.vhu. E.
manure“

P ‘5'
Muir.

 

5 1 z
t 1%

 

,. .Jr 1s}? 2 .

,I).l.‘!
. V.

x 3. ..
1.93.3331

“can 1, ,.
,3. 53...;

ill- $441.19-...ic- . . 1
.a‘uwhﬂﬁﬁéwzaﬁzéz $ ‘.......2...§.. a. 5w; ‘ ﬁgéw

 

Lac)?

This is to certify that the
dissertation entitled

COLLABORATING AND COMPETING? UNCOUPLING
VALUE CREATION AND VALUE APPROPRIATION IN

PhD.

STRATEGIC ALLIANCES

presented by

Federico Aime

has been accepted towards fulﬁllment
of the requirements for the

degree in Management

 

 

    

 

 

{7 ’ Major Professor’ Signat
W /‘ 3007
/
D e

MSU is an Afﬁrmative Action/Equal Opportunity Institution

 

<-.— —.-.-.-A--—»--.-.-u—.-

.---—.¢— u..- -._.

 

_UBRARY
Michigan State
University

 

 

 

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

 

COLLABORATING AND COMPETING? UNCOUPLING VALUE CREATION
AND VALUE APPROPRIATION IN STRATEGIC ALLIANCES

By

Federico Aime

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Management

2007

ABSTRACT

COLLABORATING AND COMPETING? UNCOUPLIN G VALUE CREATION
AND VALUE APPROPRIATION IN STRATEGIC ALLIANCES

By
Federico Aime
A broadly supported general ﬁnding from collaboration research is that
organizations involved in inter-ﬁrm collaboration usually achieve above average
performance and survival, although with substantial variation in outcomes (Burt, 1983;
Hagedoorn & Schakeraad, 1994; Mitchel & Singh, 1996; Owen-Smith & Powell, 2003;
Singh & Mitchell, 2005; Stuart & Podolny, 1999; Uzzi, 1997). Explanations for those
large variations in outcomes (e. g., structural and capability based approaches) offer
conﬂicting predictions and ﬁndings as to their effect on alliance performance (Ahuj a,
2000; Bae & Gargiulo, 2004; Burt, 1992; Chung, Singh, & Lee, 2000; Coleman, 1988,
1990; Gulati & Singh, 1998; Hargadon & Sutton, 1997; Stuart, 2000; Walker, Kogut &
Shan, 1997). I identify two distinct theoretical dimensions of strategic alliances:
collaboration (i.e., value creation) and competition (i.e., value appropriation) (Burt, 1991;
Hamel, Doz, & Prahalad, 1989; Inkpen, 2001; Lax & Sebenius, 1986; Teece, 1986). In
the literature these dimensions are combined in the construct of performance and yet
often have opposite effects on an organization’s ability to derive value ﬁ'om an alliance.
By formulating and testing models for these two dimensions of performance I provide an
explanation to resolve the conﬂicting guidance of both the structural and the capability
based approaches about alliances effect on performance. For example, I theoretically
develop and test the idea that network closure may result in value creation within

alliances while structural holes may facilitate value appropriation by individual partners.

Similarly, I study how a ﬁrm’s partners’ resources may predict value creation but
differential resource endowments between partners may result in dissimilar value
appropriation potential by individual partners.

I test these models on a large longitudinal sample of strategic alliances with
sequential cross-nested multilevel techniques. This analytical approach addresses
methodological difﬁculties implicit in the study of networks data (e. g., dependencies
among observations, cross-level interactions) that are not addressed in previous strategic
alliances research and is capable of dealing with the interdependencies between ﬁrms in
the sample. My reformulation of theory and methods to study alliance performance
contributes to realizing the potential of resource based and structural approaches as

powerful explanations and sources of guidance for ﬁrms’ strategic alliance choices.

F or Elsa

iv

TABLE OF CONTENTS

CHAPTER 1 -

INTRODUCTION ............................................................................................................... 1

CHAPTER 2

THE IMPACT OF PARTNERS’ AND ALLICANCE ATTRIBUTES ON

ALLIANCE VALUE CREATION: BACKGROUND AND PREDICTIONS ................. 12
Capabilities and Alliance Value Creation ................................................................. 12
Strength of Ties and Alliance Value Creation .......................................................... 16
Structural Positions and Alliance Value Creation .................................................... 18
Network Size and Alliance Value Creation .............................................................. 23

CHAPTER 3 .

THE IMPACT OF PARTNERS’ AND ALLIANCE ATTRIBUTES IN

PARTNERS’ APPROPRIATION: BACKGROUND AND PREDICTIONS .................. 26
Relative Strategic Stakes and Firm Value Appropriation ......................................... 28
Relative Partner Capabilities and Firm Value Appropriation ................................... 30
Relative Availability of Alternative Partners and Firm Value Appropriation .......... 31
Relative Structural Holes and Firm Value Appropriation ........................................ 32

CHAPTER 4

METHODS ........................................................................................................................ 35
Data and Sample ....................................................................................................... 35
Measurement ............................................................................................................. 38
Dependent Variable .................................................................................................. 41
Explanatory Variables ............................................................................................... 41
Control Variables ...................................................................................................... 46
Analysis ..................................................................................................................... 47
Value Creation Model ............................................................................................... 48
Appropriation Model ................................................................................................ 50

CHAPTER 5

RESULTS .......................................................................................................................... 52

CHAPTER 6

DISCUSSION AND CONCLUSIONS ............................................................................. 61
Structure .................................................................................................................... 62
Firms’ Capabilities .................................................................................................... 65
Firm Power ................................................................................................................ 68
Limitations ................................................................................................................ 70
Conclusions ............................................................................................................... 71

REFERENCES .................................................................................................................. 76

LIST OF TABLES

Table 1
Variable Names and Descriptions ...................................................................................... 39
Table 2
Parameter Development for the Relative Stakes Measure ................................................. 45
Table 3
Tests for Hypotheses .......................................................................................................... 54
Table 4
HCM2 Results for Models 1 and 2 (Dependent Variable = Yin”! +1 ................................... 57
Table 5
HCM2 Results for Models 3 and 4 (Dependent Variable = Y.- — m +1 ................................ 59

vi

LIST OF FIGURES

Figure 1
Interaction Effect between Alliance Capabilities and Tie Strength ................................... 55

vii

CHAPTER 1
INTRODUCTION

What explains differences in value creation that result from inter-ﬁrm
collaboration and the successive appropriation of that value by particular ﬁrms involved
in collaborative strategies? A broadly supported general ﬁnding from collaboration
research is that organizations involved in inter-ﬁrm collaboration usually achieve above
average performance and survival, although with substantial variation in outcomes (Burt,
1983; Hagedoorn & Schakeraad, 1994; Mitchel & Singh, 1996; Owen-Smith & Powell,
2003; Singh & Mitchell, 2005; Stuart & Podolny, 1999; Uzzi, 1997). In this dissertation,
I explore the conditions that shape value creation and value appropriation that result from
inter-ﬁrm collaboration strategies. I adopt a social networks framework to theorize and
test a multilevel model of value creation and value appropriation in inter-firm
collaborations

Inter-ﬁrm collaboration is a broad concept that includes the many, usually formal,
external linkages that may connect organizations for the purpose of cooperation. These
linkages can take many forms (e.g., vertical relationships, research consortia, joint
ventures, strategic alliances). In this study I focus on strategic alliances, deﬁned as
voluntarily initiated cooperative agreements between ﬁrms for the purpose of
exchanging,tsharing, or co-developing resources or capabilities towards an alliance goal
(Harrigan, 1985; Gulati, 1999).

As organizations commit to differential arrays of alliances, some patterns of

alliance strategy are likely to produce more value than others. The goal of this

dissertation is to account for the value of alliance strategy choices by showing how
partner speciﬁc attributes (i.e., ﬁrm level attributes, such as capabilities), and particular
combination patterns of such partner attributes in alliances (i.e., alliance level attributes,
such as tie strength) interact to impact the value creation resulting from alliances and the
subsequent appropriation of that value by ﬁrms. For example, a partner’s speciﬁc
endowment of resources and capabilities may (Stuart, 2000; Zaheer & Bell, 2005) or may
not (Bae & Gargiulo, 2004) explain value creation in alliances, depending on the strength
of the tie between the partners, the relative commitment of the partners to this particular
alliance as opposed to other alliances, and the level of interconnection (density) of the
overall network of alliances that surrounds this particular alliance.

In this study, I address ﬁve main limitations and contradictory ﬁndings of
previous research with the goal of explaining the value of alternative alliance strategy
choices for ﬁrms. In the next few paragraphs, I outline these limitations and contradictory
ﬁndings of previous research and the related new directions I propose for advancing the
discussion of the relationship between alliances and ﬁrm performance.

First, previous research on the relationship between alliances and performance
failed to address the implications of alliances being simultaneously cooperative and
competitive relationships (Burt, 1991) for ﬁrm performance. This characteristic of
alliances may help understand better the differences in ﬁndings of previous research.
Firm attributes, alliance attributes and network positions may each have differential
effects on value creation and value appropriation for ﬁrms that engage in a strategic
alliance. While an alliance is a cooperative agreement by deﬁnition, it is also necessary to

recognize that “all alliances involve a negotiated bargain between the partners” (Inkpen,

2001: 418). By focusing exclusively on ﬁrm outcomes, previous research on alliances
assumes full capture of value created in alliances by participating ﬁrms. This assumption
hides two distinct theoretical dimensions of strategic alliances, collaboration (i.e., value
creation) and competition (i.e., value appropriation) (Burt, 1991; Hamel, Doz, &
Prahalad, 1989; Inkpen, 2001; Lax & Sebenius, 1986; Teece, 1986), which the literature
combines in ﬁrm performance outcomes of alliances. Yet, these dimensions often have
opposite effects on an organization’s ability to derive value from an alliance. For
example, having a partner with large marketing capabilities may imply better potential
for distributing a ﬁrm’s product through that alliance, but may also imply that the large
partner has a strong bargaining position in the alliance to capture most of that value. The
size of the appropriation of value that each ﬁrm derives from an alliance is the result of a
c00perative game (Brandenburger & Stuart, 2001; Lippman & Rumelt, 2003) between
partners who cooperate and compete simultaneously (Lax & Sebenius, 1986; Hamel,
Doz, & Prahalad, 1989) and, therefore, is the outcome of exchange processes, bargaining,
competition, relative power, information differences and the overall structure of the
relationship.

To deal with this limitation of previous research, in this study I separate value
creation from value appropriation by making predictions at both the alliance and the
individual ﬁrm level. I ask: what explains differences in value creation that result from
strategic alliances? And then, I further specify the effects of alliances for ﬁrms by
questioning: What explains the successive appropriation of that value by the particular
ﬁrms involved in strategic alliances? Separating these questions has early antecedents in

Teece’s (1986) discussion of complementarities and appropriation, and in Lippman and

Rumelt’s (2003a; 2003b) payments and bargaining perspectives. Their work implies the
need to recognize that alliances are not all created equal. As organizations enter alliances,
some of their alliances will do better than others. I explain dyadic level outcomes with
multilevel explanations that take into account partners’ attributes, dyadic attributes and
their interaction within a complete network of alliances. By explaining dyadic level
outcomes, I take a ﬁrst step to identify and explain the conditions under which alliances .
have more potential to create value, and the differential appropriation of value among the
partners in the alliance.

Second, I integrate structuralist and resource based approaches to address two
main criticisms of collaboration research: (a) criticism by critics claiming that the
tendency of network analysis to focus on the structure of relationships (e. g., density of
ﬁrm’s alliances network) and neglect the content of ties (i.e., capabilities of the partners
in a ﬁrm’s alliances netWork) leads to treating all ties as comparable and thus
disregarding networks’ content or context (Goodwin & Emirbayer, 1994; Smith-Doerr &
Powell, 2005; Stinchcombe, 1990); and (b) criticism by critics claiming that an
overemphasis on the content of ties leads to treating all relationships as comparable
(Granovetter, 1985; Gulati, 1995). Recent work on alliances has produced some steps in
addressing these critiques by providing a series of studies that integrate the strategic
needs (i.e., external resource based view) approach with the structural approach to
understand the outcomes of alliances (Ahuj a, 2000; Bae & Gargiulo, 2004; Stuart, 2000;
Zaheer & Bell, 2005). They do this by bringing together partners’ attributes and ﬁrm
speciﬁc alliance network attributes in explaining ﬁrm performance. I extend these efforts

and model inter-ﬁrm collaboration with multilevel techniques to allow explanations at

different levels to affect each other in clarifying the relationship between ﬁrm and
network capabilities and value creation.

Previous research fails to take advantage of one of the key beneﬁts of network’s
approaches: the multilevel nature of the data (e.g., the ﬁrm, the alliance, the network).
For the purpose of analysis most networks data are analyzed separately at different levels
precluding theoretical developments across levels (Jones, Hesterly, & Borgatti, 1997;
Monge & Contractor, 2003). For example, all of the studies that have addressed the need
to integrate RBV and network constructs to explain ﬁrm performance resulting from
alliances (e. g., Ahuja, 2000; Bae & Gargiulo, 2004; Stuart, 2000; Zaheer & Bell, 2005)
have done so by constructing network variables as ﬁrm level predictors (the density of a
frrm’s network), therefore rendering themselves unable to explain the value of speciﬁc
alliance choices. This single level approach accounts, in part, for the contradictions in
their ﬁndings. I believe the recent literature on alliances has paid more attention to
portfolios of alliances at the expense of individual choices in a way that may have
obscured some of the potential drivers of the value of alliances for ﬁrms. By aggregating
their explanatory variables at the individual (i.e., ﬁrm) level of analysis, explanations
based on the overall network of alliances of a ﬁrm has been inadequately utilized to
provide explanations for the value of speciﬁc alliance choices. There are no studies of
alliances, to the best of my knowledge, which develop their theory and tests to take
advantage of the multilevel beneﬁts of network data. Understanding, for example, how
both individual partners’ attributes (i.e., ﬁrm level constructs) and the alliance
conﬁguration (i.e., alliance level constructs) interact to produce alliance outcomes within

a large array of other alliances in which both partners also beneﬁt has not been addressed

by the alliances literature to date. In this study I explicitly model alliance outcomes
through a multi-level approach that fully recognizes the role of ﬁrm and structural
attributes within ﬁrms’ alliance networks.

Third, I seek to reconcile the contradictory predictions and ﬁndings about the
impact of network closure and structural holes on the performance effects of alliances
(Ahuja, 2000; Burt, 1992; Coleman, 1988, 1990; Gulati, 1995; Gulati & Singh, 1998;
Hargadon & Sutton, 1997; McEvily & Zaheer, 1999; Uzzi, 1997; Walker, Kogut & Shan,
1997; Zaheer and Venkatramarr, 1995). Speciﬁcally, I address contradictory ﬁndings in
previous research, some ﬁnding a positive relationship between structural holes and ﬁrm
performance (e.g., Hargadon & Sutton, 1997; McEvily & Zaheer, 1999) and
contradictory research that suggests closure enhances performance (Ahuj a, 2000; Zaheer
& Venkatrarnan, 1995). I propose that, instead of being contradictory, these results reﬂect
different processes operating at different levels. By formally recognizing the multi-level
nature of strategic alliances’ outcomes these apparent contradictions can be resolved.

Fourth, I seek to understand the degree to which alternative combinations of
partner capability endowments could be appropriately regarded as the normative ideal
(Bae & Gargiulo, 2004; Chung, Singh, & Lee, 2000; Stuart, 2000; Walker, 1889).
Findings about the capability endowments of partners are also conﬂicting in the alliances
literature, showing that capability endowment of partners may (e. g., Stuart, 2000) or may
not (e.g., Bae & Gargiulo, 2004) be signiﬁcant for ﬁrm performance. I argue that
capability endowments have multiple meanings for the performance of ﬁrms that
participate in an alliance. Related with the distinction between value creation and value

appropriation, I theorize that a ﬁrm’s partners’ capabilities may have distinct implications

for value creation at the alliance level and for value appropriation at the ﬁrm level that
may help explain the contradictions in previous ﬁndings. For example, partners’
resources may predict value creation, but differential resource endowments between
partners may result in differential value appropriation by individual partners. I provide an
explanation to resolve the apparently contradictory implications of partners’ capability
endowments for ﬁrm performance.

In addressing these initial four main limitations and contradictory ﬁndings of
previous research I develop two conceptual models to explain the value of alternative
alliance strategy choices for ﬁrms. In the ﬁrst model, I develop and test a multi-theoretic
explanation for which alliances are likely to create more value. In developing this model,
I build on several theoretical traditions in developing my arguments for this model. Since
resources and capabilities are critical to the pursuit of competitive advantage (Henderson
& Cockbum, 1994; Nelson, 1991; Teece, Pisano, & Shuen, 1997, Barney, 1991), I start
by building on both internal and external capability based approaches in developing
arguments to explain alliance outcomes. Traditional internal RBV approaches focus on
resource and capability endowments within the ﬁrm, and view differences in ﬁrm
capabilities to be a result of imperfections in factor markets, chance, resource immobility
(Barney, 1986, 1991), path dependence (David, 1985; Penrose, 1959), or causal
ambiguity (Lippman & Rumelt, 1982; Red & DeFillippi, 1990; Rumelt, 1984). More
recently, research has focused on external RBV arguments that take into account the
resources and capabilities that ﬁrms may be able to access through their external
networks (Gulati & Gargiulo, 1999; Gulati, Nohria, & Zaheer, 2000; McEvily & Marcus,

2005; Nohria, & Garcia-Pout, 1991; Rao, 1994; Zaheer & Bell, 2005).

Next, I draw from on social capital theory (Burt, 1992, 1997, 2001; Coleman,
1988, 1990; Granoveter, 1973, 1985; Lin, 2001; Simmel, 1955) to explore how the
strength of the ties between partners with particular capability endowments and how the
density (i.e., closure and structural holes) of the networks in which the alliances are
embedded jointly affects the value generated in particular alliance arrangements. Finally,
I use density dependence arguments ﬁ‘om ecological processes theorizing (Harman &
Freeman, 1987, 1988, 1989; Harman & Carroll, 1992) to support predictions for a
curvilinear relationship between the number of alliances partners are involved in and the
value creation within a strategic alliance.

In the second model, I examine how partner and alliance characteristics enhance a
partner’s ability to appropriate more value from the alliance. In this model, I use a
bargaining perspective (Lippman & Rumelt, 2003a, 2003b), to confront arguments from
structural holes and closure perspectives in social capital theorizing (Burt, 1992, 1997;
Coleman, 1988, 1990), and resource based and social exchange theorizing (Bae &
Gargiulo, 2004; Gulati & Gargiulo, 1999; Gulati, Nohria, & Zaheer, 2000; Zaheer &
Bell, 2005), in explaining value appropriation by individual ﬁrms. In addition, I draw on
social exchange theory (Blau, 1964; Emerson, 1962; 1972a, 1972b; Homans, 1950, 1974)
and its more recent developments known as Network Exchange Theory (NET) (Cook &
Whitemeyer, 1992; Markovsky, Willer, & Patton, 1988; Skvoretz & Willer, 1993; Willer
& Skvoretz, 1997; Willer, Walker, Markovsky, Willer, Lovaglia, Thye, & Simpson,
2002), to develop predictions about relative bargaining power between alliance partners

and their ability to appropriate the value created in an alliance.

Finally, the ﬁfth limitation of previous research on alliances is it’s failure to
address some methodological issues that emerge from alliances data. Since alliance (e. g.,
network, dyads) data is relational, it is composed, by deﬁnition, of non-independent
observations. Also, alliances data is essentially multilevel. Therefore, “standard”
statistical approaches that assume independent subjects, such as regression analysis and
ANOVA are not appropriate (Holland & Leinhardt, 1981; Lazeaga & Van Duijin, 1997;
Van Duijin, 1995; Wasserrnan & Pattison, 1996). To analyze and understand the
multilevel explanations in my proposed models, and to acknowledge the issues of
interdependence in network data, I use an innovative set of methods. My ﬁrst conceptual
model is a multilevel cross-nested random effects model with predictions at the dyadic
level (i.e., the alliance) of analysis. The cross-nested structure is necessary to reﬂect the
fact that lower-level units (i.e., ﬁrms) in the study of alliances are usually cross-classiﬁed
by two or more higher-level units (i.e., alliances). A ﬁrm may take part in more than one
alliance at a time, therefore making it necessary to cross-nest that ﬁrm within the multiple
alliances it takes part in. In this way, the model can provide explanations based on ﬁrm or
alliance level predictors, while accounting for the overall network of relationships in
which each ﬁrm is involved. Multilevel methods in networks analysis have only recently
been deve10ped and have basically addressed the issue of selection (who selects whom
for a relationship of any sort), with individual level (usually binary) dependent variables.
I build on the logic of such models (e. g. the p 1 model (Holland & Leinhardt, 1981), the p.
model (Wassennan & Pattison, 1996), and the p; model (Van Duijin, 1995; Lazeaga &
Van Duij in, 1997)) and develop a multilevel performance model (MPM) with dyadic

level dependant variables that is fully estimable with likelihood multilevel algorithms.

My second conceptual model is also multilevel but has directed predictions. I utilize a
multilevel selection model with predictions at the individual level that can be ﬁtted using
linear mixed model algorithms.

In smnmary, in this dissertation I attempt to explore two basic questions: First, I
ask: What explains differences in value creation that result from inter-ﬁrm collaboration?
And then I ﬁrrther specify the effects of alliances for ﬁrms by questioning: What explains
the successive appropriation of that value by the particular ﬁrms involved in collaborative
strategies? To achieve this, this dissertation is organized as follows.

Chapter 2 of this dissertation discusses how different partners’ attributes (e. g.,
capability endowments, size, experience in alliances, industry they belong in) combine
both between partners and with alliance level attributes (e.g., level of commitment to the
alliance of partners, size differences between partners) to inﬂuence value creation at the
level of the alliance. Hypotheses regarding multilevel main and interaction effects of
ﬁrms and alliance level attributes on alliance value creation are then advanced.

Chapter 3 of this dissertation discusses how different partners’ attributes (e. g.,
capability endowments, size) combine both between partners and with alliance level
attributes (e. g., structural position of the alliance in the overall partners’ alliance network,
size differences between partners) to inﬂuence value appropriation for the ﬁrms involved
in alliances. Hypotheses regarding multilevel main and interaction effects of ﬁrms and
alliance level attributes on ﬁrm speciﬁc performance are then advanced.

Chapter 4 describes the methods used to test this dissertation’s predictions.
Sampling, data collection, measures, and data structure decisions are ﬁrst explained and

discussed. Then, analytic strategies for both studies are described in detail.

10

Chapter 5 describes the study results and discusses the extent to which the
dissertation’s hypotheses are supported. Finally, Chapter 6 then provides an in-depth
discussion of the ﬁndings, what they mean for the larger literatures on alliances,
resources and capabilities, networks and social exchange, and what they suggest in terms

of avenues for future research.

11

CHAPTER 2

THE IMPACT OF PARTNERS’ AND ALLIANCE ATTRIBUTES ON ALLIANCE
VALUE CREATION: BACKGROUND AND PREDICTIONS

What explains differences in value creation that result from inter-ﬁrm
collaboration? In this chapter I present arguments and develop hypothesis for the effect of
ﬁve ﬁrm or alliance attributes on alliance value creation. I start by discussing previous
ﬁndings based in resource approaches for alliance value creation. Then I develop
arguments for how the strength of the relationship implied by the alliance (i.e., strength
of ties approaches, Granoveter, 1985; McEvily & Zaheer, 1999) may have direct and
interaction effects on value creation for the alliance. Third I discuss structural positions of
partners’ in the alliance (i.e., closure and structural holes) and their effect on alliance
value creation. Finally, I discuss how the size of partners’ networks may also affect the

value created in alliances.

Capabilities and Alliance Value Creation

Resource-based views (RBV) of the ﬁrm have a long tradition of research in
strategic management and provide one of the core explanations for differential ﬁrm
performance (Barney, 2001; McGahan & Porter, 1995; Peteraf, 1993; Rumelt, 1991). The
focuses of this literature are resources and capabilities that are internal to the ﬁrm and
that provide the ﬁrm with sustainable competitive advantages resulting in differential
rents (Makadok, 2003; Miller & Shamsie, 1996; Yeoh & Roth, 1999). However,
competitive or rapidly changing environments make it difﬁcult for a single ﬁrm to own

all resources and capabilities needed to develop and sustain actual and ﬁrture competitive

12

advantages (Child & Faulkner, 1998; Dyer & Singh, 1998; Harrison, Hitt, Hoskisson, &
Ireland, 2001; Pfeffer & Salancik, 1978). Therefore, recent research has explored the
strategic value of external resources available to the ﬁrm through its alliances (Bae &
Gargiulo, 2004; McEvily & Marcus, 2005; Gulati 1999; Stuart, 2000; Zaheer & Bell,
2005)

Collaborative relationships provide access to a wider scale and scope of external
resources and capabilities (e.g., information, technology, manufacturing or marketing
capabilities, and ﬁnancial resources) than is normally available to purely internally
focused ﬁrms (i.e., ﬁrms that rely exclusively on their own resources and capabilities
rather than seeking external resources and capabilities through alliance strategies). The
co-specialization of these external resources and capabilities with ﬁrm speciﬁc resources
and capabilities either in the form of pooling or provision is a fundamental reason to enter
an alliance. It is important to look at how these external resources and capabilities may
offer a complementary explanation to the purely internal nature of co-specialization
studied by traditional resource based views. As such, external resources become a
fundamental strategic choice for organizational decision makers. If alliances are a means
to co-specialize ﬁrm resources and capabilities with resources and capabilities that are
accessed through the alliances in order to better achieve ﬁrm objectives, their
effectiveness should be contingent upon the size and quality of the resources and
capabilities of the alliance partners (Bae & Gargiulo, 2004; Chung, Singh, & Lee, 2000;
Stuart, 2000; Walker, 1889). In this line of thought, recent research (Stuart, 2000) ﬁnds
that sales growth and innovation rates for ﬁrms in a large sample of semi-conductor

producers were positively related with the size and innovativeness of their aggregate pool

13

of partners. Similarly, Zaheer and Bell (2005) ﬁnd that innovative capabilities of both the
focal ﬁrm and its alliance network have a positive relation to performance in Canadian
mutual fund companies. In contrast, Bae and Gargiulo (2004), fail to ﬁnd a relationship
between partners’ resource or capability endowments and ﬁrm performance.

A major limitation of these studies is that virtually all of them make predictions at
the ﬁrm level, potentially confusing value creation and value appropriation effects. In
order to decouple these two effects, I argue that larger capability endowments of partners
may not only result in more value creation at the level of the alliance but also result in
more appropriation of that value by partners with larger capability endowments relative
to a focus ﬁrm. Note here that large ﬁrms are also more likely to ally with partners that
have larger capability endowments therefore confounding these two drivers of alliance
performance in previous research. Moreover, by analyzing capabilities in a one level
model of performance, it is difﬁcult to separate the direct effect of capabilities on
performance from the effects of capabilities on performance that may be attributed to the
alliance. I therefore create a series of capability based explanations for alliance
performance in this paper. In this section I discuss how the combination of each partner’s
capabilities may affect alliance value creation. Later, in the value appropriation section of
the paper, I will discuss how relative endowments of capabilities between partners may
have an effect on how much of that value is likely to be appropriated by each partner in
the alliance.

I deﬁne capabilities, following Amit and Schoemaker, as “capacity to deploy
resources ...using organizational processes” (1993:35) to produce outcomes. Given that

the amount or the quality of each of the partner’s capabilities (Stuart, 2000; Zaheer &

14

Bell, 2005) and the organized interaction between their capabilities in an alliance should
have positive effects on value creation that is independent of the direct effect of
capabilities in each ﬁrm’s performance, I predict that partners’ individual capabilities will

have a positive effect on value creation at the level of the alliance.

Hypothesis 1A: The value created at the alliance level of analysis increases with the

capabilities controlled by each alliance partner.

Additionally, it is important to note that partner’s capabilities in an alliance do not
only operate independent of each other, but are co-specialized to achieve the alliance
purposes. Therefore, it is important to consider that this co-specialization should produce
additional value up and above the value the resources of each partner could have
produced in an arm-length relationship. Therefore, I argue that the dyad level interaction
of the partners’ capabilities, will have a separate positive effect on value created at the
alliance level of analysis. The co-specialization of a frrrn’s capabilities with better
partner’s capabilities should have more than a pure additive effect on value creation
given that better partner’s capabilities should allow for more productivity of a ﬁrm’s own
capabilities in the alliance. Therefore, I hypothesize that the value created in alliances is
as much a result of the capabilities of each of the alliance partners as of the co-
specialization of the capabilities of both partners in the alliance. I call alliance capabilities

to the co-specialization of each of the partner capabilities in the alliance.

Hypothesis 1B: The value created at the alliance level of analysis increases with the

alliance capabilities controlled by both partners in the alliance.

15

Strength of Ties and Alliance Value Creation

The idea of tie strength (Granoveter, 1973, 1974, 1982) is a foundational element
of network research that focuses on the ties level within networks. When there is a tie or a
relation between two actors (e.g., an alliance between two ﬁrms), tie strength indicates
the quantity of the relation (i.e., amount of, for example, interaction, intimacy, or
reciprocal services between actors in a tie or partners in an alliance) (Monge &
Contractor, 2003). Tie strength has been typically understood as a function of frequency
of interaction (Granoveter, 1973, 1974, 1982), reciprocal obligations, commitment and
investment in close relationships (Lund, 1985), empathy and unconditional regard
(Cramer, 1986), and intimacy (W egener, 1989). Strong ties represent particularistic
relationships and particularistic knowledge (Marsden, 1990). When ties are stronger,
liking, information about the other, and trust create informal interdependence (Marsden
& Campbell, 1984; McEvily & Zaheer, 2005b; Mitchell, 1987; Podonly & Baron, 1997).
In contrast, weaker ties are lower in intensity. They are less personal, less reciprocal, and
more substitutable (Granovetter, 1973, 1982; Lin, Ensel, & Vaughn, 1981).

Many researchers in the interorganizational networks literature have recently
emphasized the importance of strong ties (Kraatz, 1998; Lazerson, 1995; McEvily &
Zaheer, 1999, 2005; Suarez, 2005; Uzzi, 1997) to access both redundant and non-
redundant capabilities. For example, McEvily and Zaheer (1999) demonstrated that new
ideas, information, and opportunities can be sourced through contacts that are
simultaneously strong and non-redundant. Similarly, long-term relationships (i.e., strong
ties) between large manufacturers and small family run artisanal knitwear. ﬁrms in Italy

resulted in a successful value strategy for the fashion industry (Lazerson, 1995).

16

Alliances that imply strong ties are therefore better able to create more value since trust,
commitment of resources and capabilities, and reduced opportunistic behavior, allow for
more productive interactions between the ﬁrms in such alliances. This may lead to
increased commitment of valuable resources to the alliance and, thus, greater value
creation potential. Firms can reduce uncertainty and create more predictable
environments (Cook, 1977), develop trust, commitment and responsiveness (Dore, 1983)
as in the “just in time” strategies of Japanese manufacturers, and, in general, facilitate
organizational attempts to adapt to environmental change (Kraatz, 1998), by developing
strong ties with other organizations. Based on these arguments, I hypothesize that
strength of ties between partners (i.e., Alliance Strength) should have a positive effect on

value creation at the level of the alliance.

Hypothesis 2a: The value created at the level of the alliance increases with the tie

strength between the partners in the alliance.

Furthermore, since as hypothesized in hypothesis 1 the value created in an
alliance should be positively affected by alliance capabilities, and since tie strength may
be a relevant measure of the commitment of partners to that interaction, 1 also argue here
that the impact of alliance capabilities on value created at the level of the alliance will be
enhanced by the strength of the relationship between the partners. For example, in an
alliance, the strength of the tie between partners may result in improved commitment of
the existing resources or capabilities, therefore promoting the creation of value at the
level of the alliance. When ties are stronger, interpersonal commitments and trust create

informal interdependence (Marsden & Campbell, 1984; Mitchell, 1987; Podonly &

17

Baron, 1997) therefore promoting a larger speciﬁc commitment of partners capabilities

and resources to the alliance.

Hypothesis 2b: The relationship between alliance capabilities and the value created at
the level of the alliance is positively affected by tie strength between the partners in the

alliance.
Structural Positions and Alliance Value Creation

Several studies have indicated that the position ﬁrms occupy in inter-
organizational networks inﬂuences ﬁrm behaviors and outcomes (Gulati et al., 2000;
Portes, 1998, Powell, Koput, & Smith—Doerr, 1996; Reagans & Zuckerman, 2001, Tsai,
2002; Walker, Kogut, & Shan, 1997; Zaheer & Bell, 2005) and therefore should be
relevant to the analysis of value creation in strategic alliances. As to the form of network
structures that can appropriately be regarded as beneﬁcial, guidance from the literature is
unclear. Two main conﬂicting theories provide opposite claims on how network structure
affects the relationship between alliances and performance: network closure and
structural holes theory. According to the former view, densely interconnected networks
with many connections between a ﬁrm’s partners, or “network closure” are seen as
advantageous (Ahuja, 2000; Coleman, 1988, 1990; Gulati, 1995;-Gulati & Singh, 1998;
Walker, Kogut, & Shan, 1997; Uzzi, 1997; Zaheer & Venkatraman, 1995). Partners in
“closed” networks can better trust each other to fulﬁll ex-ante commitments or go
through with greater relationship-speciﬁc investments (Walker, Kogut, & Shan, 1997;
Zaheer & Venkatraman, 1995). They are also more likely to develop and accept shared

norms and interaction routines that improve knowledge and information sharing,

18

understanding, and joint problem solving (Ahuj a, 2000; Dyer & Noboeka, 2000; Gulati,
1999; Uzzi, 1997; Walker, Kogut, & Shan, 1997). Closed networks, therefore, generally
reduce opportunistic behavior (Coleman, 1988; Rowley, Behrens, & Krackhart, 2000;
Walker, Kogut, & Shan, 1997).

In contrast, according to structural holes theory (Burt, 1992), actors that can
develop networks of disconnected partners may obtain structural advantages derived from
the brokerage opportunities available to them. Burt (1992) built on the concept of the
“Tertius Gaudens” (i.e., the third who beneﬁts) (Simmel, 1950) to develop the concept of
the structural hole. Structural holes are the borders in social space, gaps in resources or
capabilities (e. g., information) ﬂows between ﬁrms or clusters of ﬁrms that are linked to
a focal ﬁrm but not linked to each other (Burt, 1992). Burt uses the expression “structural
hole”, both for the gap between two entities or clusters of entities and for the ties between
those two entities and a third entity that “bridges” the gap. Some scholars prefer to
separate both into the structural hole itself and the bridging tie (McEvily & Zaheer,
1999). For clarity, I use this last approach through the paper. A structural hole indicates
that ﬁrms or people on different margins of the hole have access to different ﬂow of
resources and capabilities (Ahuja, 2000; Burt, 1992; Hargadon & Sutton, 1997) through
their networks. A bridging tie refers to people or organizations that bridge these structural
holes by obtaining information, capability and control advantages over others and in so
doing gain improved performance.

As stated, the value of structural holes versus network closure remains unclear.
For example, Ahuja (2000) ﬁnds that increasing structural holes has a negative effect on

innovation, and Zaheer and Venkatraman (1995) ﬁnd that network closure allows for

19

greater relationship speciﬁc investments and lower monitoring costs. Alternatively,
Hargadon and Sutton (1997) show how ﬁrms may exploit their position as bridging ties
to develop new products, and McEvily and Zaheer (1999) ﬁnd that structural holes
improve ﬁrms’ ability to acquire capabilities. Also, both Bae and Gargiulo (2004) and
Zaheer and Bell (2005), ﬁnd a signiﬁcant positive relationship between structural holes
and ﬁrm performance.

To solve such contradictory predictions, previous research has relied on a
contingency approach (Ahuj a, 2000; Rowley et al., 2000), or on an interaction effect
between the structural component and the partners’ attributes (Bae & Gargiulo, 2004).
For example, Ahuja (2000) discusses how the contradictory results may be due to
differences in contexts. By contrasting his results with those of Hargadon & Sutton
(1997), he concludes that the explanation of the differences in results may be due to the
fact that his sample, contrary to the non-adversarial nature of Hargadon and Sutton’s
sample, includes alliances between competitor ﬁrms in the same industry. Therefore, the
adversarial nature of his sample may have resulted in a greater need for trust in alliances
and therefore in a positive effect for network closure. Alternatively, Bae and Gargiulo
(2004) looked at the interaction between the importance of a fum’s partners (e.g., the
shares of resources in an industry controlled by alliance partners) and network closure. In
their study, network closure makes the relationship between partner importance and
performance positive and signiﬁcant. Both the contingency view and the exploration of
the interaction between partner attributes and network structure point to the same

solution: namely that closure matters more when there are more reasons for opportunism.

20

In this study, I propose a third approach to address the contradictory predictions in
the literature. Speciﬁcally, I suggest that network closure promotes ﬁrms and overall
network fulﬁllment of ex-ante commitments that may result in value creation. In contrast,
structural holes may be more relevant to the distribution or appropriation of value created
between ﬁrms that are partners in an alliance. Although operationalizations of both
structural holes and closure tend to use the same measures (e. g., density of a ﬁrm’s
alliance network) but with opposite signs in most of the alliances literature (e.g., Zaheer
& Bell, 2005), there exists a fundamental difference in the mechanisms by which both
create advantages for network members.

The structural holes argument describes the value of bridging ties as a function of
brokerage opportunities. Structural holes are thus an opportunity to broker the ﬂow of
resources and capabilities between organizations or people, and to control the projects
that bring together people or organizations from opposite sides of the hole (Burt, 2000).
For example, production studios in the ﬁlm industry may have power in their negotiation
with independent producers due to their access to distribution channels. Therefore, the
value of structural holes can be said to correspond with that of “control over the
negotiation” (Burt, 1992:83) of a transaction. A ﬁrm bridging structural holes has control
and information advantages over its partners that allow it to appropriate more value, give
it bargaining power, or in Burt’s terms determine it’s “share of that proﬁt” (Burt,
1992:83).

Alternatively, closure has two main effects for people in a closed network. First, it
has stabilizing value, in that more consistent ﬂows of resources (e. g., information) are

available to participants in a closed network. For example, Baker (1984) argues that

21

markets with more closed networks between producers stabilize market prices. Second,
and the most emphasized aspect of closed networks is the fact that network closure
facilitates the development of norms and sanctions and makes it less risky for participants
in closed networks to trust one another (Biggart, 2000; Burt, 2000; Coleman, 1988,
1990;). For example, Zaheer and Venkatraman (1995) ﬁnd that network closure allows
for greater relationship speciﬁc investments and lower monitoring costs.

There is an important conceptual difference between bargaining and norms.
Bargaining is agent based and therefore individualistic in nature. Norms are group-based
and therefore less purposeful or agent driven and more socially constructed. Because of
this distinction, structural holes may be more relevant to the distribution or appropriation
of value created between ﬁrms (that I discuss in chapter 3 of this dissertation) through
individual strong bargaining power or brokerage positions. Burt (1992), for example,
considers the situation of two ﬁrms negotiating a transaction and how the relative amount
of holes bridged by each ﬁrm, may result in control over the negotiation of outcomes.
Closure, instead, should be more relevant to the normative effects that promote ﬁrms and
overall network fulﬁllment of ex-ante comrrritrnents that may result in more value
creation for the alliance. Therefore I argue that network closure should promote value
creation at the level of the alliance. I will explore the effect of structural holes in value
appropriation in the next chapter when dealing speciﬁcally with appropriability issues.
My prediction for value creation at the level of the alliance is consistent with ﬁndings
linking closure to greater relationship speciﬁc investments (Walker, Kogut, & Shan,
1997; Zaheer & Venkatraman, 1995), lower monitoring costs (Zaheer & Venkatraman,

1995), and the development of shared norms and routines (Ahuja, 2000; Dyer &

22

Noboeka, 2000; Gulati, 1999; Uzzi, 1997; Walker, Kogut, & Shan, 1997) therefore
improving performance for alliances. Closure at the alliance level, should promote
normative fulﬁllment of ex-ante commitments by organizations and their network of

partners and therefore more value creation at the level of the alliance.

Hypothesis 3: The value created at the alliance level of analysis increases with the

alliance network closure.

Network Size and Alliance Value Creation

Two alternative approaches have been used to explore the relationship between
the number of alliances ﬁrms have and alliances value creation. First, if valuable
resources, such as information (Dyer & Singh, 1998), and unique capabilities (Gulati,
1999) may be inherent within the collaborative relationships between ﬁrms, then more
alliances imply that, other things equal, an organization is more likely to receive more
and less redundant access to external resources and capabilities resulting in strategic
advantage (Burt, 2000). In fact, the extent of a ﬁrm’s alliance network has been often
understood as a result of the organization’s effort to control uncertainty and to bring
together large sets of resources and capabilities (Baker, 1990; Burgers, Hill, & Kim,
1993. For example, Stuart, Hoag and Hybels (1999) ﬁnd a positive relationship between
number of alliances they hold (a control measure in their study) and the performance of
entrepreneurial ventures. Similarly, Goerzen and Beamish (2005) control for network size
and ﬁnd a strong positive relationship between network size and performance in a
multinational alliance sample, and Ahuja (2000) ﬁnds positive relationships between

number of network ties and innovation.

23

Alternatively, both transaction costs economics and, population ecology
perspectives, would suggest inverted-U relationships between the addition of new
alliances to a ﬁrm’s alliance portfolio and value creation in the alliance. Transaction cost
econorrrics (Coase, 1937; Williamson, 1985) argues for a positive relationship between
numbers of alliances and performance but only up to the point at which the cost of
organizing an additional alliance (i.e., to produce a transaction) becomes equal to the cost
of carrying out the same transaction through a market relationship (Williamson, 1985),
and ﬂat to negative thereaﬁer. Population ecology (Harman & Freeman, 1987; 1988;
1989; Harman & Carroll, 1992) makes a similar prediction regarding the expected drivers
of alliance value creation; Resource, legitimacy and capability rich organizations attract
alliance opportunities. As alliances unfold, the resources and the legitimacy of the focal
organization increase but only up to a point. Further increases beyond that point increases
competition for resources, capabilities and managerial attention by partners, therefore
decreasing the value creation of alliances. As Baum and Oliver (1992) note, there is a
carrying capacity to alliances in that most organizations can successfully support only a
limited amount of connections. The argument, similar to that of transaction cost
econorrrics, speciﬁes an inverted u—shaped curve to describe the relationship between
number of alliances and alliance value creation. For example, Goerzen and Beamish
(2005) hypothesize such an inverted u-shaped relationship for the relationship between
the number of unique partnerships a ﬁrm has and performance.

However, alliances, by deﬁnition, have at least two partners, and value is created
in the interaction between these two partners. The implication of this is that resource

constraints in one partner ﬁrm may be compensated by its partner’s slack resources.

24

Therefore, alliance speciﬁc resources needed for value creation can originate from both
sides of the alliance. At the level of the alliance, the interaction between the levels of
commitment to other alliances of both partners is, therefore, the construct of interest to
explore the inverted u-shaped relationship between partners’ commitment to other
alliances and value created at the alliance level of analysis.

Thus, I argue here that value creation at the level of the alliance will be positively
related to the number of alliances each partner holds, up to a point at which capabilities at
the alliance level may become overstretched with respect to the capabilities partners
involved in many alliances can bring into each new alliance. Indirect support for this
argument can be found in Ahuja (2000), who ﬁnds a negative relationship between the
interaction of direct and indirect ties, and performance (e. g., innovation). Direct ties are
the amount of partners a ﬁrm has and indirect ties are the amount of alliances the ﬁrm’s
partners hold. Therefore, I predict an inverse-U shaped relationship between the total
commitment to new alliances by both partners (a dyadic or alliance level explanation)
and alliance value creation. Commitment to new alliances is deﬁned here as the
combined comrrritrnent to new alliances by the partners in an alliance relative to their
overall commitment to alliances. It is a construct that intends to reﬂect the weight of the

partners’ comrrritrnent to new alliances relative to their overall pool of alliances.

Hypothesis 4: The value created at the alliance level of analysis has an inverse-U shaped

relationship with the partners ’ commitment to new alliances.

25

CHAPTER3

THE IMPACT OF PARTNERS’ ATTRIBUTES IN PARTNER’S VALUE

APPROPRIATION: BACKGROUND AND PREDICTION S

In this chapter I adopt a bargaining perspective to explain what the differential
appropriation of alliance value by partnering ﬁrms. As stated in the previous chapter,
valuable resources, such as information (Dyer & Singh, 1998), and unique capabilities
(Gulati, 1999) are inherent in the collaborative relationships that ﬁrms hold with other
ﬁrms, making alliances a source of access to external resources and capabilities to
complement strategic needs (Burt, 2000; Eisenhardt & Schoonhoven, 1996) and create
value. But also, alliances are an attempt to maximize surplus by assigning or reassigrring
a ﬁrm’s resources and capabilities to tasks (Lippman & Rumelt, 2003) for a pay. “All
alliances involve a negotiated bargain between the partners” (Inkpen, 2001: 418). The
size of that payment, the appropriation of value that each ﬁrm derives from an alliance is
the result of a cooperative game (Brandenburger & Stuart, 2001 ; Lippman & Rumelt,
2003) between partners who cooperate and compete simultaneously (Lax & Sebenius,
1986; Hamel, Doz, & Prahalad, 1989).

The appropriation of value by each partner in an alliance is a function of both the
value created in the alliance, as discussed in the previous chapter, and the bargaining
power of each partner to capture a portion of that value. To discuss value appropriation, I
will start by discussing the theoretical antecedents that may help explain the outcomes of
an exchange between partners in an alliance. Exchange theory (Blau, 1964; Emerson,

1962; Homans, 1950, 1974) provides the core theoretical foundation up on which rests

26

my reasoning about value appropriation in alliances. Exchange theory was originally a
dyadic level formulation attempting to explain both the likelihood and the outcomes of
exchanges based on the resources that each part of any relationship, whether it be a
transaction or an alliance, had to offer. Emerson’s (1972b) chapter on exchange relations
and exchange networks introduces social structure as a primary variable in exchange
theory and suggested the need to examine contextual power inﬂuences arising ﬁom the
broader network in which relationships are usually embedded. Speciﬁcally within the
management literature, Resource Dependency Theory (Pfeffer & Salancik, 1978)
extended some of these ideas to the organizational level. By applying exchange theory to
the broader social context, Emerson’s (1962, 1972a, 1972b) work became the antecedent
to what is now commonly referred to as network exchange theory (Cook & Whitrneyer,
1992; Cook & Yamagishi, 1992; Markovsky, Willer, & Patton, 1988; Skvoretz & Willer,
1993; Willer, Walker, Markovsky, Willere, Lovaglia, Thye, & Simpson, 2002;
Yamagishi, Gillmore, & Cook, 1988).

In its latest formulation, network exchange theory argues that bargaining power is
a function of the extent to which actors in a relationship are vulnerable to exclusion from
exchanges within the network. In other words, the bargaining power of a ﬁrm with
respect to a partner in an alliance is related to the likelihood that it will exclude that
partner by selecting another partner ex ante, or exclude it by replacing the partner ex-post
(Markovsky, Willer, & Patton, 1988; Willer, 1999). Bargaining power refers to a
bargainer’s ability to inﬂuence the outcome of a negotiation (Schelling, 1956) or to

obtain beneﬁts from the other party (Dwyer & Walker, 1981; Tung, 1988).

27

When focusing on the dyad, several factors have been identiﬁed as inﬂuencing
bargaining power of the partners. First, the stakes (i.e., the level of dependence) of the
bargainers in a negotiation will help shape their bargaining power (Bacharach & Lawler,
1984). For example, a ﬁrm that is experiencing bankruptcy risk and depends on the
access to resources through a particular alliance to avoid bankruptcy may have little
bargaining power at the time of negotiating the alliance due to its high dependence on the
deal going through. Second, the possession or control of critical resources (Pfeffer &
Salancik, 1978; Harrigan, 1986; Harri gan & Newman, 1990) and the availability of
alternative sources of critical resources to both parties (Bacharach & Lawler, 1984;
Fisher & Ury, 1981; Root, 1988:76) will also help determine their bargaining power.
Third, the position in the exchange network may provide bargaining power to a person or
organization, like in the case of organizations that occupy structural holes (Burt, 1992). In
this chapter, I develop hypotheses for how the relative strategic stakes, availability of
alternative partners, partners’ capabilities, and the structural positions of the partners in
an alliance, may help determine the bargaining power of alliance members in

appropriating value from alliances relative to their partner.
Relative Strategic Stakes and Firm Value Appropriation

A high payoff for cooperation is likely when ﬁrms are in vulnerable strategic
situations (Eisenhardt & Schoonhoven, 1996). When ﬁrms are in difﬁcult market
situations or are engaging in large or particularly risky investments (Conner, 1994)
alliances can provide needed resources and capabilities (e. g., cost sharing, legitimacy,

knowledge). Empirical evidence shows that such situations promote the development of

28

strategic alliances. For example, Shan (1990) found that entrepreneurial ﬁrms in highly
competitive environments were more likely to enter more alliances, and Singh (1992)
found that the vulnerability of late entrants into the medical diagnostics industry was
positively associated with alliance formation.

Vulnerable strategic situations raise the strategic stakes for ﬁrms intending to
engage in alliances that may support their position. I defrne relative strategic stakes as the
relative importance of the alliance to a focal partner with respect to the importance of the
alliance to the other partner involved in the alliance. Firms with high strategic stakes will
have lower bargaining power in alliances because they are relatively more dependent on
the alliance (Bacharach & Lawler, 1984; Yan & Gray, 1994). But the effect of strategic
stakes depends on the strategic stakes that the alliance implies for both partners involved
in forming the alliance. Therefore, bargaining power will depend on the relative strategic
stakes on ﬁrms intending to organize an alliance. That is, when alliance partners have
unequal stakes they have unequal bargaining power. The higher the strategic stakes
implied by an alliance in a particular market situation for one of the potential partners
with respect to the other, the lower its bargaining power, and therefore the weaker it’s
bargaining ability to appropriate value in the alliance.

Therefore, I hypothesize that higher relative stakes in an alliance for a partner
with respect to the other will result in lower relative appropriation of value in the alliance

for that partner

Hypothesis 5: The value appropriated from an alliance by a partner will be negatively

associated with the relative strategic stake of that partner.

29

Relative Partner Capabilities and Firm Value Appropriation

Resource dependence theory (Pfeffer & Salancik, 1978) suggests that the
possession and control of key resources and capabilities by one ﬁrm in an alliance may
make that ﬁrm’s partner dependant on that ﬁrm. If a ﬁrm brings more critical resources
and capabilities to an alliance than its partner, it’s option of contributing or withholding
its cooperation in the alliance can be used by the ﬁrm as leverage in bargaining with its
partner (Pfeffer, 1981), therefore shaping the distribution of value between the partners.
The previous hypothesis about strategic stakes and the resource dependence theorizing
that supports this section may seem similar in nature but there is a basic difference
between both approaches. Strategic stake refers to the implicit value of the alliance to a
ﬁrm given the frrm’s situation and refers to bargaining implications of a ﬁrm’s situation
independent of the value it brings to an alliance. Resource dependence refers to what
each partner contributes to the alliance and therefore to proportionality between value
drivers and payments in a relationship. For example, a ﬁrm going into a partnership to
distribute its products in another region through a partner that has the best distribution
network in that region is dependent on that partner for such an extensive distribution and
therefore may have to forgo more margin in the deal.

The distribution of value between partners is determined by what each brings to
the alliance (Harrigan, 1986; Schelling, 1956; Yan & Gray, 1994). I use the expression
relative capabilities here to refer to the relative capabilities that one ﬁrm involved in an
alliance has with respect to the total capabilities of both ﬁrms involved in the alliance.

Therefore, I argue that the relative capabilities and resources between ﬁrms involved in

30

an alliance should be positively related to their individual value appropriation from the

alliance.

Hypothesis 6: The value appropriated from an alliance by a partner will be positively

associated with the relative capabilities of that partner.

Relative Availability of Alternative Partners and Firm Value Appropriation

Several theoretical traditions suggest that the availability of alternative partners
with access to the quality and amount of resources and capabilities needed by a ﬁrm is
another important determinant of a ﬁrm’s bargaining power in an alliance. In other
words, it is not only the need for a partner’s resources but also the absence of alternative
sources for those resources that determines the bargaining power of the partners in an
alliance. This view is core to all periods and fashions of exchange theorizing including
Social Exchange Theory (Emerson, 1971, 1972), bargaining theory (Bacharach &
Lawler, 1981; Lawler & Bacharach, 1979), resource dependence theory (Pfeffer &
Salancick, 1978), and more recently in networks exchange theory (Willer, 1999).

The essential argument is that if a partner of a focal ﬁrm has control over critical
resources and capabilities, and those resources and capabilities are difﬁcult to replace for
the focal ﬁrm because of the lack or shortage of alternative partners, that partner will be
able to appropriate extra value from the alliance (Root, 1988; Bae & Gargiulo, 2004;
Fisher & Ury, 1981; Yan & Gray, 1994). A ﬁrm that has more alternatives with which to
partner, relative to the alternatives available to their alliance partner has more power in
the relationship and should appropriate more value ﬁ'om the alliance. The Relative

availability of alternative partners is deﬁned as the availability of alternative partners for

31

a ﬁrm relative to the availability of alternative partners for its partner in the alliance.

Therefore:

Hypothesis 7: The value appropriated from an alliance by a partner will be positively

associated with the relative availability of alternative partners.

Relative Structural Holes and Firm Value Appropriation

In Chapter 2, I presented both structural holes and network closure arguments and
discussed the contradictory predictions and ﬁndings that the literature presents with
regards to these structural descriptions. Brieﬂy stated, according to network closure view,
densely interconnected networks with many connections between a ﬁrm’s partners, or
“network closure” are seen as advantageous (Ahuja, 2000; Coleman, 1988, 1990; Gulati,
1995; Gulati & Singh, 1998; Walker, Kogut, & Shan, 1997; Uzzi, 1997; Zaheer &
Venkatraman, 1995). Partners in “closed” networks can better trust each other to fulﬁll
ex-ante comrrritrnents or go through with greater relationship-speciﬁc investments
(Walker, Kogut, & Shan, 1997; Zaheer & Venkatraman, 1995), develop and participate in
shared norms and interaction routines that improve knowledge and information sharing,
understanding, and joint problem solving (Ahuja, 2000; Dyer & Noboeka, 2000; Gulati,
1999;Uzzi, 1997; Walker, Kogut, & Shan, 1997), and, in general, reduce opportunism
(Coleman, 1988; Rowley, Behrens, & Krackhart, 2000; Walker, Kogut, & Shan, 1997).

Alternatively, according to the structural holes view, actors that can develop
networks of disconnected partners may obtain structural advantages derived from the
brokerage opportunities available to them (Burt, 1992). A structural hole indicates that

ﬁrms or people on different margins of the hole have access to different ﬂow of resources

32

and capabilities (Ahuja, 2000; Burt, 1992; Hargadon & Sutton, 1997) through their
networks. People or organizations that bridge these structural holes may obtain
information, capability and control advantages over others.

As stated, the value of structural holes versus network closure in the alliance
literature remains unclear, with empirical support for both the network closure (Ahuja,
2000; Zaheer & Venkatraman, 1995) and the structural holes arguments (Hargadon &
Sutton, 1997; McEvily & Zaheer, 1999; Bae & Gargiulo, 2004; Zaheer & Bell, 2005). As
noted earlier, a potential resolution to this contradiction results ﬁ'om a better
understanding the mechanisms implied by both arguments and their implications at
different levels of analysis. Speciﬁcally, I hypothesized that network closure and not
structural holes should be more relevant to value creation at the level of the alliance
where the normative effects of a closed network may promote ﬁrms and overall network
ﬁrlﬁllment of ex-ante commitments. In contrast, I suggest that structural holes may be
more relevant to the appropriation of value created by ﬁrms participating in the alliance
through individual strong bargaining power or brokerage positions.

Therefore, the relative amount and importance of the structural holes that a focal
ﬁrm in an alliance has with respect to a partner will be related to its value appropriation
from the alliance. In line with social exchange (Emerson, 1972) theorizing and research,
ﬁrms that mediate the access of partner ﬁrms to more external capabilities (i.e., ﬁrms that
bridge more structural holes relative to a partner) can derive bargaining power from their
position in the overall network structure that surrounds the alliance. Also, structural holes

theorizing (Burt, 1992) supports the idea that there are control and information

33

advantages available to ﬁrms that bridge structural holes, which provide those ﬁrms with
control over the negotiation of the share of the value created through the alliance.

But as holes around partners provide information and control beneﬁts that result
in negotiation power for a focal ﬁrm, holes around the focal ﬁrm can also beneﬁt
potential partners in the negotiation. Higher negotiation control for a ﬁrm depends on
having numerous structural holes around partners but none attached to itself. Therefore, it
is important to consider the availability of structural holes a ﬁrm has around an alliance
with respect to the holes it’s partner has around that ﬁrm. Said another way, the ﬁrm that
bridges more structural holes in an alliance should have superior bargaining power.
Therefore, ﬁrms with more structural holes around an alliance relative to their partners in

the alliance will appropriate more value in alliances.

Hypothesis 8: The value appropriated from an alliance by a ﬁrm in the alliance will be

higher for ﬁrms that bridge more structural holes relative to their partner.

34

CHAPTER 4

METHODS

Data and Sample

I used four criteria in selecting alliances for this study. First, I restricted the
population of interest to public ﬁrms (Bae & Gargiulo, 2004). I chose to focus on
publicly traded ﬁrms to guarantee consistent criteria in the ﬁnancial data utilized in the
study. Second, each ﬁrm selected from the population had to have engaged in at least one
marketing alliance with another public ﬁrm between 1988 and 2002. Considering 15
years provides sufﬁcient length of time to cover different macroeconomic environments,
providing generalizability to other periods, and to capture multiple alliance events by
most ﬁrms in the sample. I included information for all ﬁrms in the sample for up to four
years before the period of interest to control for possible left-censoring of the data and
forward up to the latest data available for 2004 so as to allow for lagged performance
effects. Therefore, the population of interest included data ﬁ'om 1984 to 2004. Third,
given my research questions, it was necessary to deﬁne an initial population of ﬁrms that
operate in industries where alliances are generally considered an important element of
ﬁrm strategy. The following industries have been identiﬁed in previous research as
falling into this category: telecommunications, chemicals, electronics, and services
(Culpan & Eugene, 1993; Kale, Dyer, & Singh, 2002; Alliance Analyst, 1996) Industry
categories were identiﬁed on the basis of SIC codes (e.g., sic code 357 for computer
hardware). Fourth, I utilized a two wave snowball sampling design (Capobianco & Frank,

1982; Frank, 1979, 2005) to develop a random sample of alliances for the study.

35

Snowball designs are of particular interest in network research. Conceptually, a wave of
snowballing implies determining an initial random sample of individual ﬁrms, called a
seed sample, and then identifying all their direct relationships or partners that are not part
of the seed sample. The ﬁrst wave sample is then the sum of all the individuals in the
seed sample plus their direct relationships. A second wave identiﬁes the direct
relationships or partners of the ﬁrms identiﬁed in the ﬁrst wave. Thus, the ﬁnal sample
includes the initial random selection of ﬁrms, all their alliance partners, and all the
partners of these partners. Speciﬁcally, I did the following. First, I selected a random
sample of alliances in which at least one ﬁrm was included in my basic set of industries
for the study and identiﬁed those ﬁrms as my seed sample for the snowball sampling
scheme. Second, I looked for all their alliances in the period of interest with ﬁrms that
fulﬁll the criteria for being in the population of interest. I aggregated the ﬁrms in those
ﬁrst two groups and then looked for all the alliances the ﬁrms in this new data set have
for each period in which they are part of the sample and restricted to those alliances that
fulﬁll my ﬁrst two criteria (i.e., public and having a marketing alliance in the period) for
being in the population of interest. The ﬁnal sample, then, includes all the alliances the
sample members have that fulﬁll the criteria for each period in which they appear in the
sample. Formally, letting So be a random seed sample chosen from a population of
interest, S1 be a sample after snowball 1, and S; be the sample after the second wave of

snowballing, we can describe it as:

S1=SoUA(51)Sr¢So

$2 = SIUA(52)provided 52 ¢ 51, where

A(S1)= {j e A(Sl)iff Zy- =1 for somei 6 SO} and
A(Sz)= {i e A(Sz)iff Zij =1 for somei 6 St}

36

where A(S,,)is the sub-sample originated in a snowball wave and Zij =1 implies the

existence of a tie between two ﬁrms in the population.

The resulting sample includes a total of 495 alliances among the ﬁrms sampled.
This data was organized in a matrix representing the existence or absence of an alliance
between each pair of ﬁrms for each year in which they are included in the sample. Data
on alliance and ﬁrm characteristics was collected from two primary sources: Mergers,
Acquisitions and Alliances’ database of the Securities Data Corporation (SDC Platinum)
provided by Thompson Financial and COMPUSTAT by Standard and Poors. The SDC
Platinum data has been used extensively to study alliances (Bae & Gargiulo, 2004;
Casciaro & Piskorski, 2005; Sampson, 2005) and maintains an extensive list of reported
alliances for all years beginning in 1988. Two limitations need to be noted about this data
source. First, companies do not necessarily report all alliances they engage in (Kale,
Dyer, & Singh, 2002), and SDC is likely to under represent small private organizations
due to its public sources of information (e. g., corporate gazettes, business journals)
(Anand & Khanna, 2000). Alliances that are not reported by companies due to tactical or
strategic reasons are not likely to be quoted in the press or elsewhere and are therefore
beyond the scrutiny of this research. Also, since my sample is restricted to alliances
between public ﬁrms, the potential under representation of alliances with small partners
is not likely to have any signiﬁcant inﬂuence in the results of this research. It does
however limit generalizability to alliances between public ﬁrms. Second, SDC is likely to
report the same alliance multiple times (Bae & Gargiulo, 2004). To minimize reporting
errors I crosschecked SDC reports with those of Lexis-Nexis, and verify the nature and

existence of reported alliances.

37

Constructing the dataset for this study presented several challenges (e. g.,
integrating data from multiple sources, constructing the networks). First, I used CUSIP
information for integrating the alliances and ﬁnancial data. Second, I developed yearly
adjacency matrixes or sociograms (Moreno, 1938; Wasserman & Faust, 1994). An
adjacency matrix is a two way matrix or sociogram for a dichotomous relationship. I
created matrixes that include data for each alliance and its partners for each alliance event
in the year it starts. Data was organized to represent alliance (pair level) factors, and
partner level attributes in three separate sets, on for alliances and one for each partner in

an alliance.

Measurement

In this section I describe the variable used in the study. Table 1 summarizes the

variables and their formal measurement.

38

Table 1: Variable Names and Descriptions

 

 

(alliance network l
overall constraint)

Commitment to COM
alliances k
Relative Stakes4 P ASTKPRD

 

 

 

Level 1 (Pair)

Level 1 (Pair)

Level 1 (Pair)

Level 1 (Pair)

Level 1 (Pair)

 

Variable name Label Level Variable Description

Capabilities Cp Level 2 (ﬁrm) Market Share

Tie strength 1 TSl Level 1 (Pair) Dummy denoting that partners were partners
in the past 4 years

Tie strength 2 T82 Level 1 (Pair) Number of actual alliances between the
partners

Network Closure CLOSE Level 1 (Pair) Count of complete minimum sub-graphs with

a maximum of four nodes in the overall
alliance network

(Count of new alliances for Firm i * Count of
new alliances for F inn 1") / (Count of total
alliances for Firm i "' Count of total alliances
for Firm i’)

Ordinal variable where:

2: indicates performance above the
Aspiration level1 for Firm i versus
Bankruptcy risk2 by Firm i'.

1: indicates either performance below
aspiration level by F irm i versus
Bankruptcy2 risk by Firm i’ or
Performance above aspiration level1 by
Firm i versus performance below

. . 3 . .,
aspiration level by Frrm l .

0: indicates similarity in performance
situation between partners.

-1: indicates either Bankruptcy1 risk by
F irm i versus performance below the
. . 3 . .,
aspiration level by F um i or
performance below the aspiration level3
by Firm i versus performance above the
aspiration level1 by Firm i’.

 

 

l Z-scores above the lowest 10% of the sample and lagged performance at or above their aspiration level

Glam )< (KO/1m)

2 Z-score in the lowest 10% of sanrple

3 Z-scores above the lowest 10% of the sample and lagged performance below their aspiration level

(ll—071m)> (R0413)

4 Alternatively will consider a similar measure based on Industry average performance.

39

 

 

 

 

Variable name Label Level Variable Description
Level 1 (Pair) -2: indicates Bankruptcy risk2 by Firm i vs.
performance above the aspiration level1
by Firm i ’.
Relative capabilities ALLCZD Level 1 (Pair) (Sales Firm i ' Sales Firm 1") / (Sales Firm i +
size Sales Firm i’)
Relative capabilities ALLCPD Level 1 (Pair) (Market Share Firm i " Market Share Firm i 'i /
(Market Share Firm i + Market Share Firm i’)
Relative availability RAAp Level 1 (Pair) (2 Market share of ﬁrms in same industry as
ofalternatives Firm i’ with same or larger market share than
Firm 1' ’ (RAMSHi) - 2 Market share of fu'ms
in same industry as F inn i with same or larger
market share than Firm i (RAMSHi’) )/
(RAMSHi + RAMSHi’)
Relative structural RSHd Level 1 (Pair) (Count Of ﬁrm’s partners 110‘ shared (through
h0]es-direct (directed direct tICS) With Firm 1') minus (Count Of Firm
betweenness i ”s partners not shared (through direct ties)
centrality with direct with Firm 2')
constraint)
Controls:
Tie ALL Level 1 (Pair) Dummy indicating a tie between partners
Assets ASST Level 2 (ﬁrm) Assets
Advertising ADVexp Level 2 (ﬁrm) Advertising expense
expenditures
R&D Expenditures R&D Level 2 (ﬁrm) R&D Expense
Lagged DV PredP Level 1 (Pair) Lagged DV for each of the models
Leverage LVGE Level 2 (ﬁrm) Debt/equity ratio
Alliance type
Joint Venture JV Level 1 (Pair) Dummy variable indicating Joint Ventures
Supplementary CA Level 1 (Pair) Dummy Variable indicating partners operate

 

 

 

in the same industry

 

Note: All ﬁrm level variables are specified for both partner i and partner i' in each alliance.

4O

 

. Dependent Variable

Although it is widely recognized that ﬁrms pursue multidimensional goals, and that
“performance,” “value creation” or “value appropriation” can take many meanings for
individuals and organizations, the mainstream view in strategy research, and the focus of
this study, is that ﬁrms are fundamentally concerned with economic results (Goerzen &
Beamish, 2005). A review of alliance studies revealed a set of measures that have become
well accepted in the literature: sales (e. g., Stuart, 2000), and proﬁtability expressed as
return on assets (ROA), return on investment (ROI), or return on Sales (ROS) (e.g., Bae
& Gargiulo, 2004; Goerzen & Beamish, 2005). In this study I use the net income
component of the proﬁt measures as my dependent variable (assets are a control in the
models). Net Income is the proﬁt from all the company's operations, calculated as total
revenues minus total expenses and losses, as reported in the Compustat database.
Analyses at the dyadic level use linear aggregation of individual ﬁrm level net incomes
for both ﬁrms in an alliance as the dyadic unit (i.e., the alliance) dependent variable. The
dependent variables for both models are deﬁned at the dyadic level. Model 1 uses the
linear aggregation of individual ﬁrm level net incomes for both ﬁrms in an alliance as the
dependent variable. Model 2 uses the difference in net Income between ﬁrms included in

the alliance as its dependent variable.

Explanatory Variables

Table 1 summarizes the description of the explanatory variables in this study. I
brieﬂy describe here some of the variable decisions that I make in the study. I argue that

partners with more and better capability endowments with respect to a marketing alliance

41

(e.g., ﬁrms with extensive market coverage or ﬁrms that have a demonstrated ability to
outsell their market) should be better partners with respect to value creation. Two
measures of partners’ capability endowments that have been used in the alliance literature
are Sales (e.g., Stuart, 2000) and Market Share (e.g., Bae & Gargiulo, 2004). I use market
share as a measure of capability endowments to reﬂect the market effectiveness of a
partner compared to others in its particular industry. When a ﬁrm enters a marketing
alliance with a partner that holds large market share in its own market, the ﬁrm gains
access to a powerful market position, and therefore the opportunity to access that
partner’s market.

Tie strength reﬂects the level of comnritrnent, frequency of interaction and trust
implicit in a relationship. Therefore, following guidance from the literature I
operationalize tie strength as a measure of the number of alliances (for any purpose) that
partners have between them concurrently with the actual marketing alliance observation.

I measure commitment to new alliances with the product of the counts of new
alliances by each ﬁrm scaled by the product of the total number of alliances each ﬁrm

has. The actual measure is:

Count of new alliances for partner a x Count of new alliances for partner b
Count of total alliances for partner a x Count of total alliances for partner b

 

This measure of commitment to new alliances represents the stress new alliances
apply to the carrying capacity of the ﬁrm. I then create a squared version of the variable
to model the expected inverted-u shaped relationship between total commitment to

alliances at the alliance level and value creation.

42

For alliance network closure, I use an overall alliance network constraint function
based on constrains on entrepreneurial opportunities. The entrepreneurial opportunity of a
ﬁrm i is constrained (e.g., normative pressures emerge) to the extent that (a) a partner q
of a ﬁrm is also (b) a partner of partner i’ of the ﬁrm (Burt, 2000). This implies that
constraints emerge from closed subgraphs. I therefore measure alliance network closure
as aggregate constraint for the pair’s network as a whole. I deﬁne an alliance network as
the set of alliances in which each or both ﬁrms in an alliance are involved. Speciﬁcally,
the measure for closure is the count of complete minimum sub-graphs with a maximum
of four nodes in the overall alliance network or the amount of “closed” triangles
(Newman, Watts, & Strogatz, 2002) or quadrangles in the network.

Relative capabilities performance is measured as (F inn.- Market Share -Firm ,-v
Market Share) / (Firm ,- Market Share + Firm ,--Market Share) and is a measure of the
relative market power potential a partner brings to an alliance. Both measures are scaled
to be complementary between partners.

For relative stakes I ﬁrst develop a categorical variable based on the relative risk
situations of the partners. March and Shapira (1992) model risk as performance
variability. Since risk is a possible measure of the strategic stakes involved in a strategy
choice (e. g. an alliance) for a ﬁrm, I ﬁrst follow Miller and Chen (2004) and March and
Shapira (1992) in their categorization of ﬁrm risk into three levels: (a) ﬁrms threatened
by bankruptcy (TB), (b) ﬁrms not directly threatened by bankruptcy but performing
below their aspiration levels (LP), and (c) ﬁrms not directly threatened by bankruptcy and

performing at or above their aspiration levels (HP). For the threat of bankruptcy risk, I

43

use Altman’s Z (Altman, 1983; Deephouse & Wiseman, 2000; Miller & Reuer, 1996).

The calculation of Altman’s Z is:

(1.2 X working capital divided by total assets) +

(1.4 X retained earnings divided by total assets) +

3.3 X income before interest expense and taxes divided by total assets) +
(0.6 X market value of equity divided by total liability) +

1.0 X sales divided by total assets)

where lower Z-scores imply higher likelihood of bankruptcy. Following Miller
and Chen (2004) I consider ﬁrms with Z-scores in the lowest 10% of sample to be ﬁrms
in the threatened by bankruptcy category (TB). F inns in the remaining 90% of the sample

are then categorized based on whether their lagged performance (ROAi, t-1) was below

(LP) or above (HP) their predicted performance (Mn — 1) based on Bromiley’s ( 1991)
lagged performance plus growth factor which Miller and Chen (2004) suggest as a
reasonable proxy for aspiration level. Alternatively, I develop a similar measure looking
at comparisons to their industry performance as the aspiration level point of reference.
Second, I deﬁne an ordinal ﬁve point variable to represent the relative stakes
between partners in an alliance. The ﬁve values are deﬁned to represent a measure of the
relative stakes a ﬁrm experiences in forming an alliance. Therefore, each value reﬂects
the distance in terms of the importance or stakes involved in the alliance for each ﬁrm in
an alliance with respect to the other ﬁrm in the alliance. Therefore, two ﬁrms that are in
similar situation (e.g., both are operating above their aspiration level, or both are in
bankruptcy risk) may not have a bargaining advantage due to their stakes in the alliance
with respect to each other (i.e., a value of 0). Alternatively, an alliance between two ﬁrms

that are in very different situations (e.g., a ﬁrm operating above its aspiration level with

respect to a ﬁrm in bankruptcy risk) will imply an important bargaining power difference
for the ﬁrm that is in less need of entering the alliance. The values are created so as to
reﬂect that distance based on the three groups that were defmer previously (i.e., ﬁrms
performing above their aspiration level (HP), ﬁrms performing below their aspiration
level (LP), and ﬁrms threatened by bankruptcy (TB). Table 2 provides a visual
representation of the values for this variable. In this ordinal variable, 2 represents HP
ﬁrms in alliances with TB ﬁrms (i.e., the most positive relative stakes); 1 represents HP
fums in alliances with LP ﬁrms or LP ﬁrms in alliances with TB ﬁrms; 0 represents ﬁrms
that are allied to ﬁrms with a similar level of risk (e.g., HP allied to HP); -1 represents TB
ﬁrms allied to LP ﬁrms or LP ﬁrms allied to HP ﬁrms; ﬁnally, -2 represents TB ﬁrms

allied to HP ﬁrms.

Table 2: Parameter Development for the Relatives Stakes Measure

 

 

 

 

 

HP LP TB
HP 0 1 2
LP -1 0 1
TB -2 -1 0

 

 

 

 

 

Relative availability of alternative partners is a construct that reﬂects the
substitutability of a partner for a ﬁrm. It has been previously measured as the proportion
of a ﬁrm’s partners that were incumbent in the ﬁrm’s industry in a given year (e.g., Bae
& Gargiulo, 2004, in a study of alliances between competing ﬁrms in one industry).
Since my sample includes ﬁrms in multiple industries and my model looks at alliances I

focus on the existence of similarly or more capability endowed alternative partners within

45

a partner’s industry. If a ﬁrm has the alternative to partner with alternative ﬁrms in the
potential partner’s industry that are at least equally capability endowed than the potential
partner, then, that partner has a risk of exclusion. The measure compares the relative risk

of exclusion of both ﬁrms in the alliance. The speciﬁc measure is:

(2 Market share of ﬁrms in same industry as Firm i’ with same or larger market share
than Firm i’ (RAMSHi) - 2 Market share of ﬁrms in same industry as Firm i with same
or larger market share than Firm i (RAMSHi ’)) / ( RAMSHi + RAMSHi’)

Relative measures of Structural Holes (Burt, 1992, 2000) are created comparing

the bridging of structural holes for ﬁrmi with respect to ﬁrm iv, to the bridging of

structural holes for ﬁrm ,vwith respect to ﬁrm ,- (Burt, 1992). The speciﬁc measure is:

(Count of ﬁrmi 's partners not shared (through direct ties) with ﬁrm”) -
(Count of ﬁrm” '3 partners not shared (through direct ties) with ﬁrmi)

Control Variables

Control variables are included at both the ﬁrm and alliance levels of analysis to
control for ﬁrms performance not related to the alliance and to control for alternative
explanations for the relationships studied in our models. The complete list is included in
Table 1. Beyond the traditional controls for ﬁrm performance, several control variables
are included to account for relationships that previous alliances literature suggest as
relevant to the relationships studied in this paper.

Assets for each of the ﬁrms were included in the models as a regressor by itself
and in interaction with the ﬁrms having an alliance to control for ﬁrms’ size effects at

level 2. Research and Development were included to control for the potential effect of

46

R&D investments in technological capabilities that might affect a focal ﬁrm’s
proﬁtability at level 2. The Lagged Dependent Variables are included in both models as a
regressor to control for previous performance effects that are independent of the alliance
under study. Same Industry Dummy Code is a dummy coded variable indicating that both
ﬁrms in the alliance participate in the same industry. I used this control to indicate that
the alliance reﬂects supplementary joint resources and a joint competitive relationship
between the ﬁrms in the alliance. Alliance type is dummy variable indicating joint
ventures, which I used to control for the type of alliance in the model. Leverage is
measured as the ratio of debt to equity. This allowed me to control for potential inﬂuence
of the capital structure on ﬁrms’ performance (Buhner, 1987; Goerzen & Beamish, 2005;

Jensen, 1989)
Analysis

Hypotheses l to 4 relate ﬁrm and alliance speciﬁc factors to value creation at the
alliances level. Hypotheses 5 to 8 relate alliance speciﬁc factors to value appropriation at
the alliances level. The questions addressed in these hypotheses present two main
statistical issues that have recently received considerable attention in the networks
methods literature: dependencies among observations and modeling multiple levels of
analysis simultaneously. Understanding the multilevel implications of networks (Jones,
Hesterly, & Borgatti, 1997; Monge & Contractor, 2003) and simultaneously addressing
the fact that network (e. g., alliances) data is relational, and therefore composed, by
deﬁnition, of nonindependent observations (Holland & Leinhardt, 1981; Lazeaga & Van

Duijin, 1997; Monge & Contractor, 2003 ; Van Duij in, 1995; Wasserman & Pattison,

47

1996) implies that “standard” statistical approaches that assume independent subjects,
such as regression analysis and ANOVA are not appropriate. Therefore, I rely on a novel
set of statistical approaches to test my hypotheses. These techniques address the issues of

independence and allow me to study the multilevel implications of networks.
Value Creation Model

My ﬁrst conceptual model is a multilevel selection model with predictions at the
dyadic level. Multilevel methods in networks analysis have only recently been developed
and have basically addressed the issue of selection (who selects whom for a relationship
of any sort), with individual level (usually binary) dependent variables. I build on the
logic of such models (e. g., the p 1 model (Holland & Leinhardt, 1981), the p. model
(Wasserman & Pattison, 1996), and the p; model (Van Duijin, 1995; Lazeaga & Van
Duijin, 1997)) to develop a multilevel performance model (MPM) fully estimable with
maximum likelihood multilevel algorithms and designed to deal with unbalanced data to
account for the possibility that some alliances may not have been reported. Essentially,
this MPM is a cross-nested selection model for continuous DVs. My model for value

creation is formally deﬁned as:
Yii'i +1 = Yi'it +1 = ,u + ait + ai'i + 6ii’t + eii't +1 for i’=1, #1, to n. (1)

where Yii'i+1 is the performance at the dyad (alliance) level at time t+1, p is a mean,

ait is a vector containing ﬁrm level effects for one of the partners in the alliance and
am is a vector containing ﬁrm level effects for the other partner in the alliance, and

6ii'i is a vector containing dyadic level effects (i.e., alliance —pair- level effects). Note

48

that Yii'i +1 = Yi'ii +1 since this is not a directed model and therefore the alliance has one
performance (i.e., the aggregate performance of both ﬁrms in the alliance) regardless of
the ordering of the partners in the matrix.

At the ﬁrm level (individual partners’ level) the model is further speciﬁed so that

vectors for individual actors are expressed as:

air = Fiyi
and (2)
ai't = Fi'yi

where Ti and Fi' are matrices with covariates for partner ﬁrm’s effects with coefﬁcients
yin and yt'n .

Similarly, at the alliance level (pair or dyadic level) the model is further speciﬁed
so that vectors for dyadic attributes are expressed as:

6ii'i = Zii'gii' (3)
where Zii' is a matrix with covariates for dyadic (alliance) effects with coefﬁcient. The

actual models at the pair and individual levels are:

Alliance (pair) level effects:

61'1”: = (ii'1(Alliance Capabilitiesi'i't) + {it '2(Alliance Tie Strengthii 't) +

(ii'3(Alliance Capabilitiesii't x Alliance Tie Strengthii't) +

{ii'4(Alliance Closureii't) + €ii'5(Change in commitment to alliancesii't) +
{ii'6(Change in commitment to alliancesii't2)+ (4)

{ii'7(Market Shareit x Alliance T ieii't)+ Cit'8(Market Sharei't x Alliance T ieii't)+

11
Z (ii'n (Controlsii'i)
n=9

49

Individual level effects:

6
air = Z yin(ControLS'ii) (5)
n=1
6
ai'r = Z 7i'n(ControlSi't) (6)
n=1
Appropriation Model

The appropriation model is designed to explore the explanations for value

appropriation at the level of the individual ﬁrm. The model is speciﬁed as:

AYti'i +1 = ,u + air + ai’t + 6ii'i + eii'r +1 for i ’=1, i ’qél, to n. (7)
where AYii'i +1 is the difference in performance between ﬁrm we and ﬁrm i’ (i.e.,

alliance level) at time t+1, ,u is a mean, an is a vector containing ﬁrm level effects for

one of the partners in the alliance and am is a vector containing ﬁrm level effects for the
other partner in the alliance, and 6a”: is a vector containing dyadic level effects (i.e.,
alliance —pair- level effects).

At the ﬁrm level (individual partners’ level) the model is further speciﬁed so that

vectors for individual actors are expressed as:

air = Fiyi
and (8)
ai': = Fi'yi

where Pi and Fr" are matrices with covariates for partner ﬁrm’s effects with coefﬁcients

yin and yi'n .

50

Similarly, at the alliance level (pair or dyadic level) the model is further speciﬁed

so that vectors for dyadic attributes are expressed as:
6ii'i = Zii'gii' (9)

where 2a" is a matrix with covariates for dyadic (alliance) effects with coefﬁcient. The

actual models at the pair and individual levels are:

Alliance (pair) level effects:

5ii 't = (ii ' 1(Relative Stakesii 't) + {it ' 2(Relative Capabilitiesii 't) +

4‘; ii '3(Relative Availability of alternative partnersii 't) +

(2)
7
9’ ii '4(Relative Structural Holesii 't) + 2 C ii 'n (ControlSii 't)
n=5
Individual level effects:
12
air: 2 yin(ControlSit) (3)
n=1
12
ai'r = Z 7in(ControlSi't) (4)
n=1

Results for these models are described and explained in Chapter 5. In the next chapter I

present and discuss the results for this dissertation.

51

CHAPTER 5

RESULTS

Tests for the hypothesis are presented in Table 3. In Table 3, the ﬁrst column is
the number of the hypothesis being tested, the second in the name of the explanatory
variable in that hypothesis, the third column is the value of the coefﬁcient for the
hypothesis to be in the right direction and the last column is the test of signiﬁcance for
the hypothesis. The appropriate test for all single parameter tests (Raudenbush & Bryk,
2002) is based on the T-ratio; that is, I compute the ratio of the estimated coefﬁcient to its
estimated standard error. Formally, this statistic is asymptotically unit normal and under
large-sample theory is tested against the Z distribution. Therefore, values of lT-ratio | Z 2
are considered signiﬁcant for the models studied in this dissertation. Tables 4 and 5
present my ﬁndings for the predictions studied in this dissertation. In the next paragraphs
I present the results for each of the hypotheses included in this dissertation. Overall,
Model 2 in Table 4 shows support for hypotheses 2B and 3, and Model 4 in Table 5

shows support for hypotheses 5, 7 and 8 of this dissertation.

52

Table 3: Tests for Hypotheses

 

 

 

 

 

 

 

 

 

 

 

 

Hypothesis Variable Coefﬁcient TEST
, :1:
H1 a Market Share, 7“ > 0 T-ratio Z 2
Alliance T 1e
Market Sharer * y“ > 0 T-ratzo Z 2
Alliance Tie
Hlb Alliance Capabilities gt. .1 > 0 T -ratio 2 2
H2a Tie Strength {ﬂ .2 > 0 T -ratio _>_ 2
H2b Alliance Capabilities Cit? > 0 T -ratio 2 2
* Tie Strength
H3 Alliance Closure €ii'4 > 0 T ~ratio Z 2
H4 Commitment to New Cali .5 > 0 T -ratio 2 2
Alliances
T -ratio < -2
< -
Squared C“ 6 0
Commitment to New
Alliances
H5 Relative Stakes 1,-1- .1 > 0 T -ratio 2 2
H6 Relative Capabilities 1,-1.7 > 0 T -ratio 2 2
H7 Relative Availability 11.1.3 > 0 T -ratio 2 2
of Alternative
Partners
H8 Relative Structural Tii’4 > 0 T -ratio _>_ 2
Holes

 

 

 

 

53

 

Table 4 presents the results for the hierarchical cross-nested linear analysis for the
value creation models. Model 1 in Table 4 shows the results for a model that includes all
the controls of the value creation model. In line with the strategic management theory,
control variables that were included to account for previous performance, and size and
capability direct effects on performance independent of the alliance, were signiﬁcant and
in the expected positive direction. Joint venture and supplementary alliance dummies do
not seem to make a difference on the predictions of the model. Model 4 in Table 5 shows
the results for the full value creation model.

Hypotheses 1A and 1B are not supported. In HlA, I predicted that the value
created at the alliance level of analysis increases with the capabilities controlled by each
partner. For HlA, the results show no signiﬁcant effect of each ﬁrms’ capabilities (T-
ratios = -1.423 and -1.400) on the value created at the alliance level of analysis.

HlB predicted that the value created at the alliance level of analysis increases
with the alliance capabilities controlled by both partners in the alliance. For HlB, the
result is signiﬁcant but the coefﬁcient has the opposite direction (see Table 3) to the
hypothesized relationship (T-ratio = -2.676). This result implies that larger combined
capabilities by both partners seem to have a negative effect on the value created at the
alliance level of analysis.

Hypothesis 2A stated that the value created at the level of the alliance increases
with the tie strength between the partners in the alliance. Hypothesis 2A is not supported.
The effect of the tie strength on value creation at the level of the alliance is not signiﬁcant

with a T-ratio of -0.759.

54

In Hypothesis 2B, I predicted that tie strength would positively affect the
relationship between alliance capabilities and value created at the level of the alliance.
Hypothesis 2B is supported (T-ratio = 3.277). Therefore, as represented in Figure l, the
strength of the tie between the partners in the alliance positively affects the effect of their
combined capabilities on value created at the level of the alliance. While the interaction
effect representation may seem to have a positive mean or main effect for capabilities, the
actual weighting of the data accounts for the negative main effect since the low strength

population is much larger than the high strength population in the sample.

Figure 1: Interaction Effect Between Alliance Capabilities and Tie Strength

 

 

 

 

 

 

Interaction between Alliance Capabilities and Tie Strength
710 1
705 ~ [I
I I I
700 « , ,’
c I
o ,’
a: I I
g 695 « /’
o , ,
0 I I
2 690 '1 h
s
o I
U 685 -‘ I’
s r’
2 680 ~
675 «
670 .
Low Alliance Capabilities High Alliance Capabilities
F—o— Low Tie Strength — + — High Tie Strength 1

 

 

Hypothesis 3 stated that the value created at the alliance level of analysis
increases with the alliance network closure. As expected, Model 2 in Table 4 shows that

the relationship between the amount of closure in an alliance network and the value

55

created at the level of the alliance is positive and signiﬁcant (T-ratio = 3.41). Thus H3 is
supported.

Finally, Hypothesis 4 predicted an inverse U-Shaped relationship between the
combined change in commitment to alliances of the ﬁrms in the alliance and value
created in the alliance. Therefore I expected a positive relationship between the
commitment to new alliances variable and value created at the level of the alliance and a
negative relationship for the squared term of the explanatory variable. Model 2 in Table 4
shows that H4 is not supported. The results were actually not signiﬁcant for both the

natural and the squared terms (T-ratios: -1.163 and 0.910, respectively).

56

Table 4. HCM2 Results for Models 1 and 2 (Dependent Variable = Yii't +1)

 

 

MODEL 1 2
Coefﬁcient T -Ratio Coefficient T -Ratio
Alliance Level
Alliance Capabilities -6282.l6** -2.67
Tie Strength -96.55 -0.76
Alliance Capabilities "‘ Tie Strength 4032.74*** 3.28
Alliance Closure 8.32*** 3.41
Commitment to New Alliances -931.41 -1.16
Squared Commitment to New 582.76 0.91
Alliances
Controls alliance level
Lagged DV 0555*" 88.18 0.56"" 88.23
Joint Venture -1 19.701 -0.44 17.70 0.07
Same Industry Dummy Code 180.722 0.79 176.32 0.80
Intercept 693.868*** 69.55 690.02*** 59.41
Firm Levels i & 1’
Market Share; * Alliance Tie -8-163 -1.42
Market Share,” "' Alliance Tie 3-766 0-74
Controls ﬁrm levels i & i’
Market ShﬂfCi "1.799”I 2.93 1.9617” 3.17
Assets; 0015*" 20.36 0015*" 20.43
Research & Development -0.073*"”" -3.86 -0.078*""" -4.13
Expenditures,-
Leveragei 5.102*** 3.37 5.103*** 3.37
Assets,- * Alliance Tie -0.001 -O. 17 0.002 -0.48
Market Share," -0.94 -l .64 1.246“ 2.37
Assets,” 0.02*** -31.74 -0.012*** 20.77
Research & Development -0.65*** -27.12 -0.083*** -5.09
Expenditures ,- '
Leverageiv 5.42" 2.79 4.355" 2.42
Assets," * Alliance Tie -0.000 -0.064 -0.004 -1 .40
Alliance Tie, Intercept -26.838 -0.257 173.701 -O.67
Total Variance & Covariance 3386874 3024079
Components
A Variance & Covariance Components 0.107

 

Signiﬁcant at *"p < 0.001; *"' p < 0.01; *p < 0.05: two tailed tests.
Degrees of freedom model 1: Level 1: 19936; Level 2i: 360; Level 2i ': 375.
Degrees of freedom model 2: Level 1: 19928; Level 2i: 360; Level 2i ': 375.

Table 5 presents the results of the hierarchical cross-nested linear analyses for my

value appropriation set of hypotheses. Model 3 in Table 5 shows the results for a model

57

management theory, control variables that were included to account for previous
performance, and size and capability direct effects on performance independent of the
alliance, were signiﬁcant and in the expected direction. Joint venture and supplementary
alliance dummies do not seem to make a difference on the predictions of the model.
Model 2 in Table 5 shows the results for the full value creation model.

In Hypothesis 5 I predicted that the value appropriated in an alliance by a partner
will be negatively associated with the relative strategic stake of that partner. Hypothesis
5 is supported. I found a signiﬁcant (T-ratio = 2.05) positive relationship between the
relative strategic stake of a partner and the difference in value appropriated in the alliance
by that partner.

Hypothesis 6 stated that the value appropriated ﬁ'om an alliance by a partner will
be positively associated with the relative capabilities of that partner. Hypothesis 6 is not
supported. Model 4 in Table 5 shows that the relationship between the relative
capabilities of a partner in an alliance and value appropriation is not signiﬁcant (T-ratio =
-0.48).

Hypothesis 7 predicted a positive relationship between the relative availability of
alternative partners and value appropriation. Model 4 in Table 5 shows that the
relationship is positive and signiﬁcant (T-ratio = 3.35), therefore supporting H7.

Finally, in Hypothesis 8 I predicted that the value appropriated from an alliance
by a ﬁrm in the alliance will be higher for ﬁrms that bridge more structural holes relative
to their partner. Hypothesis 9 is also supported. Firms that bridge more structural holes
relative to a partner in an alliance are able to appropriate more value in the alliance (T-

ratio = 3.09).

58

relative to a partner in an alliance are able to appropriate more value in the alliance (T—

ratio = 3.09).

Table 5. HCM2 Results for Models 3 and 4 (Dependent Variable = Yi- m +1)

 

 

MODEL 3 4

C oefﬂcient T-Ratio Coefﬁcient T -Ratio
Alliance Level
Relative Stakes 229.149* 2.05
Relative Capabilities -77.989 -0.48
Relative Availability of Alternative Partners 956.801 *** 3.35
(Mktshare)
Relative Structural Holes 14.108" 3.09
Controls alliance level
Lagged DV 0548*" 87.01 .547*** 8.6.94
Joint Venture -75.429 -0.27 -146.690 —0.54
Same Industry Dummy Code ~199.663 -0.87 -l73.746 -0.78
Intercept -55.534*** -5.72 -55.095*** -5.68
Firm Levels i & i’ (all controls)
Market Share,- 4.972*"”" 8.27 4992*" 8.30
Assets,- 0.017*** 23.12 0017*" 23.22
Research & Development Expenditures,- 43-13?" '7-11 9135*" '7-27
Leverage,- -1 . 120 -0.76 -1.116 -0.76
Market Share,- * Alliance Tie ‘3-62 ‘1-65 47-219" '3-96
Assets; 4- Alliance Tie 0.003 0.64 0.002 -0.47
Market Share," -l.999*** -3.90 -2.013"”'"" -3.93
Assets," -0.012"* -20.03 -0.012**"' -20.14
Research & Development Expendituresi- 0059*" 3-75 0062*" 3-89
Leveragei' -3.245 -1.85 -3.210 -1.83
Market Sharer .. Alliance Tie -0.666 -O.14 13.228“ 2.24
Assetsr * Alliance Tie 0.003 0.80 0008* 2.12
Alliance Tie, Intercept -89.383 -0.85 -105.153 -1.03
Total Variance & Covariance Components 3379458 3184443
A Variance & Covariance Components 0.058

 

Signiﬁcant at "*p < 0.001; ** p < 0.01; *p < 0.05: two tailed tests.
Degrees of ﬁeedom model 1: Level 1: 19934; Level 2i: 360; Level 2i': 375.
Degrees of freedom model 2: Level 1: 19930; Level 2i: 360; Level 2i': 375.

I graphically assessed the normality assumption for errors at level 1 by looking at

a normal probability plot for the residuals pooled across units. Non—normality of errors at

59

to outliers. If the distribution of errors were grossly non-normal, a poisson model would
be indicated. The normal residual plot served for checking normality and for the
identiﬁcation of outliers. The resulting plot is not grossly non-normal. It was a slightly
leptokurtic normal curve with some outlier values on both tails. To check for the
inﬂuence of the outliers I identiﬁed the outlier values in the residual data set and I re—run
the model without the outlier observations. The results of this control model are similar to
the complete dataset model indicating that the results are not grossly affected by the
larger residual values in the sample.

Iran the value creation model with alternative measures for tie strength and
capabilities. If tie strength is speciﬁed as the existence of an alliance between the partners
in the past, both tie strength and the interaction of alliance capabilities and tie strength do
not seem to have a signiﬁcant relationship to alliance value creation. The model including
capabilities measured as Sales, has near singularity issues and therefore does not run
because it is multicollinear. I also ran the value appropriation model with an alternative
measure for relative stakes. In this case, I characterized performance aspirations as
related to industry performance instead of previous ﬁrm performance. The relationship
between this measure and the dependent variable is not signiﬁcant and the inclusion does
not improve the model.

The next section organizes and elaborates the implications of these results, which
help to explain differences in value creation that result from inter-ﬁrm collaboration and
the related competition for value appropriation between ﬁrms involved in marketing

alliances.

60

CHAPTER 6
DISCUSSION AND CONCLUSIONS

This study examined the impact of three broad aspects of alliances — structural,
resource, and power issues - on the performance implications of alliances for their
partners. Previous approaches to the structural aspects of alliances had tried to understand
the degree to which closed or open networks could be appropriately regarded as the
normative ideal (Ahuja, 2000; Bae & Gargiulo, 2004; Burt, 1992; Coleman, 1988;
Walker, Kogut, & Shan, 1997; Zaheer & Bell, 2005). However, prior research on the
impact of alliance network structure on performance produced conﬂicting results. To
resolve this debate, I proposed that alliance network closure should be more relevant to
value creation at the level of the alliance while bridging structural holes should produce
an advantage in terms of value appropriation within the alliance. The results of my tests
show, in line with my theory, that network closure surrounding a marketing alliance
enhances the creation of value (i.e., aggregated performance) of partners in that alliance
(H3). I also found support for the prediction that the more structural holes a ﬁrm can
bridge relative to its alliance partner, the larger the portion of alliance value that ﬁrm will
appropriate from that alliance (H8).

The second aspect of alliances that has drawn attention in the alliance literature
concerns the role of ﬁrm capabilities in alliance performance. Previous approaches to the
study of the importance of partners capabilities in alliances had tried to understand the
degree to which the capability endowments of partners in alliances had a signiﬁcant

effect on performance (Bae & Gargiulo, 2004; McEvily & Marcus, 2005; Gulati 1999;

61

Stuart, 2000; Zaheer & Bell, 2005). In this study, taking a multilevel approach to the
issue, I did not ﬁnd support for the hypothesized relationships between each partner’s
capabilities (I-IlA) or the combined capabilities of partners (I-IlB) and value creation at
the level of the alliance. I did, however, ﬁnd that the relationship between the combined
capabilities of the partners and value creation at the level of the alliance is positively
affected by the presence of a strong tie between allies (H2B). Thus, higher capabilities of
each ﬁrm alone or in combination lead to higher value creation in alliances only when the
alliance partners were linked by multiple ties.

Finally, I found mixed support for the predictions about the importance of ﬁrm
power on the frrrn’s ability to appropriate value from their alliances. I found support for
this idea when power was expressed as the relative availability of alternative partners
(H5) or as the relative stakes (H7) between partners in the alliance, but did not ﬁnd
support when the test was based on differential capabilities each ﬁrm brings into the

alliance. These ﬁndings have important theoretical and practical implications.

Structure

The results in this study support the idea that alliances are simultaneously
cooperative and competitive relationships (Burt, 1991) for ﬁrm performance. Taken
together, Hypotheses 3 and 8 talk to the importance of understanding the mechanisms
that underlie the performance implications of structural positions in an alliance’s
network. These predictions imply that there may be an alternative to trying to understand
the degree to which closed or open networks (as competing explanations) could be

appropriately regarded as the normative ideal (Ahuj a, 2000; Bae & Gargiulo, 2004; Burt,

62

1992; Coleman, 1988; Walker, Kogut, & Shan, 1997; Zaheer & Bell, 2005). In strategic
alliances, the beneﬁts of increased trust and reduction of opportunism that result from
more closed or interconnected networks around the partners in the alliance result in more
potential for the alliance to produce aggregate value for alliance members. But also, ﬁrms
that bridge structural holes seem to have an advantage at the moment of appropriating the
value created in alliances. It appears that the key principle for improving performance
through collaboration has to do with the ability to simultaneously deve10p normative
pressures on an alliance’s partner (i.e., to increase the potential for the alliance to create
value) and some relative structural advantage in the relationship with that partner (i.e., to
maximize the ability to appropriate that value). Instead of one or the other, the optimizing
strategies seem to point to one and the other: closure and structural holes. Looking for a
sufﬁcient level of closure to guarantee value creation and simultaneously developing
sufﬁcient advantage in terms of structural holes to optimize the appropriation of that
value. Thus, it appears that the optimal choice of alliances for ﬁrms may have to do with
alliances that are neither overembedded in overly closed networks nor overly constrained
by bridging too few structural holes.

This view departs ﬁom previous approaches that confronted (a) the view that
networks with many connections between a ﬁrm’s partners, or “network closure” were
advantageous to strategic alliance performance (Ahuja, 2000; Coleman1988, 1990;
Gulati, 1995; Gulati & Singh, 1998; Walker, Kogut & Shan, 1997; Uzzi, 1997; Zaheer &
Venkatraman, 1995) to (b) views that supported the idea that structural holes was actually
better for alliance performance. (Burt, 1992; Hargadon & Sutton, 1997). Theorizing

about the different mechanisms that underlie closure and structural holes predictions

63

allowed me to predict and test a more integrative view that helps to explain the
contradictory ﬁndings of previous research in the management and sociology literatures.
Turning to the control variables in models 1 and 2, I also found an interesting
result related to the structural argument. A previous explanation for the contradictory
ﬁndings with respect to closure and structural holes relied on a contingency approach
(Ahuj a, 2000; Rowley et al., 2000). Ahuj a (2000) discussed how the contradictory results
may be due to differences in contexts. By contrasting his results with those of Hargadon
and Sutton (1997), he concluded that the explanation of the differences in results may be
due to the fact that his sample, contrary to the non-adversarial nature of Hargadon and
Sutton’s sample, includes alliances between competitor ﬁrms in the same industry.
Therefore, the adversarial nature of his sample may have resulted in a greater need for
trust in alliances and thus created a positive effect for network closure. To account for
this possibility, I included a control for alliances between partners in the same industry
(i.e., supplementary alliances). Results did not support this contingency argument in thatl
did not ﬁnd a signiﬁcant relationship between participation in the same industry (and
therefore potential competitiveness between partners) and alliance value creation. This
ﬁnding calls into question the contingency explanation for the contradictory results in the
previous literature. In particular, it raises questions about explanations that suggest that
alliances between competitors may be more likely to beneﬁt ﬁom closure in the network
while alliances between ﬁrms that do not participate in the same industry may beneﬁt
more from the availability of structural holes. While the study did not include any
measures of within industry muniﬁcence, this ﬁnding provides some initial evidence of

some possible limitations of that contingency argument.

64

Firms’ Capabilities

The results for the capability hypotheses of this study (HlA, HlB) are surprising.
Based on previous ﬁndings in the alliances literature I predicted that a multi-level
exploration of the phenomena would also show positive effects of each partner’s
capabilities and of the combined partners’ capabilities on value creation at the level of the
alliance. Instead, the results showed no signiﬁcant effect of each ﬁrm’s capabilities on
the value created at the alliance level of analysis. Altematively, the result for combined
ﬁrm capabilities is signiﬁcant but in the opposite direction to my prediction. Curiously
this ﬁnding implies that larger combined capabilities by both alliance partners have a
negative effect on the value created at the alliance level of analysis. I offer three possible
explanations for these surprising ﬁndings.

First, it is certainly plausible that my measure of capabilities is underspeciﬁed.
Relying on a proxy like market share to capture a ﬁrm’s marketing capabilities presents a
possible limitation of this study. However, since proxies like market share and sales have
been shown to predict performance in previous studies (e.g., Bae & Gargiulo, 2004;
Stuart, 2000), this explanation will not explain away the difference between these
ﬁndings and previous results.

Second, previous results relied on a single level approach that may have
confounded the direct effect of capabilities on performance with the alliance level effect
of capabilities on performance in the previous literature (e.g., Ahuj a, 2000; Bae &
Gargiulo, 2004; Stuart, 2000; Zaheer & Bell, 2005). Previous research fails to take
advantage of one of the key beneﬁts of network’s approaches: the multilevel nature of the

data (e.g., the ﬁrm, the alliance, the alliance network). This single level approach may

65

account, in part, for the different findings. By modeling capabilities in a multilevel
approach, my model may have eliminated the confusion between direct capability effects
on performance and alliance related capability effects on performance that may have
arisen from looking at the effect of own and partner capabilities effects on performance at
a single level of analysis. Note that, as reported in the results in Model 2 of table 4, the
relationships between Marketshare * Alliance Tie for each one of the partners (i.e. my
tests for hypothesis 1A) and value creation at the level of the alliance are non signiﬁcant.
But the direct effects of Market Share.- and Market Sharer-on performance (i.e., the
market share measures for each ﬁrm that are included as controls in the model) are both
positive and signiﬁcant. Therefore, the single level approaches of previous research may
have captured some of the actual effect of ﬁrm capabilities on ﬁrm performance that is
independent of those ﬁrms’ being in an alliance. The multilevel approach has therefore
allowed me to separate market share effects within the alliance (not signiﬁcant) ﬁ'om
market share direct effects (signiﬁcant) on performance. This demonstrates the need of a
multilevel approach to theorizing and testing when dealing with multilevel phenomena
like strategic alliances. In this study I offered a hierarchical cross-nested model
(Raudenbush & Bryk, 2002; Van Duij in, 1995; Lazeaga & Van Duij in, 1997) as such an
approach. This approach allows researchers to acknowledge the non-independent nature
of alliance data and to investigate how, for example, both individual partners’ attributes
(e.g., capabilities of each partner in an alliance), the alliance conﬁguration (i.e., alliance
level constructs) interact to explain alliance outcomes within the larger set of alliances in
which both partners may also participate. Thus, this study demonstrates how hierarchical

cross-nested linear methods may be used to investigate the effect of alliance speciﬁc

66

factors within each level in the alliance (e.g., partner A, Partner B, alliance effects and
alliance by partner effects). I hope this study will help guide ﬁrture multilevel research in
this area.

Finally, a third explanation for the results from my tests of ﬁrm capabilities takes
form when looking at hypothesis 2B. While the combined capabilities of the partners
seem to have a negative effect on the value created at the alliance level of analysis, this
relationship gets reversed when moderated by the strength of the tie between the allies.
This ﬁnding points to an integration between relational and capability issues in predicting
overall alliance performance. Previous studies generally have relied on small samples of
alliances and have not controlled for the strength of the tie between the alliance partners.
Therefore, previous tests may be underspeciﬁed and the results may be inﬂuenced by
relational issues between allies in the samples.

In conclusion, the capabilities of both partners in an alliance seem to have a
positive effect on alliance value creation at the level of the alliance only when combined
and moderated by the strength of the tie between the ﬁrms in the alliance. This ﬁnding is
consistent with a broad process literature in alliances that emphasizes the role of trust and
commitment between the allies in alliance performance (tie). It is plausible that the larger
and more successful the ﬁrms entering an alliance, the more important the strength of the
relationship is for alliance performance. Partners with large capability endowments are by
deﬁnition, less dependant on alliance performance for overall ﬁrm success and therefore
may need to rely on more relational factors for their alliances to create value. Large
partners that have strong ties between the, may develop joint routines or work

mechanisms by working together through several alliances and these routines may help

67

them overcome the problems of getting two large ﬁrms working together with one

another.
Firm Power

Finally I tested the role of ﬁrm power on value appropriation in alliances. The
mixed results for the relative power set of hypotheses in the study are also very
interesting. I examined three potential causes of bargaining power: relative availability of
alternative partners, relative capabilities and relative stakes. Of the three hypotheses I
found support for the roles of the relative stakes of partners and the relative availability of
alternative partners, on value appropriation from alliances. However, I did not ﬁnd
support for the role of the differences between the capabilities of each partner and the
appropriation of value. That is, ﬁrms with more capabilities than their partner do not
appropriate a disproportional share of value with respect to their partners. These results
suggest that not all sources of power over a partner explain value appropriation.

A common aspect of both relative availability of alternative partners and relative
stakes (both of which have signiﬁcant effects on ﬁrm value) is that both imply conditions
of vulnerability. First, for relative availability of alternative partners, the vulnerability
arises ﬁ'om the risk of exclusion. Network exchange theory (W iller, Walker, Markovsky,
Willere, Lovaglia, Thye, & Simpson, 2002; Yamagishi, Gillrnore, & Cook, 1988) argues
that bargaining power is a function of the extent to which actors in a relationship are
vulnerable to exclusion from exchanges within the network. Therefore, when there are
many alternatives to a partner for a ﬁrm to engage in an alliance, that partner is

vulnerable because it risks being excluded from the alliance in favor of another

68

alternative partner. Second, ﬁrms with high strategic stakes will have lower bargaining
power in alliances (Bacharach & Lawler, 1984; Yan & Gray, 1994) because high payoff
for cooperation is likely when ﬁrms are in vulnerable strategic situations (Eisenhardt &
Schoonhoven, 1996).

Difference in capabilities is less an expression of vulnerability and has more to do
with potential contributions of the ﬁrms to the alliance. Given no other vulnerabilities or,
better said, when bargaining power issues related to vulnerability factors are accounted
for, appropriation may be expected to be somewhat proportional to contributions and,
therefore, have no signiﬁcant effects on value appropriation within alliances. It is not the
capabilities that a ﬁrm may contribute but the lack of alternatives or urgency to go
through with a collaboration that drives power relationships, and therefore value
appropriation, in strategic alliances. Consistent with this explanation, the hypotheses in
this study that represent vulnerable situations for ﬁrms going into an alliance explain
differential value apprOpriation, while the one that is a mere description of potential
contributions does not have a signiﬁcant effect on the results. Here, a plausible extension
to the paper would be to explore the interaction between differential capabilities and the
availability of alternative partners. It may well be that differential capabilities become
important when no alternative partners are available, but do not matter as much when
there are many alternative partners from whom to get access to similar marketing

capabilities.

69

Limitations

This dissertation, like any other, has several limitations that open avenues for
future research. First, although I examined an important set of complementary outcomes,
the impact of alliances on conjoint performance and differences in performance between
ﬁrms in an alliance within a marketing alliances network, it is important to consider
alternative aspects of ﬁrm value (e. g., the sustainability of alliance effects on
performance) that may also be affected by the predictors included in this study. For
example, exercising power in a relationship to appropriate more value in an alliance may
have long term effects on the relationship between ﬁrms and, therefore, on the medium
term value of the alliance. Also, appropriation practices may lead to reputational issues
that may affect the willingness of potential partners to form an alliance with a ﬁrm in the
future.

Second, it would be important to test the generalizability of the ﬁndings to
alliances beyond marketing alliances. While marketing alliances are a signiﬁcant
proportion of the alliances ﬁrms engage in, research and development alliances may be
an interesting ground to test the generalizability of these ﬁndings since they usually
imply more speciﬁc sets of capabilities, longer terms to achieve results and, therefore,
more issues of trust and commitment to the investment than the marketing alliances that
are the focus of this study. As such, research and development alliances may need more
trusting relationships and tie strength or closure may therefore have an even larger impact
on their performance. Also, it may be the case that, given the long term demands of such
alliances, contractual issues or cross-ownership may play an important role in such

projects.

70

Third, this research takes into account the formal network that is determined by
the alliances that ﬁrms form. Future research should analyze the intersection between
these formal expressions of relationships and informal networks that may be acting as
antecedents to the formal alliances the ﬁrms go into, or of complementary effects on the
inﬂuence that formal networks have on performance. Formal and informal networks
usually co-evolve allowing, also, for ﬁrrther research on the determinants of alliance
network structures for ﬁrm. It would be important to understand if informal networks or
previous alliances performance as explained by the model I present in this study,
independently or together, provide a more complete explanation of the alliances choices
that ﬁrms make than previous research. Longitudinal selection models can be used to
study the dynamic implications of informal and performance issues on the alliances
choices that ﬁrms make.

Finally, this model is surely incomplete and underspeciﬁed. While I think this is
an important step in unifying the ﬁndings in the alliance literature, future developments
should try to understand how other aspects of partners, alliances, or their interaction, may

help determine the long term value of alliance strategy choices for ﬁrms.

Conclusions

The ﬁndings of this study are relevant for management research for a number of
reasons. The main implication of this dissertation is that ﬁrms’ decisions regarding
alliance choices should be sensitive to collaborative and competitive conditions that
characterize each potential alliance choice. In contrast to previous studies of performance

in the alliance literature that focused exclusively on individual ﬁrm outcomes (i.e.,

71

assuming full capture of value created in alliances by participating ﬁrms), this
dissertation considered value creation and value appropriation as two different
components of alliance performance each with it’s own set of relevant predictors. While
value creation results from aspects of ﬁrms and their alliances’ choices that promote
collaboration (e.g., network closure), value appropriation results from differential
exchange conditions for partners in an alliance (e. g., relative availability of alternative
partners) that deﬁne the outcomes of their competition to appropriate the results of those
alliances. The size of the appropriation of value that each ﬁrm derives from an alliance is
the result of a cooperative game (Brandenburger & Stuart, 2001; Lippman & Rumelt,
2003) between partners who cooperate to create value and compete to appropriate that
value simultaneously (Lax & Sebenius, 1986; Hamel, Doz, & Prahalad, 1989) and,
therefore, is the outcome of exchange processes, bargaining, and the overall structure of
the relationship.

Second, this study contributes to the integration of structuralist and resource based
approaches, therefore addressing some criticisms of previous collaboration research. This
study is able to take an integrated view of ﬁrm characteristics and the structure of the
relationships they are embedded in by taking advantage of multilevel analytical tools and
theory. Taking advantage of the multilevel nature of alliance data, this study is able to
show how structural and ﬁrm speciﬁc characteristics and the interaction between them,
determines the creation and appropriation of value in alliances. In contrast to previous
research that has tried to integrate ﬁrm and structural factors by constructing network
variables as ﬁrm level predictors and aggregating portfolios of alliances (as averages),

this dissertation has been able to integrate them by cross-classifying the alliances between

72

the ﬁrms that form them and to provide a cleaner set of explanations for the impact of
both ﬁrm and structural factors in an alliance.

Third, this dissertation contributes to the alliances, sociology and network
literatures by providing a synthesis of Structural Holes and Closure perspectives with the
intention of helping to unify and redirect the network capital (i.e., social capital,
relational capital, c00perative exchanges) agenda. I provide a ﬁrst set of results that show
how closure and structural holes may not be directly contradictory in the effects of
alliances on frrrns’ performance. This ﬁnding stands in contrast to the divided belief that
either closure or structural holes can be good for collaboration, but not both. Despite a
large tradition of studies that have tested these perspectives as contradictory, little effort
has been made to understand the potential interplay of both factors in promoting the value
of a relationship for a ﬁrm. I hope this dissertation may help promote research that seeks
to ﬁrrther understand the structural holes and closure constructs in the context of
alternative outcomes and contexts. For example, it would be interesting to analyze if
these ﬁndings hold for social capital implications of informal networks or for outcomes
related to information production and distribution. Indeed, I am hopeful that the present
dissertation will also motivate additional work, importantly, including qualitative
research that might provide insight into the mechanisms of value appropriation and value
creation in strategic alliances, and simulation research that might provide optimizing
hypothesis through agent based models to test in the ﬁeld.

Fourth, this study contributes to the capabilities literature as it pertains to alliances
by providing a ﬁrst multilevel test of the effect of capabilities on the impact that alliances

have on ﬁrms. Contrary to some of the most recent results in this literature, this

73

dissertation shows that previous research may have confounded direct effects of ﬁrm or
partner capabilities on alliance driven performance, because of their single level
approach. In this study, partner capabilities do not seem to have a positive impact on the
performance of alliances for ﬁrms, while they do have a direct impact in the performance
of ﬁrm. Previous ﬁndings that had relied on a single level approach should then be re-
examined in light of these ﬁndings. Future developments may want to take a more case
based approach to the capabilities issue to understand if there is an issue of speciﬁcity of
the resources to the alliance that is essential for performance. A case speciﬁc ﬁnding,
though, would be seriously impaired for quantitative testing of the sort that I performed
in this dissertation.

Finally, this study contributes to the methodological study of alliances by
providing a multilevel approach to the study of alliance performance. For the purpose of
analysis most networks data have been analyzed separately at different levels precluding
theoretical developments across levels (Jones, Hesterly & Borgatti, 1997; Monge &
Contractor, 2003). For example, some recent studies that have addressed the need to
integrate RBV and network constructs to explain ﬁrm performance resulting ﬁom
alliances (e. g., Ahuja, 2000; Bae & Gargiulo, 2004; Stuart, 2000; Zaheer & Bell, 2005)
have done so by constructing network variables as ﬁrm level predictors (the density of a
ﬁrm’s network), therefore rendering themselves unable to explain the value of speciﬁc
alliance choices. This single level approach accounts, in part, for the contradictions in
some previous ﬁndings that these study may help resolve. In this study I offered a
hierarchical cross-nested model (Raudenbush & Bryk, 2002; Van Duijin, 1995; Lazeaga

& Van Duijin, 1997) as an improvement over previous approaches. This approach

74

allowed, for example, to understand how capabilities may have different meaning within
the alliance than within the ﬁrm, and therefore qualiﬁed ﬁndings of previous research
about the value of larger capabilities of partners for alliance success. I hope this
demonstration of how hierarchical cross-nested linear methods may be used to investigate
the effect of alliance speciﬁc factors within each level in the alliance (e. g., partner A,

Partner B, and alliance effects) will guide future multilevel research in this area.

75

REFERENCES

Ahuja, G. (2000a). Collaboration networks, structural holes, and innovation: A
longitudinal study. Administrative Science Quarterly, 45(3), 425-455.

Ahuj a, G. 2000b. The duality of collaboration: Inducements and opportunities in the
formation of interﬁrrn linkages. Strategic Management Journal, 21(3), 317-343.

Alliance Analyst. 1996. Managing Alliances — Skills for the Modern Era. March 1998.
NewCap Communications: Philadelphia, PA.

Altman, E. I. 1983. Corporate Distress: A Complete Guide to Predicting, Avoiding, and
Dealing with Nankruptcy. New York: Wiley.

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

Anand, B. N., & Khanna, T. 2000. Do ﬁrms learn to create value: The case of strategic
alliances. Strategic Management Journal, 21, 295-315.

Bacharach, S. B., & Lawler, E. J. 1981. Power and tactics in bargaining. Industrial and
Labor Relations, 34, 219-233.

Bacharach, S. B. and Lawler, J. L. 1984. Bargaining: Power, Tactics, and Outcomes. San
Francisco: J ossey—Bass.

Bae, J. H., & Gargiulo, M. 2004. Partner substitutability, alliance network structure, and
ﬁrm proﬁtability in the telecommunications industry. Academy of Management
Journal, 47(6), 843-859.

Baker, W. 1984. The social structure of a national securities market. American Journal of
Sociology, 89, 775-811

Baker, Wayne E. 1990. Market networks and corporate behavior. American Journal of
Sociology 96, 589-625.

Barney, J. B. 1986. Strategic factor markets: Expectations, luck, and business strategy.
Management Science 32, 1231-1241.

Barney, Jay. 1991. Firm Resources and sustained competitive advantage. Journal of
Management. 17, 99-120.

Barney, J. B. 2001. Resource-based theories of competitive advantage. Journal of
Management. 27(6), 643-650.

76

Baum, Joel A. C., & Oliver, Christine. 1992. Institutional embeddedness and the
dynamics of organizational populations. American Sociological Review. 57, 540-
559.

Blau, Peter M. 1964. Exchange and Power in Social Life. New York: Wiley and Sons.

Biggart, N. W. 2001. Banking on each other: The situational logic of rotating savings and
credit associations. Advances in Qualitative Organization Research, 3, 129-153.

Brandenburger A, & Stuart, G. 2001. Biform games. Working paper
(www.people.hbs.edu/abrandenburger). I

Brandenburger, A. M, & Stuart, H. 2006. Bifonn games. Forthcoming in Management
Science.

 

Bromiley, P. 1991. Testing a causal model of corporate risk taking and performance. in:
Academy of Management Journal, 34, 37-59.

Buhner, R. 1987. Assessing international diversiﬁcation of West German corporations.
Strategic Management Journal 8(1), 25-37.

Burgers, Willem P., Hill, Charles W. L., & Kim, Chan W. 1993. A theory of global
strategic alliances: The case of the global auto industry. Strategic Management
Journal 14, 419-432.

Burt, Ronald S. 1983. Corporate Proﬁts and Cooptation. New York: Acaderrric Press.

Burt, Ronald S. 1991. Measuring age as a structural concept. Social Networks 13, 1-34.

Burt, R. S. 1992. Structural Holes: The Social Structure of Competition. Harvard
University Press: Cambridge, MA.

Burt, R. S. 1997. The contingent value of social capital. Administrative Science
Quarterly, 42(2), 339—365.

Burt, R. S. 2000. The network structure of social capital, Research in Organizational
Behavior, 22, 345-423. New York: JAI-Elsevier Science Inc.

Burt, R. S. 2001. Structural Holes versus Network Closure as Social Capital. In Social
Capital: Theory and Research, Lin N, Cook K, Burt RS (eds). Aldine de Gruyter:
Chicago, IL, 31-56.

Casciaro, Tiziana,& Piskorski, Mikolaj Jan. 2005. Power Imbalance, Mutual

Dependence, and Constraint Absorption: A Closer Look at Resource Dependence
Theory. Administrative Science Quarterly 50, 167.

77

Child, J ., & Faulkner, D. 1998. Strategies of Cooperation: Managing Alliances,
Networks, and Joint Ventures, Oxford, UK: Oxford University Press.

Chung, S., Singh, H., & Lee, K. 2000. Complementarity, status similarity and social
capital as drivers of alliance formation. Strategic Management Journal, 21(1), 1-
22.

Coase, R. H. 1937. The nature of the ﬁrm. Economica, 4, 385-405.

Coleman, J. S. 1988. Social capital in the creation of hmnan capital. American Journal of
Sociology, 94, 95-120.

Coleman, James. 1990. Foundations of Social Theory. Cambridge, Mass.: Harvard
University Press.

Conner, K. 1994. The resource-based challenge to the industry-structure perspective, in
D. Moore (Ed.), Academy of Management Best Papers Proceedings 1994,
Charleston, SC: Academy of Management.

Cook, Karen S. 1977. Exchange and power in networks of interorganizational relations.
Sociological Quarterly 18, 62-82.

Cook, Karen S., & eritmeyer, Joseph M. 1992. Two approaches to social structure:
exchange theory and network analysis. Annual Review of Sociology 18, 109-27.

Cook, Karen S., & Yamagishi, Toshio. 1992. The effect of latent paths on the distribution
of power in exchange network structures. Paper presented at the annual meeting
of the American Sociological Association, Pittsburgh, August 20-24.

Culpan, R., & Eugene, K.A. 1993. Cross-national corporate partnerships: trends in
alliance formation. In Multinational Strategic Alliances, Culpan, R (ed.).
Hayworth Press: New York.

David, P. 1985. Clio and the economics of QWERTY,” American Economic Review, 75,
332-337.

Deephouse, D. L., & Wiseman, R. M. 2000. Comparing alternative explanations for
accounting risk-retum relations. Journal of Economic Behavior & Organization,

42(4), 463-482.

Dore, Ronald. 1983. Goodwill and the spirit of market capitalism. British Journal of
Sociology, 34, 459-482

Dwyer, F. R., & Walker, C. C. 1981. Bargaining in an asymmetrical power structure.
Journal of Marketing, 45(1), 104-1 15.

78

Dyer, J. H., & Nobeoka, K. 2000. Creating and managing a high-performance
knowledge-sharing network: The Toyota case. Strategic Management Journal,
21(3), 345-367.

Dyer, J. H., & Singh, H. 1998. The relational view: Cooperative strategy and sources of
interorganizational competitive advantage. Academy of Management Review,
23(4), 660-679.

Eisenhardt, K. M., & Schoonhoven, C. B. 1996. Resource-based view of strategic
alliance formation: Strategic and social effects in entrepreneurial ﬁrms. i
Organization Science, 7(2), 136-150.

Emerson, R. M. 1962. Power-dependence relations. American Sociological Review,
27(1), 31-41.

 

Emerson, R. 1972. Exchange Theory. In Berger, J ., Zelditch, M., & Anderson, B. (eds), ‘E
Sociological Theories in Progress. Boston, MA: Houghton lVIifﬂin.

Emerson, Richard M. 1972a. Exchange Theory, Part I: A Psychological Basis for Social
Exchange. In Joseph Berger, Morris Zelditch, Jr., and Bo Anderson, eds.,
Sociological Theories in Progress, Vol. 2, pp. 38-57. Boston: Houghton Mifﬂin.

Emerson, Richard M. 1972b. Exchange Theory, Part 11: Exchange Relations and
Networks. In Joseph Berger, Morris Zelditch, Jr., and Bo Anderson, eds.,
Sociological Theories in Progress, Vol. 2, pp. 58-87. Boston: Houghton Mifﬂin.

Fisher, R., & Ury, W. 1981. Getting to YES: Negotiating agreement without giving in.
New York: Penguin Books.

Frank, O. 1979. Estimation of Population Totals by Use of Snowball Samples. In
Holland, P., and Leinhardt, S. (eds.) Perspectives on Social Network Research
New York: Academic Press, 319-347.

Frank, O. 2005. Network Sampling and Model Fitting. In Carrington, P., Scott, J. and
Wasserman, S. (eds.) Models and Methods in Social Network Analysis. New
York: Cambridge University Press, 31 ~56.

Freeman, John, & Harman, Michael T. 1989. Commentary and debate. The American
Journal of Sociology, 95, 425-439.

Goerzen, A., & Beamish, P. W. 2005. The effect of alliance network diversity on
multinational enterprise performance. Strategic Management Journal, 26(4), 333-
354.

Goodwin, Jeff, & Emirbayer, Mustafa. 1994. Network analysis, culture, and the problem
of agency. American Journal of Sociology, 99, 1411-54.

79

Granovetter, M. S. 1973. Strength of weak ties. American Journal of Sociology, 78(6),
1360-1380.

Granovetter, Mark. 1974. Getting a Job. Cambridge: Harvard University Press.

Granovetter, M. 1985. Economic-action and social-structure - the problem of
embeddedness. American Journal of Sociology, 91(3), 481-510.

Gulati, R. 1995a. Does familiarity breed trust - the implications of repeated ties for
contractual choice in alliances. Academy of Management Journal, 38(1), 85-112.

Gulati, R. 1995b. Social structure and alliance formation patterns: A longitudinal
analysis. Administrative Science Quarterly, 40(4), 619-652.

Gulati, R. 1999. Network location and learning: The inﬂuence of network resources and
ﬁrm capabilities on alliance formation. Strategic Management Journal, 20(5),
397-420.

Gulati, R., & Gargiulo, M. 1999. Where do interorganizational networks come from?
American Journal of Sociology, 104(5), 1439-1493.

Gulati, R., Nohria, N., & Zaheer, A. 2000. Strategic networks. Strategic Management
Journal, 21(3), 203-215.

Gulati, R., & Singh, H. 1998. The architecture of cooperation: Managing coordination
costs and appropriation concerns in strategic alliances. Administrative Science
Quarterly, 43(4), 781-814.

Hagedoom, J ., & Schakenraad, J. 1994. The Effect of Strategic Technology Alliances on
Company Performance. Strategic Management Journal, 15(4), 291-309.

Hamel, G., Doz, Y. L., & Prahalad, C. K. 1989. Collaborate with your competitors and
win. Harvard Business Review, 89(1), 133—139.

Harman, M. T., & Freeman, J. 1977. Population ecology of organizations. American
Journal of Sociology, 82(5), 929-964.

Harman, Michael, T., & Freeman, John. 1987. The ecology of organizational morality:
American labor unions, 1836-1985. American Journal of Sociology, 92, 910-43.

Harman, Michael, T., & Freeman, John. 1988. The ecology of organizational founding:
American labor unions, 1836-1985. American Journal of Sociology, 94, 25-52.

Harman, Michael T., & Carroll, Glenn R. 1992. Dynamics of Organizational Populations.
New York: Oxford University Press.

80

 

 

Hargadon, A., & Sutton, R. I. 1997. Technology brokering and innovation in a product
development ﬁrm. Administrative Science Quarterly, 42(4), 716-749.

Harrigan, Kathryn R. 1985. Strategies for Joint Ventures. Lexington , MA: Lexington
Books.

Harrigan, Kathryn R. 1986. Managing for Joint Ventures Success. Lexington, MA:
Lexington Books.

Harrigan, Kathryn R., & Newman, W. H. 1990. Bases of interorgarrization cooperation:
Propensity, power, persistence. Journal of Management Studies, 27, 417-434.

Hanison, J. S., Hitt, M. A., Hoskisson, R.E., & Ireland, RD. 1991. Synergies and post-
acquisition performance: Differences versus similarities in resource allocations.
Journal of Management, 17, 173-190.

Henderson, R., & Cockbum, I. (1994). Measuring competence? Exploring ﬁrm effects in
pharmaceutical research. Strategic Management Journal, Winter Special Issue,
15, pp. 63-84.

Holland, P. W., & Leinhardt, S. 1981. An exponential family of probability distributions
for directed graphs. Journal of the American Statistical Association, 76, 33—50.

Homans, George Caspar. 1950. The Human Group. London: Routledge and Kegan Paul.

Homans, George Caspar. 1974 [1961]. Social Behavior: Its Elementary Forms. Rev. ed.
New York: Harcourt Brace J ovanovich.

Jensen, M. 1989. Eclipse of the public corporation. Harvard Business Review, 67(5), 61-
74.

Inkpen, A. A. 2001. Strategic Alliances. Published in Hitt, M. A., Freeman, E. E., &
Harrison, J. S. (eds.). Blackwell Handbook of Management. Malden, MA:
Blackwell Publishing.

Jones, C., Hesterly, W. S., & Borgatti, S. P. 1997. A general theory of network
governance: Exchange conditions and social mechanisms. Academy of
Management Review, 22, 911-945.

Kale, P., Dyer, J. H., & Singh, H. 2002. Alliance capability, stock market response, and
long-term alliance success: The role of the alliance frmction. Strategic
Management Journal, 23(8), 747-767.

Kraatz, Matthew S. 1998. Learning by association? Interorganizational networks and
adaptation to environmental change. Academy of Management Journal, 41, 621-
643.

81

Lawler, Edward J ., & Bacharach, Samuel B. 1979. Power dependence in individual
bargaining: The expected utility of inﬂuence. Industrial & Labor Relations
Review, 32, 196.

Lax, D. A., & Sebenius, J. 1986. The Manager as Negotiator. New York: The Free Press.

Lazeaga, E., & Van Duijin, M. 1997. Position in formal structure, personal characteristics
and choices of advisors in a law ﬁrm: a logistic regression model for dyadic
network data. Social Networks, 19, 375- 397

Lazerson, Mark. 1995. A new phoenix? Modern putting-out in the Modena knitwear
industry. Administrative Science Quarterly, 40, 34-59.

Lin, Nan. 2001. Building a Network Theory of Social Capital. In Lin, N., Cork, K., and
Burt, R. (eds.) Social Capital Theory and Research New York: Aldino De
Gruyter, 3-29.

Lin, N., Ensel, W. M., & Vaughn, J. C. 1981. Social Resources and Strength of Ties -
Structural Factors in Occupational-Status Attainment. American Sociological
Review, 46(4), 393-403.

Lippman, Steven A., & Rumelt, Richard P. 1982. Uncertain irnitability: An analysis of
interﬁrrn differences in efﬁciency under competition. Bell Journal of Economics,
1 3 , 41 8-43 8.

Lippman, Steven A., & Rumelt, Richard P. 2003a. A bargaining Perspective on Resource
Advantage. Strategic Management Journal, 24, 1069-1086.

Lippman, Steven A., & Rumelt, Richard P. 2003b. The payments perspective: Micro-
foundations of resource analysis. Strategic Management Journal, 24, 903.

Makadok, Richard. 2003. Doing the right thing and knowing the right thing to do: Why
the whole is greater than the sum of the parts. Strategic Management Journal, 24,
1043.

March, J. G., & Shapira, Z. 1992. Variable Risk Preferences and the Focus of Attention.
Psychological Review, 99(1), 172-183.

Markovsky, B., Willer, D., & Patton, T. 1988. Power Relations in Exchange Networks.
American Sociological Review, 53(2), 220-236.

Marsden, Peter V. 1990. Network data and measurement. Annual Review of Sociology,
16, 435-463.

Marsden, P. V. 1993. The reliability of network density and composition measures.
Social Networks, 15(4), 399-421.

82

McEvily, B, & Marcus, A. 2005. Embedded ties and the acquisition of competitive
capabilities. Strategic Management Journal (in press).

McEvily, B., & Zaheer, A. 1999. Bridging ties: A source of ﬁrm heterogeneity in
competitive capabilities. Strategic Management Journal, 20(12), 1133-1156.

McEvily, Bill, & Zaheer, Akbar. 2005. Does Trust Still Matter? Research on the Role of
Trust in Interorganizational Exchange. In Bachman, R., and Zaheer, A. (eds.)
Handbook of Trust Research. Cheltenham: Edward Elger.

Miller, Danny, & Jamal, Shamsie. 1996. The resource-based view of the ﬁrm in two I
environments: The Hollywood ﬁlm studios from 1936 to 1965. Academy of
Management Journal, 39(3), 519-543.

Miller, K. D., & Chen, W. R. 2004. Variable organizational risk preferences: Tests of the
March-Shapira model. Academy of Management Journal, 47(1), 105-115. ll

 

Miller, K. D., & Reuer, J. J. 1996. Measuring organizational downside risk. Strategic
Management Journal, 17(9), 671-691.

Mitchell, W., & Singh, K. 1996. Survival of businesses using collaborative relationships
to commercialize complex goods. Strategic Management Journal, 17(3), 169-195.

Monge, P. R., & Contractor, N. 2003. Theories of Communication Networks. New York,
NY: Oxford University Press.

Nelson, R. R. 1991. Why do ﬁrms differ and how does it matter? Strategic Management
Journal, Winter Special Issue, 12, 61-74.

Newman , M. E. J ., Watts , D. J ., & Strogatz, S. H. 2002. Random graph models of social
networks. Proceedings on the National Academy of Sciences, 99(1), 2566-2572.

Nohria, N., & Garcia-Pont, C. 1991. Global strategic linkages and indusu'y structure.
Strategic Management Journal, 12, 105-124.

Owen-Smith, Jason, & Powell, Walter W. 2003. The expanding role of university
patenting in the life sciences: Assessing the importance of experience and
connectivity. Research Policy, 32,1965-171 1.

Owen-Smith, J ., & Powell, W. W. 2003. Knowledge networks as channels and conduits:
The effects of spillovers in the Boston biotechnology community. Organization
Science, 15(1), 5-21.

Penrose, E. T. 1959. The Theory of the Growth of the Firm. New York: Wiley.

83

Peteraf, M. A. 1993. The cornerstones 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. Harper & Row, New York.

Podolny, J. M., & Baron, J. N. 1997. Resources and relationships: Social networks and
mobility in the workplace. American Sociological Review, 62(5), 673-693.

Portes, A. 1998. Social capital: its origins and applications in modern sociology, Annual
Review of Sociology, 24, 1-24.

Powell, W. W., Koput, K. W., & Smith-Doerr, L. 1996. Interorganizational collaboration
and the locus of innovation: Networks of learning in biotechnology.
Administrative Science Quarterly, 41(1), 116-145.

Rao, H. 1994. The social construction of reputation: certiﬁcation contests, legitimization,
and the survival of organizations in the American automobile industry: 1985-
1912. Strategic Management Journal, Winter Special Issue, 15, 29-44.

Raudenbush, S. W., & Bryk, A. S. 2002. Hierarchical Linear Models. Thousand Oaks,
CA: Sage Publications Inc.

Reagans, R., & Zuckerman, E. W. 2001. Networks, diversity, and productivity: The
social capital of corporate R&D teams. Organization Science, 12(4), 502-517.

Root, F. R. 1988. Some Taxonomies of International C00perative Arrangements. In F.J.
Contractor & P. Lorange (Eds), Cooperative Strategies in International Business:
69-80. Lexington, MA: Lexington Books.

Rowley, T., Behrens, D., & Krackhardt, D. 2000. Redundant governance structures: An
analysis of structural and relational embeddedness in the steel and semiconductor

industries. Strategic Management Journal, 21 , 369-3 86.

Rumelt, Richard P. 1984. Toward a strategic theory of the ﬁrm. Competitive Strategic
Management, 557-570.

Rumelt, Richard P. 1991. How much does industry matter? Strategic Management
Journal, 12, 167-175.

Sampson, Rachelle C. 2005. Experience effects and collaborative returns in R&D
alliances. Strategic Management Journal, 26, 1009-1031.

Schelling, Thomas C. 1956. Book reviews. National Tax Journal (pre-l986), 9: 91-93.

84

Shan, W. 1990. An empirical analysis of organizational strategies by entrepreneurial high
technology ﬁrms. Strategic Management Journal, 11, 129-139.

Simmel, Georg. 1950. The Sociology of Georg Simmel. Trans. Kurt H. Wolff. New
York: Free Press.

Simmel, G. 1955. Conﬂict and the Web of Group Afﬁliations. New York, NY: Free
Press.

Singh, K., & Mitchell, W. 2005. Growth dynamics: The bidirectional relationship
between interﬁrm collaboration and business sales in entrant and incumbent
alliances. Strategic Management Journal, 26(6), 497-521 .

Skvoretz, John, & Willer, David 1993. Exclusion and power: A test of four theories of
power in exchange networks. American Sociological Review, 58, 801-818.

Smith-Doerr, Laurel, & Powell, Walter W. 2005. Networks and Economic Life. In
Smelser, N, and Swedberg, R. (eds) The Handbook of Economic Sociology. New
York: Russell Sage Foundation, 379-402.

Stinchcombe, Arthur L. 1990. Weak structural data (review of Mizruchi and Schwartz).
Contemporary Sociology, 19, 380-82.

Stuart, T. E. 2000. Interorganizational alliances and the performance of ﬁrms: A study of
growth and innovation rates in a high-technology industry. Strategic Management
Journal, 21(8), 791-811.

Stuart, Toby E., & Podolny, Joel M. 1999. Positional Consequences of Strategic
Alliances in the Semiconductor Industry. Pp. 161-82 in vol. 16 of Research in the
Sociology of Organizations, ed. Steven Andrews and David Knoke. Greenwich,
Conn.; JAI Press.

Stuart, T. E., Hoang, H., & Hybels, R. C. 1999. Interorganizational endorsements and the
performance of entrepreneurial ventures. Administrative Science Quarterly, 44(2),
3 1 5-3 49.

Teece, D. J. 1986. Proﬁting from technological innovation: Implications for integration,
collaboration, licensing and public policy. Research Policy, 15(6), 285-305.

Teece, D. J ., Pisano, G., & Shuen, A. 1997. Dynamic capabilities and strategic
management. Strategic Management Journal, 18(7), 509-533.

Tsai, W. 2002. Knowleged transfer in intraorganizational networks: Effects of network

position and absorptive capacity on business unit innovation. Academy of
Management Journal, 44(5), 996-1004.

85

Tung, R. L. 1988. Toward a conceptual paradigm of international business negotiations.
Advances in International Comparative Management, 3, 203-219.

Uzzi, B. 1997. Social structure and competition in interﬁrm networks: The paradox of
embeddedness. Administrative Science Quarterly, 42(1), 35-67.

Van Duijin, M. A. J. 1995. Estimation of a Random Effects Model for Directed Graphs.
In Snijders, T.A.B. (Ed.) SSS '95. Symposium Statistische Software, nr. 7. Toeval
zit overal: programmatuur voor random-coefﬁcient modellen [Chance is
omnipresent: software for random coefﬁcient models], 113-131. Groningen.

Walker, G., Kogut, B., & Sharr, W. J. 1997. Social capital, structural holes and the
formation of an industry network. Organization Science, 8(2), 109-125.

Wasserman, Stanley, & Faust, Katherine. 1994. Social Network Analysis: Methods and
Applications. Cambridge University Press.

Wasserman, S., & Pattison, P. 1996. Logit models and logistic regression for social
networks: An introduction to Markov graphs and P*. Psychometrika, 61, 401-425.

Willer, D. (Ed.). 1999. Network Exchange Theory. London: Praeger.

_ Willer, David, & Skvoretz, John. 1997. Games and structures. Rationality and Society, 9,
5-35.

Williamson, 0. E. 1985. Markets and Hierarchies. New York: Free Press.

Yamagishi, Toshio, Gillrnore, Mary R., & Cook, Karen S. 1988. Network connections
and the distribution of power in exchange networks. The American Journal of
Sociology, 93, 833-851.

Yan, Ainrirr, & Gray, Barbara. 1994. Bargaining power, management control, and
performance in Uni. Academy of Management Journal, 37, 1478-1517.

Yeoh, Poh-lin, & Roth, Kendall. 1999. An empirical analysis of sustained advantage in
the US. pharmaceutical industry: Impact of ﬁrm resources and capabilities.
Strategic Management Journal, 20, 637.

Zaheer, A., & Bell, G. G. 2005. Beneﬁting ﬁ'om network position: Firm capabilities,
structural holes, and performance. Strategic Management Journal, 26(9), 809-
825.

Zaheer, A., & Venkatraman, N. 1995. Relational governance as an interorganizational

strategy - an empirical-test of the role of trust in economic exchange. Strategic
Management Journal, 16(5), 373-392.

86

         

MICHGAN STATE UNI ERSII'Y LIBRA E5
1

illiiiiliilii iii Hilliiiiiiiiiiiiiill/Hi

3 1293 02 45 8218