Stock market reactions to related diversiﬁcation moves by
service ﬁrms seeking benefits from information asymmetry and economies
of scope. Strategic Management Journal, 14(8), 569-591.

Nelson, R., & Winter, S. (1982). An Evolutionary Theory of Economic Change.
Cambridge, MA: Harvard University Press.

Palepu, K. (1985). Diversiﬁcation Strategy, Proﬁt Performance and the Entropy
Measure. Strategic Management Journal, 6(3), 239-255.

154

Palich, L. E., Cardinal, L. 8., & Miller, C. C. (2000). Curvilinearity in the
diversiﬁcation-performance linkage: An examination of over three decades
of research. Strategic Management Journal, 21 (2), 155-174.

Palmer, M. (2006). lntemational retail joint venture Ieaming. Service Industries
Journal, 26(2), 165-187.

Palmer, M., 8 Owens, M. (2006). New directions for international retail joint
venture research. The International Review of Retail, Distribution and
Consumer Research, 16(2), 159.

Pelligrini. (1994). Alternatives for growth and lntemationalization in retailing.
lntemational Review of Retail, Distribution and Consumer Research, 4(2),
121-148.

Penrose, E. T. (1959). The Theory of the Growth of the Firm. London: Basil
Blackwell.

Penrose, E. T. (1980). The Theory of the Growth of the Firm (2nd ed.). Oxford:
Oxford University Press.

Peteraf, M. A. (1993). The Comerstones of Competitive Advantage: A Resource-
Based View. Strategic Management Journal, 14(3), 179-191.

Pfeffer, J., & Salancik, G. R. (1978). The External Control of Organizations: A
Resource Dependence Perspective. New York: Harper and Row.

Picot-Coupey, K. (2006). Determinants of International Retail Operation Mode
Choice: Towards a Conceptual Framework Based Upon Evidence From
French Specialised Retail Chains. International Review of Retail,
Distribution and Consumer Research, 16(2), 215-237.

Porter, M. E. (1985). Competitive Advantage: Creating and Sustaining Superior
Performance. New York: The Free Press.

Porter, M. E. (1987). From Competitive Advantage to Corporate Strategy.
Cambridge, MA: Harvard Business School Press.

Porter, M. E. (1990). Competitive Advantage of Nations. Cambridge, MA:
Harvard University Press.

Prahalad, C., & Doz, Y. (1987). The Multinational Mission. New York: Free Press.
Ramanujam, V., 8 Varadarajan, P. (1989). Research on Corporate

Diversiﬁcation: A Synthesis. Strategic Management Journal, 10(6), 523-
551.

155

Raudenbush, S. W. (1993). A Crossed Random Effects Model for Unbalanced
Data With Applications in Cross-Sectional and Longitudinal Research.
Journal of Educational Statistics, 18(4), 321-349.

Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical Linear Models:
Applications and Data Analysis Methods (2nd ed.). Thousand Oaks: Sage.

Ravenscraft, D. J., & Scherer, F. M. (1987). Life After Takeover. The Journal of
Industrial Economics, 36(2), 147.

Reuer, J. J., & Leiblein, M. J. (2000). Downside risk implications of
multinationality and international joint ventures. Academy of Management
Journal, 43(2), 203-214.

Ricart, J. E., Enright, M. J., Ghemawat, P., Hart, S. L., & Khanna, T. (2004). New
frontiers in international strategy. Journal of International Business
Studies, 35(3), 175-200.

Robins, J., & Wiersema, M. F. (1995). A resource-based approach to the
multibusiness ﬁrm: Empirical analysis of portfolio interrelationships and
corporate financial performance. Strategic Management Journal, 16(4),
277-299.

Rodrigues, A. M., Bowersox, D. J., & Calantone, R. J. (2005). Estimation of
Global and National Logistics Expenditures: 2002 Data Update. Journal of
Business Logistics, 26(2).

Root, F. R. (1994). Entry Strategies for International Markets. London: Lexington
Books.

Rouse, M. J., & Daellenbach, U. S. (2002). More thinking on research methods
for the resource-based perspective. Strategic Management Journal,
23(10), 963-967.

Rugman, A. M. (1981). Inside the Multinationals. New York: Columbia University
Press.

Rugman, A. M. (2000). The End of Globalization. New York: McGraw-Hill.

Rugman, A. M., & Verbeke, A. (2004). A perspective on regional and global
strategies of multinational enterprises. Journal of lntemational Business
Studies, 35(1), 3-18.

Rugman, A. M., & Verbeke, A. (2005). Towards a Theory of Regional

Multinationals: A Transactions Cost Approach. Management International
Review, 45(1), 5-17.

156

Rumelt, R. P. (1974). Strategy, structure, and economic performance.
Cambridge, MA: Harvard University Press.

Rumelt, R. P. (1982). Diversiﬁcation Strategy and Proﬁtability. Strategic
Management Journal, 3(4), 359-369.

Rumelt, R. P. (1986). Strategy, Structure, and Economic Performance (revised
ed.). Boston, MA: Harvard Business School Press.

Rumelt, R. P. (1991). How Much Does Industry Matter? Strategic Management
Journal, 12(3), 167-185.

Salmon, W., & Tordjman, A. (1989). The Internationalization of Retailing.
lntemational Journal of Retailing, 4(2), 3-16.

Schmalensee, R. (1985). Do Markets Differ Much? The American Economic
Review, 75(3), 341 -351.

Schwartz, C. J. (1991). Hitting the Target: Effective Techniques for Market Share
and Response Rate Modeling Using Demographics [Electronic Version].
Quirk’s Marketing Research Review.

Scott, R., & Meyer, J. (1983). Organizational Environments: thual and
Rationality. Thousand Oaks: Sage.

Sharma, A. (1998). Mode of entry and ex-post performance. Strategic
Management Journal, 19(9), 879-900.

Simmonds, P. G. (1990). The Combined Diversiﬁcation Breadth and Mode
Dimensions and the Performance of Large Diversified Firms. Strategic
Management Journal, 11(5), 399-410.

Simon, H. A. (1991). Organizations and Markets. Journal of Economic
Perspectives, 5(2), 25-44.

Slater, S. F ., & Narver, J. C. (1995). Market orientation and the Ieaming
organization. Journal of Marketing, 59(3), 63-74.

Snow, C. C. & Hambrick, D. C. (1980). Measureing Organizational Strategies:
Some Theoretical and Methodological Problems. Academy of
Management Review, 5, 527-538.

Sorenson, O., & Sorenson, J. B. (2001). Finding the right mix: Franchising,

organizational learning, and chain performance. Strategic Management
Journal, 22(6/7), 713-724.

157

Stemquist, B. (1997a). lntemational expansion of US retailers. lntemational
Journal of Retail & Distribution Management, 25(8), 262.

Stemquist, B. (1997b). lntemationalization of Japanese Department Stores.
lntemational Journal of Commerce and Management, 7(1), 57-73.

Stemquist, B. (2007). International Retailing (2nd ed.). New York: Fairchild
Publications.

Szulanski, G. (1996). Exploring internal stickness: lmpediments to the transfer of
best practice within the ﬁrm. Strategic Management Journal, 27-43.

Szymanski, D. M., Bharadwaj, S. G., & Varadarajan, P. R. (1993). An analysis of
the market share-profitability relationship. Journal of Marketing, 57(3), 1-
18.

Szymanski, D. M., Bharadwaj, S. G., & Varadarajan, R. (1993). An analysis of
the market share-proﬁtability relationship. Journal of Marketing, 57(3), 1-
18.

Tallman, S., & Li, J. (1996). Effects of International Diversity and Product
Diversity on the Performance of Multinational Firms. Academy of
Management Journal, 39(1), 179-196.

Tanriverdi, H., & Venkatraman, N. (2005). Knowledge Relatedness and
Performance of Multibusiness Firms. Strategic Management Journal,
26(2), 97-1 19.

Teece, D. J. (1980). The Diffusion of an Administrative Innovation. Management
Science, 26(5), 464-470.

Teece, D. J. (1982). Towards an Economic Theory of the Multiproduct Firm.
Journal of Economic Behavior & Organization, 3, 39-63.

Teece, D. J. (1983). Technological and Organizational Factors in the Theory of
Multinational Enterprise. In M. Casson (Ed.), Growth of International
Business. London: Allen and Unwin.

Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities and strategic
management. Strategic Management Journal, 18(7), 509-533.

Tohmatsu, D. T. (2006). 2006 Global Powers of Retailing: Deloitte Touche
Tohmatsu

Toyne, B., & Nigh, D. (Eds). (1997). International Business: An Emerging Vision.
Columbia, SC: University of South Carolina Press.

158

Trautwein, F. (1990). Merger Motives and Merger Prescriptions. Strategic
Management Journal, 11(4), 283-295.

Varadarajan, R., & Ramanujam, V. (1987). Diversification and Performance: A
Reexamination Using a New Two-Dimensional Conceptualization of
Diversity in F inns. Academy of Management Journal, 30(2), 380-393.

Vargo, S. L., & Lusch, R. F. (2004a). Evolving to a New Dominant Logic for
Marketing. Journal of Marketing, 68(1), 1.

Vargo, S. L., & Lusch, R. F. (2004b). The Four Service Marketing Myths:
Remnants of a Goods-Based, Manufacturing Model. Journal of Service
Research : JSR, 6(4), 324.

Ramanujam, P. V. (1989). Research on corporate diversiﬁcation: A synthesis.
Strategic Management Journal, 10(6), 523-551.

Wal-Mart, C. (2006). Annual Report. Bentonville, AR.

Wemerfelt, B. (1984). A Resource-Based View of the Firm. Strategic
Management Journal, 5(2), 171 -1 80.

Wemerfelt, B. (1995). The resource-based view of the ﬁrm: Ten years after.
Strategic Management Journal, 16(3), 171-174.

Williamson, 0. E. (1985). The Economic Institutions of Capitalism: Firms,
Markets, Relational Contracting New York: The Free Press.

Williamson, 0. E. (1996). Economic organization: The case for candor. Academy
of Management Review, 21 (1 ), 48-57.

Woodcock, C. P., Beamish, P. W., & Makino, S. (1994). Ownership-based entry
mode strategies and international performance. Journal of lntemational
Business Studies, 25(2), 253-273.

Wooldridge, J. M. (2003). Introductory Econometrics: A Modern Approach (2e
ed.) Mason, OH: South-Western.

Wrigley, N. (2000). Strategic market behaviour in the internationalization of food
retailing - Interpreting the third wave of Sainsbury's US diversification.
European Journal of Marketing, 34(8), 891.

Wrigley, N., & Currah, A. (2003). The Stresses of Retail lntemationalization:

Lessons for Royal Ahold's Experience in Latin America. lntemational
Review of Retail, Distribution and Consumer Research, 13(3), 221 -243.

159

Yammarino, F. J., & Dausereau, F. (Eds.). (2004). Multi-Level Issues in
Organizational Behavior and Processes (Vol. 3). New York: Elsevier.

Zaccarin, S., & Rivellini, G. (2002). Multilevel Analysis in Social Research: An
Application of a Cross-Classiﬁed Model. Statistical Methods &
Applications, 1 1(95-108).

Zaheer, S. (1995). Overcoming the liability of foreignness. Academy of
Management Journal, 38(2), 341-363.

Zeithaml, V. A., Parasuraman, A., & Berry, L. L. (1985). Problems and Strategies
in Services Marketing. Journal of Marketing, 49(2), 33-46.

160

APPENDIX A: HISTOGRAMS, DESCRIPTIVE STATISTICS,
AND BIVARIATE CORRELATIONS

Level One Continuous Variables

Hypothesis 1

 

Frequency

Mean =0.2577
Std. Dev. =0.3820
N =197

 

 

 

. ,. X
10000 1.5000 20000
Total Diversiﬁcation

161

Frequency

 

S

 

 

2.00 4.00 6-00
In country outlets (log)

162

Frequency

 

 

-200 0.00 200
Retail Market Share 2003 (log)

163

 

 

ii

Frequency

a

 

 

 

 

 

 

 

 

  
  
     

.200 0.00 2.00
Retail Market Share 2004 (log)

  

Mean =-0.03
Std. Dev. =1.478
N =184

Hypotheses 2 8i 3

 

  
  
  

 

 

4o—
30- ..
>
0
c
0
3
m—
2
H.
10—.
Mean =1.91
Std. Dev. =o.9o7
, , N=342
o_ . . . .. . . .
0.00 1 .00 2.00 3.00 4.00 5.00

In country experience (I09)

165

 

 

 

81” 10.00

6.

00

.

4

2

ln-Country Outlets 2002 (I09)

0.00

166

Frequency

 

 

.500 -250 0.00
Retail Market Share 2003 (I09)

 

167

Mean =-0.76
Std. Dev. =1.673
N P351

 

5°—

8
I .

Frequency

w
‘?

 

-5.oo

-2.so 0.00
Market Share 2004 (lo )

168

 

Mean 8089

Std. Dev. =1.638
33

N=3

Frequency

Histograms of Corporate Level Continuous Variables

 

 

 

 

100 200
Retail Operations Experience (Years)

169

Mean =66.53
Std. Dev. =39.698
N =34

 

Frequency

 

10.00
Thousands of Employees Worldwide (log)

170

 

 

A
I

Frequency
‘1’

 

 

 

Mean =7.21
Std. Dev. =1.
N =34

5.00 6.00 7.00 3.00 9.00
Thousands of Outlets Worldwide (Log)

171

Frequency

 

 

 

1o 20
lntemational Diversiﬁcation

172

Histograms of Country Level Continuous Variables

 

    
   

on
I

   

  

Frequency
‘i

Mean =3.80
Std. Dev. =1.418
N =49

 

 

4.00 6.00
Industry Concentration (Square root)

173

Frequency

 

 

 

10.00 12.50 15.00 17.50

Retail Labor Market Size
(Log of thousands of retail employees)

5.00 7.50

174

 

 

 

 

14.00

12.00
il Market Size

(Log of thousands of outlets)

.00

10

8.00

Reta

175

 

 

 

 

 

 

 

m8 805% z 23>
8m. 2: at. m8- SE: 88.- 8a 8.9 x: 335:.
men. 0:. mt. 89. 8th 82.- 5a was- R: 8:55
own. 8? 4:. 8a.- 93m. 28; m3. 8. Sr 8.8on
new. vmor Q. r. 5 _.. Somme mmomd and 8. BF «050:5
men. 83.3 wt. 53 98m. 33. EN 8. Be cameos
m3”. 2 E at. 93 Sam. 88. m3 8.- 3F 86%
men. 8 2. mt. $3 88 an. ERR. ES; 88. Be 85
mvm. 83- mi. mm? «8.9 2.8 mm F B? 2.30
men. 23- mt. Sm. 42.3 8.8 one m Be 298
55 .5 025m Bum .3 25sz 0:285 27.:ma egg 2.me 22me

"38:3. mmoEsoxw scant/mo cups. 825me SEES—2 z
Em

 

 

 

 

 

 

 

 

 

_. £35062.— .0. noses—3m 0593009 up .05..

176

 

 

 

 

 

 

8..-.8...
a...” z 3.3
08. .8.. 8.. 8.. .- 8.8.. 88.- 0.8 8..- m8 825...
8m. 8.. . 8 .. 8o. .- .88.. N8..- 88 8.8- .8 8.0.5...
.8. 8m.- 8.. 88. 88... 8.8.8 8.8 8. 88 8.8...
88. 8..- 8.. 8..- 8.8. 0.8.. 8... 8. m8 888.
8w. 8..» 8.. .8.» «8. .o. . o .8 .588.
8N. ..N.» 8.. 8..» .8. 8. . o .8 282.8
.8. 8o. 8.. 8.... 8.. .m. . o .8 25.88
8w. 8..- 8.. .N... .8. 8. . o .8 .8...
8m. 48. 8.. .8.. m8. 8. . o .8 >..
8.. .... .- 8.. 84. .8. 8. . o .8 8..
8w. ..3 8.. 8.8 .8. 8. . o .8 88.8..
8m. .8. .- 8 .. ..N. 8... 8. . o .8 8;
8m. .88 8.. 8.8 .8. 4.. . o .8 09.2:
8.. 8.8.8 8.. .8... 8.. 8. . o .8 >42:
8m. 88.... 8.. 8... 8m. 8. . o .8 002.2:
8.. m8.- 8.. .8.. .8. 8. . o .8 005.
8... 08. . 8 .. 88. . 88. m .. . o .8 >2
8m. .8. .- 8 .. .8. 8... .8. . o .8 00.50
8.. 80... 8.. .88 88.8. 8.8.. 88.. 88. .8 88.2
98. o8. .- 8.. 8...- 8.88. .808. 88.. 88. .8 .8558
.8. .8. .- 8.. .8. .8.... 8.8 8 . .8 2.8
.8. 8... .- 8 .. 8m. 82... 8.8 8. . .8 298

.em. .90 8.8.8.0 .em. .20 8890 08.8.0 0.7.8.0 08.8.0 08.8.0 08.8.0

0.02.3. 000505 520300 50.2 8:60.05. €38.55. 2
.20

 

 

 

 

 

 

 

 

 

n a N 00005002.. .5. 0050305 02.3.0000 a. .05..

177

 

 

 

8 8.28... z 20>
800 .0. .88 0.0 0... 8 0:858
808.. 08.... .00. 8.0 8 83:50:.
888.. .003. .05. 8.0 0.. 8.0.2:.
.00. 8. . o 8 .058.
80. 00. . o 8 8.050
88.0. 08.0. 0.8 m. 0.. 880:8
. .00. 00.0w 8 . 8 0.0.8
8.6300 .06 E005. 83:5me E2555.

 

 

 

 

 

 

 

8.8080 9.3.880

178

III:

 

 

 

 

 

.0 $0255 2 20>
m .09. . mm .N... .00 o: vm «98...
003.00.. «N000. 3.3 000 V0 $7....
80800 «mam. .0 v0.0.9 mm. vm 235.63
0508.. 0000. mm. .N him- vm 22.9%
vow. om. . o vm 63:00
08.8 $.08. 08. 0N0. 3 2:5
8.. .0 $0. om. 0 V0 2&8
c2550 .05 c022 E:E_xms_ E:E_c_s_

 

 

 

 

 

mozmwaum 023E500

179

 

a». .8cm_.waxm

 

 

 

 

 

 

 

 

 

 

 

0
mucus. 0035. EcozaEmE.
an a a a a a . . a 8% 52.350820
0.3... 08.00. 8003. 02.8.. 0800... :38. ”200.0 . “3&3. 00...». 0222.5 cams.
3%. 62.85295
0222::
f 62.0.8qu
0 O . < .
09.03. .mnovo. a 2.53 3.5.682...
“0 an. L. a a m . . a 2..
82.0.. E8... .33.. $09.... 838.. 0.8%.. a .880 . “9.008. 80000.. 582 8:80.995
:0. 52.35825
3.2....
a NO .
«8000. a. 3.0
a a a a . . . 8.. 52.80.9020
828. .NNNS. 0.38. .0230 503.8 “0.003 38.652...
a” an no a w m
800$ $0.00 .033 0.4.380 85:0. 3002... .088..- k000i..- “~38? .3300..- 8.. .220 :85.
«scam 03.“.
a .25.... 0 .285. . .822 0 .26.... 0 .28.... v .82... a .822 u .822 . .25.... “mm“.

 

 

 

 

 

 

 

 

 

 

 

 

309cm .05.. 380950 "N... 030...

mPADmmm ZO_._.<O.n=wmm>_n_ .02-$022009” ”m x_ozmmn_<

180

nova .0 E00580 0. _0uoE 03030.0
0. 50:00.08 .0022 nova .0 €005.56 0_ Eco... 03030.0 9 50:00.08 .8220 .o..v0 .0 E00585. .mo.va .0 .c0o....cm_m0

 

 

 

 

 

 

 

 

 

 

 

 

 

0. 0. 0 0 0 . . 0 0 0 0.0.0.0200. 0
00.0 000 000 000 0.0 00. . 00. o. 000 0000500
000.000 000.000 000.000 .000..00 0.00.000 0000.000 0..00000 0000.000 000000.0.0 0.0.0. .
.000.. 00000. 00000. 0.000. 00.00. 000.0. 00000. .00.0. .0.00. .0000. .0092...
0.0.6..
.0
00.00.. 00000.. 00000.. 0..00.. 0.000.. 000.0.. 00000.. .0000. 00000.. .00...0 . N
000000., :00 0......5
:33
3:83.050 E20000.
00.000.- 000. .8008900 000.2
00% 65020502
3:05.030
3.00502
0 w .I a a CM. 0
0.000..- 00..0..- .0.... 0000...- 0.000..- 0. 000000.90.
a a . . . . . . Om .
000.00. 000000. 000000 000.000 m0.00.0 0..0000 0000000 ..00000 a. 0.0.30 000s.
2%.
0.0.50 .0 .0050-2

 

 

 

 

 

 

 

 

 

 

 

 

Nv 030.. 6.2.50

181

mo.v0 .0 .:005:a.0

0. .0005 00030.0 0. 50:00:80 .0005.0 00.1. .0 .:0o5:m_0 0. .0005 00030.0 0. 80:09:00 .002...o .0..v0 .0 2.0050050 .mo.v0 .0 E00535m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 0 . 0 0 0 0 0.0.0.0200 0
a 0 o 0 0 0 0..00..000 0 . 0.0000. .00 .00 0.000500
00000.000 000.0.000 00000.000 0 00000000 0 000 00. .
0000. 0000.. 0000.. 000... 0.000. 00.... ..00.. 5020.5
0 0.0..
00000.. 00000.. ..000.. 00000.. 00000.. .0000. 00000.. .0 000000., :00 0.50s
._0>0.
0.:0:0 E00 E020...
0000000.- 03080.- 000. .9205. m:.90Em_
a .0800. a .00000. 0 000000. 0 0000.0. .0... 0.0000 000.2
00%. 0.0.50 .0 30:52
00..
a .0.000. 0 000000. ...0000.. a 0.000.. 0000000.. -
6020052020 0.0000 :005.
3%. 62.00.50.029 00.0.0.6...
..0..000.. 0000000.. 0 ..0. .0.. 0000.00,. ..00.00..0 ..0000000 2.. .0005. 000.000.2020
:0 0000059020 00.0.00
0000.00.- 0... .000
0 N00000:. 0 ovmm .0. nwwmowm. .00. 00.20.). 0590.5.
.- 00.000.0- m0000000. 0 000000. .- a .00000. .- a 00.000.- “000.00.. 000M000 8.. 00020.5
00%
.00—5 .00.0.). ..0.0¢ 0a0..0><
0.00:0 00...".
0005.
0 .0005. 0 .0022 0 .0005. 0 .0005. 0 .0022 . .0005. 0000

 

 

0.00tm _0>0._ 3.5.00 .9 0.0.0..

182

APPENDIX C: ADDITIONAL ANALYSES

HYPOTHESIS 1

Corporate Model

Model 2 continues to determine other potential explanatory variables for
the diversiﬁcation model and then explore the impact of corporate level
moderators. Moreover, we replace the mean difference diversiﬁcation score with
the raw diversiﬁcation score for ease of interpretation. Because we are no longer
testing curvilinearity, it is easier to interpret raw variables than difference scores.
In this model, we extend our diversiﬁcation-market share model with Level 1
covariate to account for varying sizes of the retail chains, i.e. the number of
outlets possessed by the FBUs. We ﬁnd signiﬁcant relationships between both

diversiﬁcation ()710 =1.286508,t = 4.394) and chain size ()920 = .244419,t = 3.902)

and retail market share. The residual variance between corporations,

too = .61427, is larger than the original random ANOVA model. A range of

plausible values for corporate means, given that corporations have mean
diversification scores and average chain sizes, is between 0 and 1.71%. Finally,
we can see that corporate market share means vary signiﬁcantly after adding
diversification and chain store variables to the model

([2 =113.72729,df = 32, p < .001). Because there are important differences left to

explain, we need to introduce covariates that can help us explain the sources of
differences between corporations.

In the next series of models, we attempt to clarify the variability of the
regression equations across corporations. Our FBU-level model remains the
same as model 2. In models 3-6, we progressively incorporate corporate level
explanatory variables, namely international diversiﬁcation, corporate size,
international experience and total number of outlets. All Level 2 variables are
grand-mean centered to facilitate interpretation. In summary, we are trying to

183

explain why some corporations may have higher mean market shares than
others and why the magnitude of the relationship between certain FBU
characteristics and market share may strengthen or weaken.

Model 3 incorporates the corporation’s international diversiﬁcation
experience, namely the number of countries in which the corporation operates,
as an explanatory variable for the intercept. (The variable was tested as an
explanation for the relationship between in-country diversiﬁcation and
performance, but was not significant). The effect brought about from operating in
foreign environments has a positive effect on a corporation’s mean market share.
By adding an explanatory variable to the equation at Level 2, we change the

meaning of the residual, uoj and make the variance, too, a conditional variance of

average corporate market share after controlling for international diversiﬁcation
experience. We observe a signiﬁcant association between international
diversiﬁcation and mean market share. The residual variance between the Base

Model (100 =.4647) and this model (2'00 =.41905) is lower. Controlling for

international diversiﬁcation status, the range of plausible values for mean market
share is 0% to 1.36%. In other words, we explained 9.8% of the true between-
corporation variance in market share by adding the international diversiﬁcation
variable. Even after controlling for the effects of international diversiﬁcation,
corporations still vary signiﬁcantly in their mean market shares

(12 = 79.69298, df = 31, p < .001), encouraging further exploration of explanatory
variables.

Model 4 integrates the corporation’s retail experience, namely the number
of years that corporation has operated at the store-level, as an explanatory
variable for the mean market share. In this model, we assume that retail
experience has the same moderating effect across corporations. The residual

variance between the previous model (2'00 2 .41905) and this model (100 = .59594)

is higher. Controlling for international diversiﬁcation and retail experience, the
range of plausible values for mean market share is 0% to 1.66%. Even after
controlling for the effects of international diversiﬁcation and retail experience,

184

corporations still vary significantly in their mean market shares
(12 =102.01194,df=31,p<.001) .

Model 5 includes the corporation’s retail outlet size, namely the total
number of outlets that corporation has operates worldwide, as an explanatory
variable for the size of their local operations. The residual variance between the

previous model (700 =.59594) and this model (100 2.46462) is lower. Controlling

for international diversification, retail experience and outlet size, the range of
plausible values for mean market share is 0% to 1.43%. Even after controlling
for the effects of international diversiﬁcation, retail experience and outlet size,

corporations still vary signiﬁcantly in their mean market shares

(12 =92.34137,df=31,p<.001).

Country Model

In Model 2 we replace the mean difference diversification score with the
raw diversification score for ease of interpretation and to establish a comparable
deviance statistic. In Model 3, we extend our diversiﬁcation-market share model
with Level 1 covariate to account for varying sizes of the retail chains, i.e. the

number of outlets possessed by the FBUs. We ﬁnd positive signiﬁcant
relationships between both diversification (in, =1.000802,t = 3.833) and chain size

(772,, = .310188,t = 5.401) and retail market share. The residual variance between

countries, rm 2 .76434, is larger than the original random ANOVA model. A range

of plausible values for country means, given that FBUs within countries have
mean diversification scores and average chain sizes, is between 0 and 1.92%.
Finally, we can see that average country market share vary signiﬁcantly after
adding diversiﬁcation and chain store variables to the model

(12 = 203.47084,df = 48, p < .001).

In the next series of models (4-6), we attempt to clarify the variability of the
regression equations across countries. Our FBU-level model remains the same

185

 

as model 3. In models 46, we progressively incorporate country level
explanatory variables, namely emerging market status and market size. All Level
2 variables are grand-mean centered to facilitate interpretation. In summary, we
are trying to explain why some countries may have higher mean market shares
than others and why the magnitude of the relationship between certain FBU
characteristics and market share may strengthen or weaken.

Models 4 and 5 incorporate emerging market status and retail market size
as explanatory variables for the intercept. In Model 4, operating in certain
emerging markets may have a positive effect on a ﬁrm’s mean market share
because of the market’s technological capabilities and market opportunities
relative to the international retailers’ home markets. By adding an explanatory

variable to the equation at Level 2, we change the meaning of the residual, “o,-
and make the variance, 2'00, a conditional variance of average corporate market
share after controlling for emerging market status. We observe a signiﬁcant
association between emerging markets (070, = .634039,t = 2.499) and size of the
retail labor force (790, = —.303416,t = -5.095) and mean market share. The residual

variance between the Base Model (2'00 = .75471) and Model 5(100 = .34877) is

lower. Controlling for emerging markets and labor market size, the range of
plausible values for mean market share is 0% to 1.28%. In other words, we
explained 53.8% of the true between-country variance in market share by adding
the emerging market and market size variables. Even after controlling for
emerging market and market size effects, countries still vary signiﬁcantly in their

FBU mean market shares (12 =111.99485,df = 46, p < .001), encouraging further

exploration of explanatory variables.

186

 

HYPOTHESIS 2

Corporate Model

In our ﬁrst series of models, we are looking at the corporate level effects
on format level characteristics and strategies, and performance. Models 1
through 5 clarify the Level 1 (FBU) model. In Models 1 & 2, we establish the
governance parameters. Initially, a composite number for related and unrelated
wholly-owned subsidiaries (WOS) was used and found to be not significant.
Then, related and unrelated WOS measures were tried. Only the related (RWOS)

measure was signiﬁcant in the model (1710 = 471985,t = 2.253) , so this measure

was used. The joint venture (JV) parameter is also signiﬁcant

(fix, 2 -.616796,t = —2.539). However, we would not expect governance modes

alone to be excellent predictors of market share, so we use Models 3-5 to
explicate the Level 1 model by adding two kinds of formats (i.e., hypermarkets
and supermarkets) and the in-country operating experience in the format for the
F BU.

In Model 5 we extend our governance-market share model with Level 1
covariates to account for varying format experience and speciﬁc format
operations. We ﬁnd signiﬁcant relationships between RWOS

(flu, = 525922,: = 2.510) and JV (ﬁle = —.567424,t = —2.390) retail market share,

after controlling for supermarkets, hypermarkets and in-country experience. The
residual variance between countries, rm = .81083, is larger than the original
random ANOVA model. A range of plausible values for corporate means is
between 0 and 1.90%. Since the intercept term has a definition by default, we
can say that the average market share for all those FBU operations that are ngt:
RWOS, JVs, hypermarkets, or supermarkets would fall in that range. Finally, we
can see that average corporate market share vary signiﬁcantly after Level-1

covariates to the model (12 =110.99955,df = 33, p < .001).

187

In testing the time-invariance of the ﬁnal model, we re-ran the analysis
with market share for the same companies in 2004. This reduced the number of
useable observations to 334. The magnitudes and directionality of the parameter
coefﬁcients are similar, and the parameters are signiﬁcant.

Country Model

We use Models 2-5 to further explicate the Level 1 model. Unlike the
previous corporate model, because a “relatedness” term is not included, we can
capture the impact of being in the FBU’s most important retail segment. The
dominant term refers to the highest revenue generator in FBU group operations,
while the related business refers to other formats that have similar business
models and are contained within the same business segment as the dominant
format. In Model 5 we extend our govemance-market share model with Level 1
covariates to account for dominant business or related to the dominant business,
in-country experience and chain size. We ﬁnd signiﬁcant relationships between

ACQ (filo = .274965,t 21.729) and retail market share, after controlling for
membership in the most important segment, in-country experience and chain
size. The residual variance between countries, r00 =1.08165, is larger than the

original random ANOVA model. A range of plausible values for corporate means
is between 0 and 2.08%. Since the intercept term has a deﬁnition by default, we
can say that the average market share for all those FBU operations that are g9_t:
Acquisitions or in the dominant segment would fall in that range. Finally, we can
see that average corporate market share vary signiﬁcantly after Level-1

covariates to the model (12 = 286.98374,df = 48, p < .001) .

HYPOTHESIS 3

Corporate Model

Our ﬁrst series of analyses models corporate level inﬂuences on F BU
format operations. The Base Model through Model 5 clarify the Level 1 (FBU)

188

model. In Models 1 & 2, we establish the governance parameters. Our Base
Model consists of five of the six diversiﬁcation strategies. This suggests that the
sixth diversiﬁcation strategy is represented in the term (among other things). We

ﬁnd that parameters for RWOS 07,0 =1.026389,t = 3.710), related acquisitions
(RACQ) 0730 =1.088741,t = 4.008) and unrelated joint ventures (UJV)

0750 =1.08874l,t = 4.008) are signiﬁcant. However, the mean market share

ordering is not as expected for the related pairs. In this ﬁrst model, we ﬁnd that
RACQ>RWOS>RJV, although this ordering changes by ﬁnal model that includes
covariates. The parameters for related joint ventures (RJV) and unrelated
wholly-owned subsidiaries (UWOS) are not signiﬁcant, but at the magnitudes and
in the directions expected.

In Models 1-5, we add important covariates to the Level-1 model relating

to the format type and chain size. The format types included in the models are
hypermarkets, supermarkets, convenience stores and soft discounters. We ﬁnd

significant relationships between RWOS (filo = .936285,t =3.320), RACQ

()730 = 950419,! =3.460), UJV 0750 = —1.194557,t = —2.536) and retail market share,
after controlling for supermarkets (y70 = .587541,t = 2.316), hypermarkets

(ya, = .71737 1,1 = 2.833), convenience stores (no = —.642250,t = —1.601), soft
discount (790 =.570162,t =1.598) and in—country experience

(y100 = .269285,t = 2.815). The residual variance between countries is

too = .96708. A range of plausible values for corporate means is between '0 and

2.02%. Since the intercept term has a definition by default, we can say that the
average market share for unrelated acquisitions that are not hypermarkets or
supermarkets. Finally, we can see that average corporate market share vary
signiﬁcantly after Level-1 covariates to the model

(12 =126.10320,df :33, p < .001).

189

Country Model

The next series of analyses models country-level effects on FBU format
operations. The Base Model to Model 5 clarify the Level 1 (FBU) model. Our
Base Model includes ﬁve of the six governance pairs. We ﬁnd that parameters

for RWOS (z, =.701736,t = 2.87), RJV (7“,, = 541675,! =1.881) RACQ
0730 =1.154387,t = 4.690) and UJV (1750 = —.774334,t = —1.702) are signiﬁcant, while

the parameter UWOS is not significant. The directions for each of the
related/unrelated variables are as expected, but the magnitudes are not. In this
model, estimated mean market shares are RACQ>RWOS>RJV and the superior
performance from being an RACQ does not change in our country analyses. In
Models 1-5, we add important covariates to the Level-1 model relating to the
format type and chain size. The format types included in the models are
hypermarkets, supermarkets, convenience stores and soft discounters. We ﬁnd

signiﬁcant relationships between RWOS (97,0 = .684133,t = 3.254), RJV

(f2, = 821305,: =3.274) RACQ (2,0 = .978328,t = 4.682), UJV

()750 = —.648431,t= -1.710) and retail market share, after controlling for
supermarkets (770 =.043314,t =.244), hypermarkets (760 =.799349,t =4.785) ,
convenience stores (780 = —1.445537,t= —5.417) , soft discounter

(ygo =.136231,t = .514) and in-country experience (y100 = .477331,t = 10.545). The

residual variance between countries is 2'00 2 1.04648 . A range of plausible values

for corporate means is between 0 and 2.05%. Since the intercept term has a
deﬁnition by default, we can say that the average market share for unrelated
acquisitions that are not hypermarkets or supermarkets. Finally, we can see that
average corporate market share vary signiﬁcantly after Level-1 covariates to the

model (12 = 347.41174,df = 48, p < .001).

190

 

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

1ijijjjjmjjjjjijjjjm