 

3.41.:

‘ I!
7...

r
. nﬁwM m 1

JV

1 .{1

m

5...}. a...» .
~1vutn42w ‘
Quhnmm‘ . u,
:1. 23...”?
, Anglia

$1.233?
1.3.1: .23 I
. 3.. ,. ”kw...“ mi. .

«bans... i.
2! a V: dd.

2
3:1. . . .i‘. A .
5! .92 3‘50.“ .ﬂnf! .
lights. 5&5! am.
5...!
, I wanna: . tot?
iixﬁﬂ

.ww‘tw-Juxu..i.r. ‘
1 ‘ . #43:..‘1. a .

‘ . 91m.” ‘ tur4$mnnu.s..ur x , .
. . A. ﬂywﬁrgﬁw ,,
.w‘ ... Jami»;-

‘ iii:

 

w ﬁ.

‘t'x‘mms

8cm

This is to certify that the
dissertation entitled

ENHANCING INNOVATION CAPABILITY THROUGH
SUCCESSFUL INTER-FIRM COLLABORATIONS: TWO

LIBRARY
ichigan State
' '“i'érsftv

 

Li

PhD

ESSAYS ON R&D ALLIANCES

presented by

SHICHUN XU

has been accepted towards fulﬁllment
of the requirements for the

degree In Business Administration

 

 

(1)[’7W Oil/VLJI/J/

 

Major Professor’ 3 Signatf’l/
Q 5/ jar AL :00 5)-

Date

 

MSU is an afﬁrmative-action, equal-opportunity employer

—-~— --.--.--.o-o.-.--.-...-.-.--..-.--.-

- —--.-.-u-.-.--u-a-n--.-----.-n----.--.-:-u-a.

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:lClRC/DateDue.indd—p.1

ENHANCING INNOVATION CAPABILITY THROUGH SUCCESSFUL INTER-
F [RM COLLABORATIONS: TWO ESSAYS ON R&D ALLIANCES

By

Shichun Xu

A DISSERTATION
Submitted to
Michigan State University

in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Marketing and Supply Chain Management

2007

ABSTRACT

ENHANCING INNOVATION CAPABILITY THROUGH SUCCESSFUL INT ER-
FIRM COLLABORATIONS: TWO ESSAYS ON R&D ALLIANCES

By
Shichun Xu

Innovation capability plays a critical role in establishing a ﬁrm’s competitive
advantage. With fast changing technologies and soaring costs, R&D alliances have
become an important tool for ﬁrms to effectively conduct their new product development
activities. Most research in this area has focused on inter-ﬁrm agreements in isolation,
with a special emphasis on understanding dyadic information transfer and coordination.
This two-essay dissertation addresses the R&D alliance issues from a portfolio approach
by examining ﬁrrns’ whole inventory of alliance activities. We argue that the
sustainability of a ﬁrm’s competitive advantage comes primarily from the successful
conﬁguration of its R&D alliance portfolio more so than from the successful management
of any particular R&D alliance project.

Essay One seeks to answer the research question: how do different characteristics
of a ﬁrm’s R&D alliance portfolio affect its ability to acquire external knowledge which
in turn inﬂuences its innovation outcomes. A set of characteristics of R&D alliance
portfolio are identiﬁed from both management and marketing literatures (vertical
alliances vs. horizontal alliances, repeated partnering, multiple-partner alliances,
multiple-technology alliances, and partner dispersion). Arguments are made that these
characteristics inﬂuence both the number of knowledge domains (knowledge breadth)
and the intensity of knowledge sharing between alliance partners in any particular area

(knowledge depth).

While literature suggests both knowledge breadth and knowledge depth are
positively associated with a ﬁrm’s innovation output, a further argument is made in terms
of the differential impacts of knowledge breadth and knowledge depth on radical
innovation and incremental innovation. We hypothesize that radical innovations are
impacted more by the diversity and novelty of incoming knowledge than by its quantity
while incremental innovations are impacted more by the relatedness and quantity of the
incoming information than by its novelty or newness.

Essay Two explores the form of the relationship between the number of a ﬁrm’s
R&D alliances and its innovation output. Past literature has normally hypothesized either
a linear positive or inverse U-shaped relationship. We theoretically argue, from
organizational learning and integrative capability perspectives, that the relationship
should be inverse S-shaped. Our empirical test using data from the pharmaceutical

industry supports our hypotheses.

Copyright by
Shichun Xu
2007

To

My parents, my wife, and my daughter

ACKNOWLEDGEMENTS

I would like to express my sincere gratitude to a number of individuals for their
guidance and support towards the completion of this dissertation as well during my
graduate studies here at Michigan State University.

Foremost, I am indebted to Dr. S. Tamer Cavusgil, my mentor, advisor and
committee chair. Neither this dissertation nor my successful completion of the doctoral
program would have been possible without his constant guidance, encouragement, and
unlimited patience. More than an academic advisor, he has been a role model to me by
being unfailingly helpful, caring, and supportive during my doctoral studies.

My gratitude also goes to my committee members: Dr. Roger Calantone for his
insightﬁil comments on methodology related issues, Dr. David Grifﬁth for his attentive,
constructive, and valuable suggestions throughout this dissertation, Dr. Chris White for
his great help on the theoretical foundation of this work, and Dr. Ahmet Kirca for his
valuable comments and suggestions on the discussion of the results. They each shared
their knowledge and expertise with me to strengthen this research project. Their
constructive comments and suggestions greatly contributed to this research. I feel truly
honored to have all of them serving on my committee.

I also extend my thanks to CIBER at MSU for providing ﬁnancial support for the
collection of data to empirically test the arguments made in this dissertation.

Last but not least, I thank my family for their continuous love and understanding
during this whole process. They have been an inspiration to me all these years and I

dedicate this work to them.

vi

TABLE OF CONTENTS

LIST OF TABLES .............................................................................................................. x
LIST OF FIGURES ........................................................................................................... xi
CHAPTER 1 -

INTRODUCTION .............................................................................................................. 1
1.1 The Importance of R&D Alliances ..................................................................... 1
1.2 The Importance of Alliance Portfolio Management ........................................... 3
1.3 Purpose of This Dissertation ............................................................................... 5
1.4 Contribution of the Dissertation .......................................................................... 8

CHAPTER 2

THEORETICAL FOUNDATIONS OF STRATEGIC ALLIANCES ............................. 10
2.1 Resource-Based View and Alliance Partner Selection ........................................... 10
2.2 Knowledge-Based View and R&D Alliances ......................................................... 13
2.3 Transaction Cost Analysis and Alliance Governance ............................................. 15
2.4 Social Network Theory and Firm Performance ...................................................... 20
2.5 Organizational Learning and Knowledge Acquisition ........................................... 23

CHAPTER 3

ESSAY ONE: ALLIANCE PROTFOLIO CONFIGURATION, KNOWLEDGE
DEVELOPMENT, AND INNOVATION: THEORETICAL FRAMEWORK AND

EMPIRICAL EVIDENCE ................................................................................................ 35
3.1 Development of the Conceptual Framework .......................................................... 35
3.2 Hypotheses Development ....................................................................................... 42

3.2.1 Vertical and Horizontal R&D Alliances and Knowledge Development ......... 42
3.2.2 Partner Dispersion and Knowledge Development ........................................... 44
3.2.3 Partner Repeatedness and Knowledge Development ...................................... 47
3.2.4 Multiple Partner R&D Alliances and Knowledge Development ..................... 50

vii

3.2.5 Alliance Scope and Knowledge Development ................................................ 52

3.2.6 Knowledge Development and Innovation Output ........................................... 54
3.2.7 Innovation and Firm Financial Performance ................................................... 56
3.2.8 Control Variables ........... . .................................................................................. 57
3.3 Research Methodology ........................................................................................... 59
3.3.1 Empirical Setting ............................................................................................. 59
3.3.2 Sample .............................................................................................................. 60
3.3.3 Data Sources .................................................................................................... 63
3.4 Measures ................................................................................................................. 66
3.5 Analysis and Findings ............................................................................................. 74
3.5.1 Model Speciﬁcation ......................................................................................... 74
3.5.2 Statistical Results ............................................................................................. 75

3.5.2.1 R&D Alliance Portfolio Characteristics and Knowledge Development... 75

3.5.2.2 Knowledge Development and Innovation Output .................................... 77
3.5.2.3 Innovation and Firm Financial Performance ............................................ 80
3.6 Discussion ............................................................................................................... 82
3.6.1 Theoretical Contribution .................................................................................. 83
3.6.2 Managerial Contribution .................................................................................. 86
3.6.3 Limitations and Future Research ..................................................................... 89
CHAPTER 4
ESSAY TWO: NON-ADDITIVITY EFFECT OF R&D ALLIANCE PORTFOLIO SIZE
ON INNOVATION: AN INTEGRATION CAPABILITY PERSPECTIVE ................... 93
4.1 Introduction ............................................................................................................. 93
4.2 Received Wisdom on R&D Alliance Portfolio Size and Innovation. ..................... 96

viii

4.3 Theoretical Foundation and Hypotheses Development .......................................... 99

4.3.1 Diminishing Returns with Limited Number of R&D Alliances ...................... 99
4.3.2 Increasing Returns with Expanded Number of R&D Alliances .................... 102
4.3.3 Control Variables ........................................................................................... 110
4.4 Methodology ......................................................................................................... 112
4.4.1 Empirical Setting ........................................................................................... 112
4.4.2 Sample ............................................................................................................ 112
4.4.3 Data Sources .................................................................................................. 113
4.5 Analysis and Findings ........................................................................................... 114
4.5.1 Model Speciﬁcation ....................................................................................... 114
4.5.2 Statistical Results ........................................................................................... 115
4.6 Discussion ............................................................................................................. 118
4.6.1 Contribution ................................................................................................... 118
4.6.2 Limitations and Future Research ................................................................... 122
REFERENCES ............................................................................................................... 125

ix

LIST OF TABLES

Table 1 Select List of Literature on Strategic Alliances ................................................... 29
Table 2 Summary of Sample Characteristics .................................................................... 62
Table 3 Constructs Operationalization and Data Sources ................................................. 71
Table 4 Correlation Matrix of Key Constructs ................................................................. 73
Table 5 Alliance Portfolio Characteristics on Knowledge Development ......................... 78
Table 6 Knowledge Development and Innovation Outcome ............................................ 79
Table 7 Innovation and Firm Financial Performance ....................................................... 79
Table 8 Summary of Hypotheses Testing Results ............................................................ 81
Table 9 Number of R&D Alliances and Innovation Output ........................................... 117

LIST OF FIGURES

Figure 1: The Conceptual Model ...................................................................................... 41
Figure 2: Data Sources for Empirical Testing .................................................................. 65
Figure 3: Relationship between Number of Alliances and Innovation Output ............... 111

xi

CHAPTER 1
INTRODUCTION
1.1 The Importance of R&D Alliances

New product development plays a critical role in establishing a ﬁrm’s competitive
advantage in the market place. Successful new products often incorporate the most recent
development in technologies and thus offer superior customer beneﬁts than existing
products in the market. As such, new products act as important engines for economic
grth (2003) as well as provide an important source of sales, proﬁts and competitive
strength for organizations (Sivadas and Dwyer 2000). It is estimated that over forty
percent of the sales for most large organizations come from products introduced in the
past ﬁve years (Schmidt and Calantone 2002) and new products provide about one-third
of the proﬁts for Fortune 1000 companies (Booz and Hamilton 1982; Sivadas and Dwyer
2000). As a result, organizations devote considerable amount of resources to new product
development realizing that their survival in the market depends on their ability to
continuously introduce new products that meet customer needs over time (Dougherty and
Hardy 1996; Nerkar and Roberts 2004).

However, most ﬁrms are ﬁnding it increasingly difﬁcult to develop new products
purely internally (Sivadas and Dwyer 2000; Wind and Mahajan 1997). First of all, new
product development is a highly demanding endeavor in terms of both ﬁnancial and
human capital resources. Firms need to make huge resource commitment upfront without
the guarantee that a successful new product will be developed. It is estimated that the
average cost of developing a new drug in the United States is over 800 million dollars

and this ﬁgure is increasing by over 10 percent annually (GAO 2006). With an

investment this big, ﬁrms are not guaranteed any conceivable result for
commercialization. In fact, it is estimated that over forty percent of new product
development projects fail to come up with anything new (Schmidt and Calantone 2002).
As such, ﬁrms with ﬁnancial constraints may ﬁnd it difﬁcult to initiate new product
development projects all by itself. Secondly, the risk associated with new products
extends beyond the development stage. Only half of the introduced new products will
survive the market test during commercialization (Page 1993; Schmidt and Calantone
2002; Zirger and Maidique 1990). As such, ﬁrms face considerable risk in their new
product development endeavors.

However, the most important reason ﬁrms cannot take on new product
development activities all by themselves may have something to do with the changing
competitive environment in the market place. In order to develop a truly innovative
product, ﬁrms need to pool knowledge from different areas(Cohen and Eliashberg 1997;
Kogut and Zander 1992; Prabhu et al. 2005). Increasing complexity in new product
development and fast advancing technologies often mandate ﬁrms to look beyond their
boundaries to access knowledge and technology for their new product development
purposes (Sood and Tellis 2005; Wuyts et al. 2004). In the agricultural industry, Cargill’s
Chief Technology Ofﬁcer indicates that to bring an innovative product to the market
place requires “. . .so much cooperation and integration of knowledge that you just can’t
get it done unless you pick partners” (Forbes Magnetic, 40, 2001, P. 66;Ireland et a1.
2002)

As a result, inter-ﬁrm collaborations on R&D activities have become common

practice for ﬁrms to: overcome the ﬁnancial constraints; share the inherent risk associated

with new product development; speed up the NPD process; and to get access to the most
advanced technologies not available within a ﬁrm (Sivadas and Dwyer 2000; Wuyts et al.
2004). R&D alliances are deﬁned as the ongoing, formal cooperative arrangements
between two or more ﬁrms to achieve new product development goals (Sivadas and
Dwyer 2000). Echoing the popularity of R&D alliances in business, marketing scholars
have contributed considerable understanding to the issues of R&D alliances over the past
two decades (e. g. Mohr et al. 1996; Rindﬂeisch and Moorman 2001; Rindﬂeisch and
Moorman 2003; Sivadas and Dwyer 2000; Wuyts et al. 2004). Most of these studies have
treated R&D alliances in isolation as the unit of analysis is normally at the project level
(Wuyts et a1. 2004), with particular emphasis on partner selection, individual alliance
management, and knowledge transfer and coordination (Wuyts et al. 2004). Their
emphasis is thus on the understanding of how each individual R&D alliance can be
effectively managed.
1.2 The Importance of Alliance Portfolio Management

In spite of the extent literature on alliance management, very few studies have
focused on the management of alliance portfolios as a whole (Bamford and Ernst 2002;
Wuyts et al. 2004). Firm performance are very much tied to the successful management
of entire portfolios of inter-ﬁrm agreements rather than the successful management of
any one particular alliance (Wuyts et al. 2004). The sustainability of a ﬁrm’s competitive
advantage requires a ﬁrm’s ability to beneﬁt from its alliances activities over time.

Organizations enter into dozens of R&D alliances. A conservative estimate would
be 30 alliances for most of them and many have more than 100 (Bamford and Ernst

2002). For example, in the aerospace industry, United Technologies is involved in over

100 worldwide collaborations (Ireland et a1. 2002). For both practitioners and academia,
however, there is a lack of understanding about how ﬁrms can successfully manage their
alliances portfolios as a whole to achieve their competitive goals (Bamford and Ernst
2002; Hoffrnann 2007; Wuyts et al. 2004). According to Bamford and Ernst (2002, p.
30-37):

A company’s alliance portfolio too often grows into a random mix of

ventures assembled over the years by a variety of business units... Even if a

deal made sense when ﬁrst negotiated, the portfolio is unlikely to be as good

as it could be in View of the current strategy of the company or even of a

business unit. Companies often fail to recognize performance patterns across

their alliance portfolios- patterns concerning particular deal structures, types

of partners, or functional tasks. A failure to spot and ﬁx such recurring

problems can be costly. . .. Once a company better understand how its

portfolio is performing, it can conduct a top-down review of overall strategy

in order to ensure not only that its alliance portfolio is conﬁgured in the best

possible way and contributes sufﬁciently to its performance but also that it

has ranked new opportunities in a clear order of priority.
They further suggest that:

(To achieve this), the ﬁrst consideration is the performance of the portfolio

as a whole. With a coherent strategy and an appropriate linking of partners,

a portfolio can be more than its sum.

Portfolio’s conﬁguration is the second consideration. Does the company
have the right set of alliances and the appropriate level of commitment to
each?

The ﬁnal element in a strategy review is to rank future initiatives in order of
priority.

It has been shown in the literature that the portfolio descriptors have an important
inﬂuence of a ﬁrm’s innovation outcomes (Bamford and Ernst 2002; Wuyts et al. 2004).
A portfolio approach, which examines the overall effect of the whole collection of
strategic alliances of a ﬁrm, might be a more appropriate unit of analysis to our question
(Wuyts et al. 2004). A ﬁrm’s competitive position is more of a result of the history and
accumulation of knowledge rather than a result of single alliance outcomes. Therefore,
competitive advantage is and should be viewed as a result of successful management of
alliance portfolios (Slowinski 2001).

Another research area that demands further attention is the differential
roles of incremental innovation and radical innovation. Even though scholars have
documented the different characteristics and performance impacts of these two
types of innovations (Prabhu et al. 2005; Sood and Tellis 2005; Wuyts et al.

2004), it is not clear though what knowledge requirement differences exist for the
development of these two different types of innovation.
1.3 Purpose of This Dissertation

This two-essay dissertation aims to facilitate the understanding of successful

R&D alliance portfolio management. The ﬁrst essay focuses on identifying key R&D

alliance portfolio characteristics and explores their inﬂuence on ﬁrms’ ability to develop

knowledge reservoir on both breadth and depth dimensions. We further investigate the
differential impact of knowledge breadth and knowledge depth on incremental and
radical innovation development.

We take an exploratory approach, which involves an extensive literature review
and in-depth discussions with industry executives, to identify key strategic decisions
managers need to make in formulating their R&D alliance portfolios. The identiﬁed
alliance portfolio characteristics include mangers’ preferences of incorporating vertical
partners vs. horizontal partners into their portfolio (Dussauge et al. 2000; Rindﬂeisch and
Moorman 2001), tendency to work with the same partner over time (Wuyts et al. 2004),
likelihood of building multiple technologies into their agreements (Khanna et al. 1998;
Oxley and Sampson 2004), likelihood of bringing in unfamiliar partners (Goerzen and
Bearnish 2005), and propensity of involving multiple partners in their R&D alliances
(Zeng and Chen 2003).

Due to the different levels of knowledge heterogeneity and governance challenges
associated with each of these characteristics, a conceptual framework is proposed based
on arguments ﬁ'om knowledge-based view and transaction cost analysis, which identiﬁes
the relationship between these key characteristics and a ﬁrm’s knowledge breadth and
depth development. In order to help organizations align their alliance portfolio
management with their corporate strategies, ftnther arguments are made in terms of the
differential impact of knowledge breadth and depth on incremental innovation and radical
innovation(Atuahene-Gima 2005; Chandy et al. 2003; Chandy and Tellis 2000; Chandy
and Tellis 1998; Cooper 2000; Sood and Tellis 2005; Sorescu et al. 2003; Wuyts et a1.

2004). Due to the different level of innovativeness in a new product, incremental

innovation and radical innovation differs in their development processes in terms of
which knowledge dimension (i.e. breadth and depth) plays a major role in the
development success. Finally, in view of the inconsistent ﬁndings in the literature, we
hypothesize relationships between innovation and ﬁnancial performance.

Drawing on the combinative capability literature, which refers to ﬁrms’ ability to
integrate different knowledge domains in their innovation effort, the second essay aims to
seek answer to one key unattended question: the nature of the relationship between a
ﬁrm’s number of R&D alliances and its innovation output. Existing literature on this
question offers divergent ﬁndings. Some studies have found a positive liner relationship
between the number of R&D alliances of a ﬁrm and its innovation outcome (e. g., Baum
et a1. 2000; Rothaerrnel and Deeds 2004; e. g., Shan et al. 1994). Other studies though
have argued a non-linear relationship between the size of a ﬁrm’s R&D alliance portfolio
and its innovation output (Rothaerrnel and Deeds 2004). As such, it is very inconclusive
as to the true ﬁrnctional from of the impact of the number of R&D alliances on
innovations.

We argue that literature in this area has ignored an important capability of
organizations, their ability to integrate knowledge sectors across partners (Henderson and
Cockburn 1994). Based on this literature, we hypothesize an inverse S-shaped
relationship between a ﬁrm’s number of R&D alliances and its innovation output.
Speciﬁcally, we argue that the marginal return of adding a new alliance to the portfolio
would be increasing once a critical mass of knowledge is reached, which allows for

options of multiple combinations among the newly added technology and existing ones.

To overcome the low response rate and common method bias that are prevalent in
survey based studies, this dissertation uses secondary data to empirically test the
hypothesized relationships. We collect data from multiple industry sources in order to
provide accurate proxies for our constructs for empirical testing purposes.

1.4 Contribution of the Dissertation

This dissertation contributes to the literature by offering a framework guiding the
successful management of R&D alliance portfolios. By explicitly linking each
managerial preference to knowledge development outcomes, it offers guidelines as to
what type of portfolio would help organizations achieve their knowledge development
goals. It offers ideas about what type of partner agreement ﬁrms should pursue as well as
those they need to avoid in their portfolio conﬁguration process.

In addition, this dissertation also offers new insights about the different
knowledge development strategies that would help ﬁrms achieve their innovation
objectives. Because of differences in the corporate strategic emphasis on incremental vs.
radical innovations, ﬁndings in this dissertation will help ﬁrms align their knowledge
development strategies with their innovation goals. This, combined with insights of
relationships between alliances attributes and knowledge development, should provide a
complete picture about strategies ﬁrms should pursue in their innovation effort.

This dissertation is structured as follows: Chapter 1 illustrates the motivation and
importance of the dissertation. Chapter 2 will provide an overview of existing streams of
literature on R&D alliance management. Chapter 3 focuses on Essay one. More
speciﬁcally, a knowledge-based conceptual model of successful R&D alliance portfolio

management is presented. In this chapter, literature on knowledge-based view and

transaction cost analysis are integrated to support the conceptualization and hypotheses
development. Research methodology, statistical analysis, ﬁndings, and discussion of the
ﬁrst essay are also included in this chapter. Chapter 4 focuses on the second essay.
Hypotheses development, research methodology, statistical analysis, ﬁndings, and

discussions of the results of the second essay are all included in this chapter.

CHAPTER 2
THEORETICAL FOUNDATIONS OF STRATEGIC ALLIANCES

This section presents a comprehensive review of recent theoretical research on
strategic alliance management literature, with an emphasis on R&D alliances. Strategic
alliance embraces a diversity of causes, collaborative forms, and consequences (Grant
and Baden-Fuller 2004). The proliﬁc phenomenon associated with alliances challenges
scholars’ ability to develop an all-encompassing theory (Grant and Baden-Fuller 2004).
As such, a wide array of theories has been applied in scholarly works in the area. The
most commonly used theories in the alliance research are: l) Resource-based review; 2)
Knowledge-based view; 3) Transaction cost analysis; 4) Social network theory; and 5)
Organizational learning theory. To date, researchers have used these theories to
understand the formation, management, and performance implication of strategic
alliances.
2.1 Resource-Based View and Alliance Partner Selection

The resource-based view of ﬁrm holds that a ﬁrm is a collection of heterogeneous
resources (Wernerfelt 1984). Firms can obtain sustained above normal returns only when
they possess superior resources which are protected by some form of isolating
mechanisms tying the resources within the boundary of a ﬁrm (Barney 1991; Barney
2001; Kor and Mahoney 2004; Penrose 1959). As such, resources heterogeneity is the
foundation for competitive advantage, and management’s task, according to RBV, is to
identify, develop, and deploy the key resources to maximize its return (Barney 1991).

True resources of a ﬁrm should have the following characteristics:

10

1) Valuable in that they help ﬁrms to conceive or implement strategies more
effectively or more efﬁciently than competitors.

2) Rare in that the resource is only distributed among a limited number of
competitors in the industry so that the rent cannot be diffused across ﬁrms.

3) Imperfectly imitable in that competitor cannot easily copy away the resources
due to their unique development history, causal ambiguity, and social
complexity.

4) Non-substitutable in that no equivalent strategic resources that by themselves
either not rare or imitable.

As the market conditions become increasingly complex, firms seldom possess all
the necessary resources to create competitive advantage or even achieve a particular
marketing objective such as new product development all by itself (Ireland et al. 2002).
As such, resource based view help us understand the motivation behind inter-ﬁrm
alliances (Eisenhardt and Schoonhoven 1996).

It is agreed that a dominant motivation behind the formation of alliances is both to
gain access to partner-speciﬁc resources and to jointly develop new resources through
collaborations between partners (Ireland et al. 2002; Reid et al. 2001). Alliances provide
access to technology, ﬁnancial assets, information, and markets that a ﬁrm will ﬁnd
difﬁcult to obtain otherwise (Hitt et al. 2001). Accessing and developing these resources
is key to a ﬁrm’s competitive position in the market. Glaister and Buckley (1996) found
that resource access has a greater explanation power than other causes such as risk
sharing and development of economies of scale in alliance formations (Ireland et al.

2002)

11

Besides shedding light on the general motivation of strategic alliance formation,
the resource-based View also explains issues associated with partner selection. It is found
that ﬁrms are more inclined to partner with ﬁrms with complementary resources than
with ﬁrms possessing similar resources (Chung et al. 2000; Gulati et al. 2000). This
ﬁnding applies to both large, resource rich ﬁrms from developed countries as well as
small, resource scarce ﬁrms from developing countries (Hitt et al. 2000). As such, the
resource composition of a ﬁrm relative to the market need plays a central role in its
alliance activities (Stuart 2000).

Both a ﬁrrn’s resource inventory and the market need will have motivational
values for alliance formations. For ﬁrms to be sought after as partners, they must be
endowed with resources that other ﬁrms value but do not have (Das and Bing-Sheng
2000). On the other hand, ﬁrms without valuable resources will not be desirable partners,
which inhibit their ability to locate the optimal resource conﬁguration in which the value
of their resource could be maximized due to their inability to offer the reciprocal beneﬁts
(Ahuja 2000b). Newly emerging market and innovative industries will prompt ﬁrms to
actively seek alliances to maintain completive in the market place (Eisenhardt and
Schoonhoven 1996).

The resource endowment of alliance partners also inﬂuences the outcome of the
collaboration. Dussauge, Garrette and Mitchell (2000) observe more frequent
reorganization and takeover when partner ﬁrms share similar resources. Harrison, Hitt,
Hoskisson and Ireland (1991) also argue that alliances between partners with similar
resources would not perform very well as they provide opportunities for exploiting based

collaborations but prevents exploration based collaborations. This ﬁnding is supported by

12

Madhok and Tallman (1998) who argue that partners with complementary resources offer
the potential for integrating complementary resources and thus create synergies to
achieve new competitive advantage. Rindﬂeish and Moorman (2001) also ﬁnd that ﬁrms
with similar resources will experience inferior knowledge development outcome.

2.2 Knowledge-Based View and R&D Alliances

Even though resources can take on many forms such as assets, technologies,
information, process, or even relationships, most scholars agree that true competitive
advantage comes from a ﬁrm’s access to its intangible idiosyncratic resources, those that
are tacit and knowledge-based (Dussauge et al. 2000). As knowledge gains its focal
position in explaining ﬁrm performance differences, a “knowledge-based view” emerged
in the literature (Grant 1997; Grant 1996; Grant and Baden-Fuller 1995). As an
outgrth of the traditional resource-based view, knowledge-based view focuses on
knowledge as the most strategically important resource of a ﬁrm. Fundamental to the
knowledge-based view is that knowledge represents the single critical input in production
and the primary source of value (Grant 1996). However, because of the bounded
rationality, “efﬁciency in knowledge production requires that individual (ﬁrm)
specializes in particular areas of knowledge” (Grant 1996 p.112 ).

While resource-based view offers theoretical explanation to the general alliance
formations such as franchising (ﬁrms with product resources ally with ﬁrms possessing
market resources) and outsourcing (ﬁrms with intellectual resources work with ﬁrms with
production resources), knowledge-based view offers great insights to a particular type of
inter-ﬁrm collaboration, the R&D alliances (D02 1998; Grant and Baden-Fuller 2004). In

explaining the formation and motivation issues of R&D alliances, several studies have

13

identiﬁed accessing knowledge as their major objective (Dyer and Nobeoka 2000; Grant
and Baden-Fuller 2004; Inkpen and Crossan 1995; Lyles and Salk 2007; Sirnonin 1999).

The knowledge-based view distinguishes two dimensions of knowledge
management through alliance formations: knowledge acquiring and knowledge
accessing. First, alliances may be formed for the purpose of augmenting a ﬁrm’s
knowledge inventory. This could be achieved either by learning from alliance partners or
by jointly developing new knowledge with partners. This is referred to as “exploration”
by March (1991) or “knowledge generation” by Spender (1992). Second, ﬁrms may be
motivated to form alliances primarily for the purpose of deploying existing knowledge to
create value without actively changing the knowledge base. This is referred to as
“exploitation” by March (1991) or “knowledge application” by Spender (1992). Firms
simply pool resources together in this case without actively learn or internalize
knowledge from partners.

Grant and Baden-Fuller (2004) identiﬁed circumstances where alliances could
serve better mechanisms for knowledge related activities than pure market transactions
and better than merger and acquisitions. First, alliances can avoid many of the costs
associated with knowledge transaction by limiting opportunism since both parties need to
foster investment in trust (Simonin 1997; Teece 1992). Second, under highly diversiﬁed
knowledge domains, alliances could be better integration mechanism than the market or
ﬁrm. According to Grant and Baden-Fuller (2004) “as the range and diversity of
knowledge increases, so integration mechanisms need to be increasingly differentiated,
resulting in rising marginal costs of knowledge integration within the ﬁrm. In these

circumstances, efﬁciency of integration may be maximized through separate ﬁrms

l4

integrating knowledge at the component or subsystem level, with overall integration
through an alliance between ﬁrms” (p.69). Third, alliances offer better mechanism for
knowledge utilization when the product domain of the ﬁrm and the knowledge domain of
the ﬁrm are highly incongruent (Grant and Baden-Fuller 2004).

Knowledge transfer across partners is another heavily researched topic in alliance
literature. Effective transfer of knowledge from partner to partner or from parent ﬁrms to
the joint venture is critical for alliance success (Mowery et al. 1996). The effectiveness of
knowledge transfer is heavily inﬂuence the nature of the knowledge (Simonin 1999).
Explicit knowledge is much easier to transfer than tacit knowledge and knowledge
ambiguity inhibits the success transfer of knowledge (Simonin 1999). Explicit knowledge
can be easily codiﬁed into written forms, which allows for a much easier channel of
communication of the content. Tacit knowledge, on the other hand, has codiﬁcation
difﬁculties. The transfer of tacit knowledge thus requires more frequent and more
personal contacts between partners in order to fulﬁll a successful transfer.

The overall conclusion concerning the transferability of knowledge is that
tacitness, complexity, cultural distance and organizational distance between partners are
all positively related to ambiguity which in turn inhibits knowledge transfer (Simonin
1999). However, alliance partners could overcome these transfer difﬁculties by
instituting structural mechanism such as training, internal consulting, and assistance
(Ireland et al. 2002; Lyles and Salk 2007). Inkpen and Dinur (1998) advanced the
importance of knowledge management strategies as certain types of strategies outperform
others when knowledge types are different.

2.3 Transaction Cost Analysis and Alliance Governance

15

Transaction cost analysis maintains that the minimization of total production and
transaction costs will determine the boundaries of ﬁrms (Hemphill and Vonortas 2003;
Williamson 1991). Williamson (1991) states that certain dimensions of transactions give
rise to transaction costs that allow the evaluation of alternative transaction structures or
governance mechanisms. Due to environmental uncertainty which give rise to adaptation
problem; behavioral uncertainty which leads to performance evaluation problem; and
asset speciﬁcity which induces safeguarding problem, transactions incur different levels
of costs depending on the completeness of contract to be agreed upon upfront
(Williamson 1975; Williamson 1998).

One important behavioral assumption in transaction cost analysis is
opportunism. It assumes that given the opportunity, ﬁrms may seek to serve their own
interest at the partner’s cost (Rindﬂeisch and Heide 1997; Williamson 1994).
Opportunism poses a problem to the extent that relationships may require relationship-
speciﬁc investment from parties and this would create safeguarding problem (Rindﬂeisch
and Heide 1997).

Choice of governance mechanisms will be a function of the transaction cost and
the production cost advantage of the market (Rindﬂeisch and Heide 1997). In particular,
two types of governance mechanisms are emphasized: ﬁrms and market. Under perfect
market competition where costs associated with adaptation, safeguarding and
performance evaluation is low, given the high-powered price incentives, market is the
choice of transaction by default. However, when market is imperfect, costs associated
with contract enforcing and performance monitoring will increase which necessities a

uniﬁed hierarchical governance structure, ﬁrm (Williamson 1991).

16

High transaction cost is a result of incomplete contract as the contract fails to fully
specify all possible actions of each party in all states of the world (Hemphill and
Vonortas 2003). Alliance partners often increase the complexity of their contracts when
high asset speciﬁcity is involved (Reuer and Arino 2007). The circumstances around
strategic alliances though, especially R&D alliances where technologies are the focal
resources, often are conducive to high costs for writing a complete contract (Hemphill
and Vonortas 2003). In an alliance where the shared goal is to develop a new technology,
incomplete contracting for property rights is almost inevitable as the contracted asset do
not even exist at time of contracting (Oxley 1997).

Both environmental uncertainly and behavioral uncertainty are present in
technology based alliances. Environmental uncertainty such as the market and technology
uncertainties due to the ﬂuid stage of development, lack of dominant design, and ill-
deﬁned technological trajectories makes writing a complete contract almost impossible
(Hemphill and Vonortas 2003). Firms in an alliance, especially a technology-based
alliance, are also subject to opportunistic behaviors from their partners (Oxley 1997).
These opportunistic behaviors could range form delivering less technology than speciﬁed
in contract to using partner’s technology in areas other than speciﬁed in the contract
(Oxley 1997). When uncertainties present, writing a complete contract that would
incorporate all the possible courses of future thus become difﬁcult due to the bounded
rationality of management (Rindﬂeisch and Heide 1997).

Knowledge as the major resource for alliance activities also brings about some
unique difﬁculties. Knowledge, especially tacit knowledge, has approriability problems

(Grant 1996). Appropriability refers to the ability for the owner of a resource to receive a

17

return equal to the value of the creation of the resource (Grant 1996; Oxley 1997). Both
tacit and explicit knowledge suffer from appropriability problems. For tacit knowledge, it
is not appropriable because it cannot be directly transferred. It can only be appropriated
through it application to productive activities (Grant 1996). For explicit knowledge,
anyone who acquires it can resell it without losing it (Arrow 1971; Grant 1996). Second,
the marketing process itself would in effect make it available to potential buyers (Arrow
1971; Grant 1996). Thus unless the knowledge is protected by legal establishment such as
patents or copyright, knowledge is not appropriable by means of market transactions
(Grant 1996).

Due to the appropriability hazards in alliance activities involving tacit know-how,
full speciﬁcation of the assets to be transferred and monitoring partner performance and
behavior become problematic since both parties have incentives to act opportunistically.
When parties face these problems, they are inclined to mitigate against the contracting
hazard by moving up the market-hierarchy continuum of alliance forms (Oxley 1997). As
a result, an equity-based alliance would emerge as it offers the shared equity, additional
monitoring rights, and increased administrative control. In fact, any factors that may add
to the difﬁculty of writing a complete contract such as number of partners, technology
scope, as well as number of ﬁrms involved, may give alliance partners incentives to move
to a more hierarchical governance structure (Oxley 1997).

Equity alliance governance is not the only solution to the possible opportunism
problems. Alliance governance mechanism studies also drew from the relational
exchange theory in which trust is a focal construct. Trust is deﬁned as a ﬁrm’s belief that

its transaction party will perform actions that will result in positive outcomes for the ﬁrm,

18

and will not take unexpected actions that would result in negative outcomes (Anderson
and Narus 1990). The essence of trust is to alleviate opportunism (Gulati 1995).
Predictability, dependability, and faith are the key components of trust (Andaleeb 1992).
As such, trust suggests that alliancepartrrers can depend on each other’s actions without
worrying about opportunistic behaviors (Ireland et a1. 2002). Relational exchange theory
suggests that trust evolves from past experiences and as such the history of the
cooperation between partners is negatively related to the perceived opportunistic behavior
of their partners (Parkhe 1993).

The trust-based approach to alliance management has received considerable
empirical support. Studies have shown that trust between alliance partners emerge ﬁom
prior collaborations that would be working as a substitute for formal governance (Reuer
and Arino 2007). Oxley (1997) demonstrates that ﬁrms with prior collaborations are less
likely to choose equity alliances than partners without prior collaborations. Japanese
alliance partners, due to their inclination to ally with prior partners, experience less
transaction costs than those in the United States (Dyer 1997).

To summarize, transaction-cost analysis have served as a powerful tool for
scholars to understand how alliances could be effectively governed from opportunistic
behaviors. The general conclusion is that when there is high uncertainty both in terms of
partner actions and future environmental development, a more hierarchical governance
structure is preferred to prevent any unwanted actions from either party involved in the
alliance (Oxley 1997). In addition, ﬁrms can develop relational norms such as trust as a
substitute for this formal governance structure through repeated partnering experiences

(Dyer 1997).

19

2.4 Social Network Theory and Firm Performance

As most ﬁrms either form multiple alliances with different partners or involve
multiple partners in a single alliance agreement, forming alliances actually serve as a
gateway for ﬁrms to get into a broader network system through both direct and indirect
ties (Gulati 1998). Recently scholars have brought a network perspective in
understanding the beneﬁts and challenges associated with alliance management (Gulati
1999). Research applying social network theory largely falls into two major streams:
network structure analysis and network content analysis.

Two opposing views were expressed in the literature in terms of what structure
offers members within the network the greatest beneﬁts (Walker et al. 1997). Proponents
of closed network emphasize the importance of social capital. Social capital is deﬁned as
the established norms of cooperation among ﬁrms within a network system (Coleman
1993). Social capital is regarded as a resource that helps the development of social norms
for acceptable behaviors among members (Walker et al. 1997). Due to the high ﬁ'equency
of interactions among members within a closed network, information regarding any
member’s deviant behavior is rapidly diffused and consequently sanctioned. As such,
self-interest seeking opportunistic behaviors are constrained among members which
enhances cooperation among members (Walker et al. 1997).

On the other hand, proponents for open networks stress the importance of
information diversity beneﬁt. They content that closed network prevents the inﬂow of
fresh information and hinders the innovative efforts of members (Burt 1992). One
construct, the structural hole, has gained considerable attention in the literature. Structural

holes refers to the sparse regions between the dense network regions (Burt 1992; Walker

20

et a1. 1997). Burt’s (1992) argument is that ﬁrms in a closed network system lose
independence in terms of their partner selection. However, an open network offers greater
latitude in terms of whom they want to work with. This becomes an important attribute of
open network system in innovation-driven ﬁrms, as they are more likely to obtain up to
date technology and fresh information (Burt 1992).

At the micro level, studies have shown that ﬁrrns’ position within an alliance
network has important implications to both their behaviors and performances (Powell et
al. 1996). Members within a network differ from each other in terms of their location and
number of ties they have relative to others (Freeman 1979; Freeman 1992). Centrality is a
measure of the contribution of network position to the importance, inﬂuence, and
prominence of an actor in a network (Freeman 1979).

Centrally located members within a network normally have more ties than those
peripheral members and this offers them certain competitive advantage. For example,
centrally located ﬁrms have alternatives to access resources due to their connections with
multiple actors in the network and their dependency on any particular actor is thus greatly
reduced. They also have access to a more diversiﬁed knowledge set because of their
ability to work with more partners. In addition, their central location offers them the
ability to identify unique opportunities for collaborations, themselves involved or not.
They may simply refer third parties and beneﬁt from these brokerage activities (Freeman,
1979). Centrally located ﬁrms in a network system also enjoy high visibility and
reputation, which enhances their ability to attract talent and opportunities (Powell et al.

1996).

21

Firms in a network also differ in terms of how many direct ties and indirect ties
they have. Ahuj a (2000a) speciﬁcally identiﬁed three beneﬁts of direct ties that would
help a ﬁrm’s innovation effort: knowledge sharing, complementarity, and scale. When
ﬁrms collaborate with each other, the resultant knowledge developed is normally greater
than knowledge that could be developed by any partner independently and it would be
available to both parties. In addition, direct collaboration also brings together
complementary resources and technologies as breakthrough innovation normally requires
the simultaneous use of different sets of skills and knowledge base in the innovation
process (Powell et al. 1996). The third beneﬁt of maintaining a large number of direct ties
in a network is the scale economy as large projects undertaken by multiple ﬁrms together
generates signiﬁcantly more knowledge than smaller ones (Powell et al. 1996).

The innovation output of a ﬁrm is also inﬂuenced by indirect ties, those indirect
contacts ﬁrm has though its direct partners (Powell et al. 1996). A ﬁrm’s direct ties bring
the knowledge and skills from their interactions with other ﬁrms to the interactions with
the focal ﬁrm (Gulati and Gargiulo 1999). As such, direct ties offer extended linkage to
the focal ﬁrm by offering access to the resources held by the partner’s partners. When
ﬁrms need particular resources for innovation activities, they normally ﬁrst seek help
from their direct ties in the network. However, when this fails, their partners may offer
referral services by linking them with their own direct ties who has possession of the
focal resource (Powell et al. 1996). As such, ﬁrms also prefer to increase their indirect
ties by working with partners who have extensive linkages by themselves.

Scholars also have started to understand the inﬂuence of the content of alliances

networks, that is the characteristics of the nodes and the qualitative nature of the

22

relationships on ﬁrrn’s performance (Goerzen and Bearnish 2005). Two particular content
factors have been identiﬁed as inﬂuential to ﬁrrn’s performance: repeated partnership and
network diversity.

When ﬁrms need to form an alliance, they have the choice of working with either
familiar partners or unfamiliar partners. Repeated partnering refers to ﬁrms’ preferences
to allay with partners that they have allied with before. From a network perspective, ﬁrms
with linkages to a diversiﬁed network of partners enjoy the beneﬁt of freshness and
novelty in information acquisition. On the contrary, ﬁrms that have a great degree of
redundancy in their partner portfolio are limited in terms of the range of technology,
skills, and marketing information access. As a result, ﬁrm performance will be negatively
affected by a high level of partner redundancy (Goerzen 2007).

Another important network content characteristic in the literature is the network
diversity. Three beneﬁts were identiﬁed for a diversiﬁed network. First of all, ﬁrms in a
diversiﬁed network can access to a large number of partners. Secondly, ﬁrms normally
enjoy timing advantage in terms of information in the market when they reside in a
highly diversiﬁed network. Finally, ﬁrms also beneﬁt from potential referrals as their
partners may identify partnering opportunities for them. As such, a highly diversiﬁed
network is initially found to be positively associated with ﬁrm performance (Goerzen and
Beamish 2005). However, this positive effect will eventually start to diminish as the
marginal beneﬁt is overwhelmed by the marginal costs of management. To conclude, the
relationship between network diversity and ﬁrm performance is inverse-U shaped
(Goerzen and Bearnish 2005).

2.5 Organizational Learning and Knowledge Acquisition

23

Grant (1996) argues that the beneﬁt of strategic alliances is in resource accessing.
However, most scholars argue that ﬁrms not only access resource, they learn and
internalize resources from partners when possible (Khanna et al. 1998). As alliance
partners need to partially open their) resources to their partners and because of the
frequent communication and interaction among partners, learning from each other
inevitably happens in alliances. This is more so in alliances that are based on
complementary resources than do alliances created for economies of scale (Ireland et al.
2002). Firms are motivated to learn from each other because the internalization process
will greatly reduce its dependency level on its partner.

Organizational learning falls into two major categories in alliance literature,
internal learning (learning to manage alliance activities), and external learning (leaming
knowledge from partners). Kale and Singh (2007) proposed a four-stage learning theory
for alliance management. The ﬁrst step is articulation of alliance management know-how.
Alliance managers are supposed to extemalize their personally held, tacit knowledge
about alliance management by using different techniques such as spoken or written
words, metaphors, analogies, or models. This is an important step in that articulated
knowledge is easy to access and store and this would greatly enhance the learning process
(Nonaka, 1994). It helps ﬁrms to create a record of its alliance history, to facilitate ex
post sense making of actions and decisions, and to better understand causal relationships
about alliance management (Kale and Singh, 2007).

The second step is codiﬁcation of alliance management know-how. Going beyond
simple documentation of existing knowledge, Kale and Singh (2007) deﬁned codiﬁcation

as “creating and using knowledge objects or resources such as alliance guidelines,

24

checklists, or manuals to assist actions or decision making in future alliance situation”.
This process is expected to enhance a ﬁrm’s decision-making and actions in its alliance
management and consequently lead to better alliance results in the ﬁrture.

The third step is sharing of alliance management know-how. It is a process
deﬁned as “exchanging and disseminating individually and organizationally held alliance
management knowledge, which is both tacit and/or codiﬁed, through interpersonal
interactions within the organization”. Face-to-face interactions among individuals are
particular important in this stage as they provide opportunities to share individually held
tacit know-how among members of an organization.

The fourth step is internalization of alliance management know-how. This
involves individual’s effort to facilitate absorption of accumulated organizational level
knowledge (Nonaka 1994). It goes beyond simple dissemination process by emphasizing
the absorption of relevant knowledge by individual receivers (Kale and Singh 2007).
Organization members, especially those managing alliances on a daily basis, need to
understand both the why and the how aspects of alliance management strategies.

Another stream of research in the alliance literature focuses on ﬁrms’ active
learning from their partners. In fact, one of the motives for forming alliances is the
acquisition of new technologies, skills, or capabilities from partners (Mowery et a1.
1996). Alliances offer great platform for this type of learning as it combines the incentive
structure of market and the administrative controls associated with hierarchy to overcome
the incomplete contracting problem when uncertainty regarding the focal technology or
capability is involved (Mowery et al. 1996). Because knowledge and technological

capabilities are the building block of competitive advantage, ﬁrms can simultaneously

25

prevent inertia while promoting environmental adaptation by actively learning from their
alliance partners (D02 1998; D02 1996; Ireland et al. 2002).

When ﬁrms are motivated to learn from their partners, learning race occurs in
which each partner is trying to outleam the other (Hamel 1991). The faster learner may
achieve its learning goal and dissolve the alliance even though the other party has not
accomplished its learning goal yet (Hamel 1991). This is especially true when alliance
partners possess unequal learning abilities (Makhija and Ganesh 1997). As such, an
alliance is simultaneously cooperative and competitive oriented (Hamel 1991). This
creates a greater problem to partners in the same industry as the knowledge learned by a
partner may be used directly against the knowledge originator.

Khanna, Gulati, and Nohria (1998) proposed a model to understand the alliance
dynamics when cooperation and competition co-exists by introducing the concept of
private and common beneﬁts. Private beneﬁts refer to those that a ﬁrm can pick up
unilaterally by learning from its partner and apply them to its own operations in areas that
are not related to the alliance activities. Common beneﬁts, on the other hand, refer to
those that accrue to each alliance partner from the collective application of learning that
both parties go through by being in the alliance and these must be from operations in
areas of the ﬁrm that are related to the alliance (Khanna et al. 1998).

The idea is that the ratio of private to common beneﬁts will inﬂuence how the
alliance evolves as well as ﬁrrns’ incentives to allocate resources in a particular area
(Khanna et al. 1998). Different ratio of private to common beneﬁts will result in different
resource allocation patterns. For alliances with only common beneﬁts, ﬁrms have no

incentive to get ahead of the other in terms of learning, as both parties must advance their

26

learning of each other’s knowledge before any common beneﬁt could be realized. This
will result in a pure cooperation situation and resource allocation decision are made
jointly by partners (Khanna et al. 1998).

On the other extreme when only private beneﬁts exist, ﬁrms are motivated to
engage in pure competitive behaviors (Khanna et al. 1998). Because each ﬁrm is
interested in accessing knowledge from its partner and using it consequently in situations
in which it can reap beneﬁts that accrue only to itself and there is no common beneﬁt to
seek, each partner is inclined to terminate the alliance once it has achieved it learning
goals. As such, each partner will try to achieve the leader position in this leaning race and
competition mode controls the whole dynamic (Khanna et al. 1998).

Most alliances will involve both common and private beneﬁts and the situation
thus becomes more complicated than the two extreme cases discussed above. However,
the general conclusion is that the competition will intensify and cooperation will diminish
as the ration of the private beneﬁt to common beneﬁt increases (Khanna et al. 1998).

Several factors were identiﬁed as either facilitating or hindering organizational
learning from partners. These include cultural distance (Barkema et al. 1996), absorptive
capacity(Cohen and Levinthal 1990), partners’ initial conditions (D02 1996), relational
capital between partners (Kale and Singh 2000), social capital (Y li-Renko et al. 2001),
and age and size of ﬁrm (Stuart 2000). Of these, absorptive capability is a key construct
that warrants some particular attention.

The basic idea of absorptive capability is that ﬁrms need to have a prior related
knowledge base in order to successfully identify, acquire, assimilate, and exploit the

external knowledge to which they are exposed (Cohen and Levinthal 1990). Evidence of

27

absorptive capability largely comes from the cognitive and behavioral sciences. As
individuals use linkage to its existing memory to facilitate leaning of a new object, the
more objects and patterns in the memory the easier for the new information to be mapped
into the existing memory structure (Cohen and Levinthal 1990). As such, learning is a
function of what is already known in the related areas. Learning in a novel area is much
more difﬁcult than learning in a known area.

To summarize, different theoretical lenses have been used in understanding
different aspects of strategic alliances. While each theory is primarily used to answer one
particular question in alliance literature, their linkages should not be underestimated. For
example, resource-based view is linked with organizational learning theory informing us
what is being learned in the alliance process. Network theory, when linked with
organizational learning theory may provide direct leaning and indirect learning
differences and explain how much learning could take place in an alliance. Transaction
cost analysis could also be combined with organizational learning theory in explaining
what governance structure may facilitate the learning process given the approriability
hazards associated with the focal resource, knowledge. As all the discussed theories are
intrinsically linked with one another, scholars sometimes need to use multiple theories in
building a comprehensive model to understand issues surrounding alliances. Table 1
presents a select list of literature using these theoretical lenses in understanding strategic

alliance issues.

28

Table 1 Select List of Literature on Strategic Alliances

 

 

 

 

Study Theories Employed Major Topic

Afuah (2000) Resource-based View The inﬂuence of technological change on
ﬁrms’ relationship with alliance partners

Chung et al. Resource-based View The impact of resource complirnentarily, social

(2000) capital, and status similarities on alliance
formation potential

Eisenhardt and Resource-based View Explanation of strategic alliance formation

Schoonhoven motivations including stage of market,

(1996) innovative strategy, size and previous industry

exposure

 

Harrison, et al.
(2001)

Resource-based View

Firms form alliances in order to gain access to
strategic resources. The complementarities of
partners resources are of great importance

 

Hitt et al.
(2000)

Resource-based View

Understanding the different partner selection
criteria between ﬁrms from developing
countries and developed countries (ﬁnancial
assets, willingness to share, and technological
capabilities vs. market knowledge and access)

 

Lampel and
Shamsie
(2000)

Resource-based View

Explains the impact of a ﬁrm’s dominant logic
on its manner in alliance forming and
termination decisions

 

Simonin
(1997)

Resource-based View

Proposes a model emphasizing the importance
of converting experiences into know-how in
order to inﬂuence ﬁrm alliance beneﬁts

 

Grant and
Baden-Fuller
(2004)

Knowledge-based
View

Argument about the importance of alliance
formation lies in its ability of knowledge
accessing rather than acquiring knowledge

 

 

 

Inkpen and Knowledge-based Depending on the type of knowledge
Dinur (1998) View considered, different type of knowledge
transfer strategies might lead to different
knowledge transfer outcomes. Knowledge and
strategy need to be aligned
Larn (1997) Knowledge-based How the socially embedded nature of
View knowledge impede collaboration and

 

 

knowledge transfer

 

29

 

Table 1 (Cont’d)

 

 

 

 

 

 

 

 

 

 

 

Study Theories Employed Major Topic

Lyles and Salk Knowledge-based Examination of the organizational, structural,

(2007) View and contextual factors in inﬂuencing a ﬁrm’s
ability to acquire knowledge from foreign
parent in IJVs.

Reid et al. Knowledge-based Provides a detailed stage-based model of

(2001) View issues surrounding the alliance formation

Simonin Knowledge-based Knowledge ambiguity is negatively related to

(1999) View knowledge transfer. Tacitness, speciﬁcity,
complexity, partner protectiveness,
organizational distance, and cultural distance
increase the ambiguity of knowledge.
Experiences is negatively related to ambiguity

Tiwana and Knowledge-based Demonstrates the role of peripheral knowledge

Keil (2007) View in complementing control mechanisms in
technology outsourcing alliances

Wiklund and Knowledge-based Knowledge-based resources positively

Shepherd View inﬂuence ﬁrm performance. This relationship

(2003) is moderated by entrepreneurial orientation

Glaister and Transaction Cost Strategic motivations for international alliance

Buckley Analysis formation. Factors identiﬁed include relative

(1996) partner size, primary geographical location,
and the industry

Gulati (1995) Transactional Cost Shared R&D component and different national

Analysis location increase the use of equity-based

alliances. Experience with the partner,
decreases the likelihood of equity alliances

McGee et Transaction Cost Examines the inﬂuence of different strategic

al.(1995) Analysis focus and manager’s alliance experiences on
ﬁrm performance

Oxley (1997) Transaction Cost Examination of the governance properties of

 

Analysis

 

different alliance types. Focused on the
appropriability hazard in knowledge based
alliances

 

30

 

Table l (Cont’d)

 

 

 

 

 

 

 

 

 

 

 

Study Theories Employed Major Topic
Oxley and Transaction Cost Proposes an alternative governing mechanism
Sampson Analysis to cope with the incomplete contract problems
(2004) in strategic alliances: limiting the scope of
collaboration
Parkhe (1993) Transaction Cost Examines the inter-linkage among the
Analysis perceived “ﬁrture”, opportunistic behavior,
history of collaboration, commitment, and
payoff in strategic alliances
Reuer and Transaction Cost Examines the dimensionality of the contractual
Arino (2007) Analysis complexity construct and the determinants of
ﬁrm’s decision to adopt various contractual
provisions
Santoro and Transaction Cost Governance is directly inﬂuenced by partner,
McGill (2005) Analysis task, and technological uncertainty and by
interactions among asset co-specialization,
partner uncertainty, and task uncertainty
White and Lui Transaction Cost The greater joint task complexity and inter-
(2005) Analysis partner diversity, the greater the cooperation
costs. Both cooperation costs and transaction
costs inﬂuence the level of time and effort
managers spend on managing alliances
Young-Ybarra Transaction Cost Examines the effect of asset speciﬁcity,
and Analysis hostage arrangement, and dependence on
Wiersenma ﬁrms’ strategic ﬂexibility when alliances are
(1999) performing poorly
Ahuja (2000a) Social Network The impact of ﬁrm’s direct ties, indirect ties,
Theory and structural holes in the network on its
innovation output
Baum et al. Social Network The inﬂuence of the initial network set-ups of
(2000) Theory start ups on their early performance and
strategies to enhance the outcome
Gulati (1999) Social Network Determinants for ﬁrms entering into an
Theory alliance include the level of resources of the

 

 

ﬁrm from prior alliances and its alliance
formation capability

 

31

 

Table l (Cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

Study Theories Employed Major Topic
Kraatz (1998) Social Network Explain how network structure inﬂuences
Theory ﬁrm’s ability to adapt to environmental

changes.

Modhavan et Social Network Explains the network position dynamics when

al. (1998) Theory major industry events occur such as structure
reinforcing event and structure loosening event

Powell et a1. Social Network Examines the impact of the number of direct

(1996) Theory ties, indirect ties, and structural holes in a
ﬁrm’s alliance network on its performance

Tsai and Social Network The centrality and shared vision of a business

Ghoshal Theory unit improves its trustworthiness which in turn

(1998) impacts its resource exchange with other units.

Walker et al. Social Network Proposes social capital as a better indicator of

(1997) Theory cooperation over time than structural holes
propositions

Barkema et al. Organizational The inﬂuence of cultural distance on the

(1996) Learning Theory longevity of foreign joint ventures and the
learning effect in mitigating this inﬂuence

Cohen and Organizational Discussion of absorptive capability and how

Levinthal Learning Theory this concept inﬂuences ﬁrm’s allocation of

(1990) resources on innovation

D02 (1996) Organizational The mediating role of learning in the

Learning Theory relationship between the initial condition of the

alliance partners and alliance outcomes

Kale and Singh Organizational Explains the learning process of successﬁrl

(2007) Learning Theory alliance management. The four stages include:
articulation, codiﬁcation, sharing, and
internalization

Khanna et al. Organizational Explains how the tension between cooperation

(1998) Learning Theory and competition affects the dynamics of

 

 

learning alliances. The ratio of common
beneﬁt and private beneﬁt would be the major
determinant of the cooperation level

 

32

 

Table 1 (Cont’d)

 

 

 

 

 

Study Theories Employed Major Topic
Parkhe (1991) Organizational Develops a multilevel typology of inter-ﬁrm
Learning Theory diversity. Examines how learning and
- adaptation moderate the relationship between
diversity and longevity/ diversity and
effectiveness
Sampson Organizational Examines the effect of prior alliance
(2005) Learning Theory experience on current alliance performance.
Included are the factor such as uncertainty and
level of experience
Ahuja (2000b) Resource-based The creation of inter-ﬁrm linkages through
View/ Social Network collaboration
Theory
Anand and Resource-based Why and how the learning effect of alliance
Kharma (2000) View/Organizational management experience is important to ﬁrms
Learning Theory

 

Dussage et al.

Resource-based

The outcomes and durations of strategic

 

 

 

 

 

(2000) View/Organizational alliances among competing ﬁrms, using
Learning Theory alliances outcomes as indicators of leaming by
partner ﬁrms
Dyer and Social Network The creation and implementation of learning
Nobeoka Theory/Knowledge- networks. Providing guiding coordination
(2000) based View principles that facilitate knowledge transfer
among partners
Goerzen and Social Network Competing hypotheses were set up regarding
Beamish Theory/Transaction the alliance network diversity based on
(2005) Cost Analysis network theory and transaction cost analysis.
Results support network hypothesis
Goerzen Social Network Competing hypotheses were set up regarding
(2007) Theory/Transaction the repeated partnerships based on network
Cost Analysis theory and transaction cost analysis. Results
support network hypothesis
Hamel (1991) Resource-based Extent and means through which alliances

 

View/Organizational
Learning Theory

 

may lead to reappointment of skills of
members. Unequal learning capabilities may
change the stability and longevity of alliances

 

33

 

Table 1 (Cont’d)

 

 

 

Study Theories Employed Major Topic
Hemphill and Transaction Cost Applied multiple theories to recount business
Vonortas Analysis/Resource- motives to engage in R&D alliances and point
(2003) based , out that the differences among these
View/Knowledge- theoretical arguments may not as sharp as they
based seem to be
View/Organizational
Learning Theory
Kale and Singh Knowledge-based Factors that enhance learning from partners
(2000) View/Organizational and protecting core capabilities from partners
Learning Theory (relational capital and integrative conﬂict

handling)

 

Lorenzoni and

Transaction Cost

Articulates the importance of ﬁrm’s capability

 

 

 

 

Lipparini Analysis /Resource- in interacting and coordinating knowledge
(1999) based View/ sharing with other ﬁrm
Knowledge-based
View/Organizational
Learning Theory
Mowery et al. Knowledge-based Examines how equity based alliances and
(1996) View/Organizational absorptive capability facilitate knowledge
Learning Theory transfer among partners
Stuart (2000) Network Examines how partners’ resources and
Theory/ Organizational performance inﬂuence the performance of the
Learning Theory focal ﬁrm. The moderating effect of ﬁrm size
and age is noted
Yli-Renko et Organizational Social capital facilitates the acquisition of
al. (2001) Learning knowledge from external sources
Theory/Social
Network Theory

 

 

 

34

 

CHAPTER 3
ESSAY ONE: ALLIANCE PROTFOLIO CONFIGURATION, KNOWLEDGE
DEVELOPMENT, AND INNOVATION: THEORETICAL FRAMEWORK AND
EMPIRICAL EVIDENCE
3.1 Development of the Conceptual Framework

Even though dyadic relationship studies have dominated the literature of
alliances, scholars have started to realize the importance of managing alliance portfolio as
a whole (Hofﬁnann 2007; Powell et al. 1996; Reuer and Arino 2007; Reuer and
Ragozzino 2006). Both scholars and practitioners have realized that competitive
advantage of a ﬁrm seldom relies on one single successﬁrl alliance but on its ability to
successfully conﬁgure its alliance portfolio (Bamford and Ernst 2002; Hofﬁnann 2007).
As such, the conﬁguration of the whole alliance portfolio becomes an important strategic
issue for ﬁrms with numerous alliances (Hoffrnann 2007).

Hoffrnann (2007) identiﬁed both strategic and environmental determinants of
conﬁguration and evolution of strategic alliance portfolio. Strategic alliance portfolio
conﬁguration is thus deﬁned as the structural and relational characteristics of the alliance
portfolio (Hoffrnann 2007). Drawing on knowledge based view and transaction cost
analysis, this essay explores how characteristics of the R&D alliance portfolio
conﬁguration inﬂuence a ﬁrm’s ability to develop its knowledge reservoir which in turn
impacts its ability to innovate.

We deﬁne R&D alliance portfolio characteristics as ﬁrms’ preference to include
any particular type of agreement or partner in their alliance portfolio conﬁguration

process. In order to compile a comprehensive array of the characteristics, a systematic

review of the alliance literature is conducted to identify managerially relevant partner and

35

agreement types. This process resulted in the following ﬁve portfolio characteristics:
partner dispersion (Goerzen and Bearnish 2005; Hoffrnann 2007; Koka and Prescott
2002; Parkhe 1991), vertical vs. horizontal alliances (Baum et al. 2000; Dussauge et al.
2000; Rindﬂeisch and Moorman 2001; Sivadas and Dwyer 2000); multiple partner
agreements (Feinberg and Gupta 2004; Lazzarini 2007; Zeng and Chen 2003), alliance
scope (Khanna 1998; Khanna et al. 1998; Oxley and Sampson 2004), and partner
repeatedness (Goerzen 2007). These ﬁve characteristics were then presented to a panel
of pharmaceutical practitioners at a strategic marketing conference and their managerial
relevance in alliance portfolio management is conﬁrmed.

Knowledge is particularly important for technology-based ﬁrms: generating and
exploiting knowledge in high technology sectors demand knowledge to be continuously
replenished (Lane and Lubatkin 1998; Yli-Renko et al. 2001). One of the most widely
cited motivations for forming R&D alliances is the acquisition of new technical skills and
technological capabilities from partner ﬁrms (Hamel 1991; Khanna 1998; Mowery et al.
1996). R&D alliances have advantages over conventional contracts and markets for this
task because ﬁrm speciﬁc technological capabilities frequently are based on tacit
knowledge and are subject to considerable uncertainty concerning their characteristics
and performance (Mowery et al. 1996). Alliances with its trait of both market transaction
and hierarchical arrangement, provide a superior means to gain access and assirrrilate the
knowledge and other complex capabilities.

Alliance portfolio management is a very important tool for ﬁrms to gauge its
existing knowledge inventory. Alliance portfolio conﬁguration is a deliberate decision

process that ﬁrms have to go through. With multiple potential partners to choose from,

36

ﬁrms need to make decisions such as which partner to work with and what form of
alliances will be employed. As a basic assumption of resources based view and
knowledge-based View of ﬁrm, resources (knowledge) heterogeneity is inherent across
ﬁrms. As such, the different choices of ﬁrms and forms of collaboration in a ﬁrm’s R&D
alliances portfolio conﬁguration will affect its exposure to external knowledge and its
opportunities for the transfer of that knowledge (Wuyts et al. 2004).

The role of inter-organizational partnerships as conduits of information,
knowledge, and learning is well accepted in the literature (Goerzen and Bearnish 2005).
Studies have suggested that key characteristics of alliance portfolio would have an impact
on a ﬁrm’s ability and opportunity to acquire knowledge from their alliance partners
(Goerzen and Bearnish 2005). The resultant knowledge development from alliance
activities is a function of both the knowledge and skill set of the partners, which
represents the opportunity for learning, and partners’ willingness to share, which
represents the actual potential for learning.

One major assumption of knowledge-based view is that ﬁrms are heterogeneous
in terms of knowledge and competencies they possess. As such, different types of
partners will bring different kinds of skills and technologies to the revelation of the focal
ﬁrm. As such, managers’ preferences in their choice of partners will determine how much
novel knowledge and skills they could be potentially exposed to. This would impact the
knowledge development of the focal ﬁrm. For example, managers’ preference for
working with new partners will ensure a constant inﬂow of fresh and novel information
into the organization. This would be advantageous compared with ﬁrms who only ally

with a limited number of partners or ally with the same partner repeatedly.

37

Transaction cost analysis maintains that knowledge as the focal resources in R&D
alliances poses considerable difﬁculties in contracting since knowledge use can not be
effectively speciﬁed and monitored once it is transferred over to partners (Oxley 1997).
As such, ﬁrms are subject to potential opportunistic behaviors from partners in
knowledge sharing. Firms are thus motivated to withhold information, especially the
critical and sensitive ones from their partners in order to avoid the risk. Therefore,
alliance governance mechanism becomes an essential component in alliance management
as it bears a direct impact on ﬁrms’ willingness to share information with each other.

Alliance governance mechanism studies drew from the relational exchange theory
and identiﬁed the trust-based approach of alliance management. The essence of trust is to
alleviate opportunism (Gulati 1995). As such, trust suggests that alliance partners can
depend on each other’s actions without worrying about opportunistic behaviors (Ireland
et al. 2002). Therefore, ﬁrms are more willing to exchange information and knowledge
with each other when trust exists between partners.

A ﬁrm’s alliance portfolio conﬁguration that fosters trust development will thus
have a positive impact on its knowledge development as partners are more willing to
share information with the focal ﬁrm. Trust nurturing alliance portfolio could be a result
of deliberate partner selection in the portfolio conﬁguration process. Relational exchange
theory suggests that trust evolves from past experiences and as such the history of the
cooperation between partners is negatively related to the perceived opportunistic behavior
of their partners (Parkhe 1993).

As such, ﬁrms’ particular preferences for partner types or agreement types in their

R&D alliance portfolio conﬁguration will have an impact on the total amount of

38

knowledge they are exposed to as well as the amount of knowledge partners are willing
to share with each other. This will consequently either enhance or hinder the knowledge
development of ﬁrms.

We distinguish two dimensions of knowledge development: knowledge breadth
and knowledge depth. Knowledge breadth refers to the range of ﬁelds over which a ﬁrm
has knowledge. Knowledge depth refers to the amount of the knowledge possessed by the
ﬁrm in any particular area (Prabhu et al. 2005). It is important to distinguish these two
dimensions of knowledge development, as we believe that the development of knowledge
breadth and knowledge depth may have both different antecedents and consequences.
While knowledge breadth development is primarily a function of the array of knowledge
sectors a ﬁrm is exposed to, knowledge depth might be more a result of how much
sensitive and in depth information partners are willing to share with each other.

Past studies in understanding the inﬂuence of alliance characteristics on ﬁrm
performance have largely ignored the different development requirements for knowledge
breadth and knowledge depth. Take partner repeatedness as an example, past studies on
its effect on ﬁrm performance have taken on both transaction cost analysis and social
network theory which suggested quite opposite effects (Goerzen 2007). The ﬁndings thus
become very much data oriented instead of theoretically driven.

However, we believe that to understand the impact of R&D alliance portfolio
characteristics’ inﬂuence on knowledge development, the two dimensions of knowledge
development needs to be clearly delineated as the same characteristic may have different
implications on the two knowledge dimensions. An illustration of this point could be

demonstrated by a closer examination of the partner repeatedness. When a ﬁrm has a

39

propensity to ally with the same partner repeatedly, the scope of knowledge the focal ﬁrm
is exposed to is rather limited compared with ﬁrms who always ally with different
partners. This would have a negative impact on its knowledge breadth development.
However, this repeated interaction with the same partner allows for the development of
trust between partners, which allows ﬁrms to share more in depth and sensitive
information with each other. As a result, the focal ﬁrm’s knowledge depth development
will be enhanced.

As ﬁrms develop their knowledge inventory through learning from partners, they
tend to add new products to their product portfolio using the newly acquired knowledge
or by combining it with their existing knowledge base (Rothaerrnel et al. 2006). The
newly acquired knowledge will allow ﬁrms to make improvements in their existing
products or create totally new products that will serve the customer needs better. This
process of knowledge creation also offer increased likelihood of success of related
products in the portfolio (Rothaerrnel et al. 2006). Due to the different characteristics of
radical and incremental innovations, we would expect knowledge breadth and knowledge
depth play different roles on radical and incremental innovations.

The conceptual model is presented in Figure l.

40

 

 

oxahutmmxm Qwum
out. 235.8%
out. «Em

”.8353; NoLEoD

 

 

 

 

 

 

Essmem

 

 

2:830
.3235.”—

 

cowgoqﬁ _S:oEo.~oE c

83865: 136mm 0

 

 

2.83:0 toga—£833....—

 

EQDQ «swap—Bog o

53on “ammo—Bog. o

 

 

 

 

2.83:6 wagon—5.332

3.82 353280 2:. 3 «EMS

 

mcggm 3333”
3on 85:2
35:? 85.89352
somﬂomma 558m
8522 BUSES:

8:32 see;

 

 

533:3.an
ozoutem 8.5:? Gun“

41

3.2 Hypotheses Development
3.2.1 Vertical and Horizontal R&D Alliances and Knowledge Development

R&D alliance partners could be classiﬁed into two major categories: ﬁrms that are
direct competitors and those that are channel members along the value chain. We deﬁne
horizontal R&D alliances as formalized new product development collaborative
arrangements among two or more partners from the same industry. We deﬁne vertical
R&D alliances as formalized new product development collaborative arrangements
among two or more partners operating at adjacent stages of the value chain (Rindﬂeisch
and Moorman 2001). At times, ﬁrms need to enter into alliances with both types of
partners in order to pursue their strategic goals. Even though it might be unexpected that
ﬁrms that are direct competitors will collaborate with each other, such horizontal
alliances do exist in practice (Dussauge et al. 2000; Rindﬂeisch and Moorman 2001).

The effect of knowledge sharing in horizontal R&D alliances stems from one
major assumption: level of redundancy of knowledge stores among partners. Redundancy
refers the degree of overlap in the knowledge base between two or more social actors or
organizations. Horizontal R&D alliance partners (i.e. partners that are operating in the
same industry) share very similar knowledge inventory as they use similar technologies
to serve similar customer needs (Dussauge et al. 2000; Rindﬂeisch and Moorman 2001).

As Rindﬂeisch and Moorman (2001) point out, horizontal alliance partners,
compared with vertical alliance partners, have a high redundancy in terms of the
capabilities and skills each partner contributes to the joint effort. They have access to
similar information because of common structural linkages through trade associations

(V ives 1990), industry based norms and procedures (von Hippel 1987), and membership

42

of common technological community (Powell et al. 1996). It should be acknowledged
that ﬁrms in the same industry may also possess speciﬁc, idiosyncratic skills and
capabilities (Dussauge et al. 2000). However, these idiosyncratic resources may be the
cornerstone for a ﬁrm’s competitive advantage on the market place. Firms may feel
reluctant to share these critical resources with their direct competitors because of the
zero-sum opportunity, increased opportunistic behavior potential, long term partner
motives, and the vulnerability of the proprietary know-how (Sivadas and Dwyer
2000).As such, a ﬁrm that has an emphasis on forming R&D alliances with direct
competitors has a limited exposure to new and novel knowledge which in turn limits its
knowledge breadth.

Vertical R&D alliances, on the other hand, are characterized by a low redundancy
of knowledge. Overlapping knowledge is the product of social actors occupying similar
structural positions in which they are exposed to similar types of information, input and
dominant logic (Bettis and Prahalad 1995; Rindﬂeisch and Moorman 2001). Vertical
R&D alliance partners, which are deﬁned as partners from different stages along the
value chain, normally occupy very different social positions compared with horizontal
alliance partners and thus result in very different knowledge composition. Partners bring
highly complementary and fresh knowledge to the collaboration. As such, a ﬁrrn’s
propensity to engage in vertical R&D alliances will have an advantage in developing its
knowledge breadth.

Hypothesis 1 .' The higher the propensity of a firm entering into horizontal over

vertical R&D alliances, the lower the knowledge breadth of the ﬁrm.

43

In order to develop knowledge depth, information freshness and novelty becomes
a lesser issue compared with the relatedness and volume of information. Information
richness, intensity of knowledge sharing, and knowledge absorption along a narrow
domain become critical. The notion of absorptive capability thus plays an important role
in a ﬁrm’s knowledge depth development. The basic idea of absorptive capability is that
ﬁrms need to have a prior related knowledge base in order to successfully identify,
acquire, assimilate, and exploit external knowledge they are seeking (Cohen and
Levinthal 1990). One key attribute of horizontal R&D alliances is the common
knowledge base partners possess. It is argued that because of the similarity of the
knowledge stores among partners in a horizontal alliance arrangement, this similarity
enhances the absorptive capability of the partners to better assimilate the knowledge
along a rather narrow scope of skills, which will positively inﬂuence the knowledge
depth of the parties.

However, ﬁrms in vertical R&D alliances normally contribute very
complementary knowledge to each other. Because of the knowledge complementary, not
knowledge similarity, absorptive capability will be limited among partners. As such,
ﬁrms that have a propensity to enter into vertical partnerships will be limited in its
knowledge depth development.

Hypothesis 2: The higher the propensity of a ﬁrm entering into horizontal over

vertical R&D alliances, the higher the knowledge depth of the ﬁrm.

3.2.2 Partner Dispersion and Knowledge Development
Partner dispersion refers to the spread of distinctive alliance partners in the

portfolio. For any given portfolio size, it essentially measures the number of distinct

44

partners. It is therefore an indicator of a ﬁrm’s propensity to bring in new/unfamiliar
partners into its R&D alliance network. A highly dispersed R&D alliance portfolio links
the focal ﬁrm to a diversiﬁed set of partners from different strategic groups and industries
(Hoffrnann 2007). Thus it offers the focal ﬁrm access to a diversiﬁed pool of skills and
knowledge.

A basic assumption of knowledge-based view is that ﬁrms are heterogeneous in
terms of the knowledge they possess to compete in the market (Barney 1991; Barney
2001; Grant 1996). Knowledge heterogeneity refers to the variety of knowledge, know-
how, and expertise to which a ﬁrm has access through its alliance activities (Rodan and
Galunic 2004). In fact, this heterogeneity is supposed to be sustainable so that ﬁrms with
superior resources and knowledge can achieve long run superior performance (Barney
1991; Wemerfelt 1984). A diversiﬁed partner portfolio of R&D alliances thus indicates a
diverse inﬂow of knowledge into the focal ﬁrm. This diversity of inﬂow of information is
thus a direct antecedent of knowledge breadth of the participating ﬁrms in the alliances.
As such, we would expect a ﬁrm’s R&D alliance portfolio dispersion to be positively
associated with its knowledge breadth development.

As for the portfolio dispersion’s inﬂuence on ﬁrms’ knowledge depth
development, different views exist in the literature. One thought maintains that
knowledge depth is essentially about the richness of information and knowledge. As
such, it focuses on ﬁrms’ overall alliance experience and its history with the current
partners (Koka and Prescott 2002). Prior alliance experiences with partners help to
establish relational norms which include collaboration, continuity expectation, and non-

coercive communication patterns (White and Lui 2005). All these bring the beneﬁt of

45

effective routines of accessing knowledge from each other (Koka and Prescott 2002).
However, a ﬁrm’s tendency to work with new, unfamiliar partners brings about the
concern of behavioral uncertainty of partners. When the uncertainty of partners’
behaviors, especially in regards to their future use of the shared knowledge and skills,
partners may feel reluctant to share any sensitive, in depth information with each other.
Alliance portfolio dispersion measures a ﬁrm’s propensity to work with
new/unfamiliar partners. A highly dispersed R&D alliance portfolio indicates that the
ﬁrm prefers to work with new partners. In this case, the relational capital does not exist
and this would hinder the assimilation of knowledge in depth. As such, portfolio
dispersion would be negatively related to the knowledge depth development of a ﬁrm.
On the contrary, another stream of research may suggest that alliance portfolio
dispersion actually enhances the knowledge depth of a ﬁrm. The diverse inﬂow of
knowledge is proven to enhance the assimilative power, enable novel associations (Cohen
and Levinthal 1990; Wuyts et al. 2004), and stimulate broader perspectives and synthesis
(Dewar and Dutton 1986). It is demonstrated that exposure to heterogeneous knowledge
should improve both the creative potential of the focal ﬁrm and its ability to implement
its ideas and to execute complex tasks in general (Rodan and Galunic 2004). It is often
times that the newly assimilated information, which seems to be unrelated to the existing
knowledge inventory, may turn out to be intriguing in new interpretation of existing
knowledge in the ﬁrture (Pelled et al. 1999). This is in nature a process of reintegration of
concepts or blending of ideas which is a fundamental cognitive process in the generation

of novel insights into the existing knowledge of a ﬁrm.

46

Our position regarding the effect of alliance portfolio dispersion is that in order
for the dispersion to have a positive effect on the knowledge depth, successﬁrl knowledge
assimilation is the necessary antecedent for this positive effect to take place. However,
without the incentive to share, the new information will not be assimilated. As such, the
knowledge that has the potential to offer new perspectives on the existing knowledge
base will remain outside the ﬁrm boundary and its effect can not be realized. Thus, we
believe that the dispersion property of an R&D alliance portfolio will have a negative
effect on a ﬁrm’s knowledge depth development.

Hypotheses 3: The higher a ﬁrm 's R&D alliances portfolio partner dispersion,

the higher its knowledge breadth.

Hypotheses 4: The higher a ﬁrm ’3 R&D alliances portfolio partner dispersion,

the lower its knowledge depth.

3.2.3 Partner Repeatedness and Knowledge Development

We deﬁne partner repeatedness as a ﬁrrn’s preference to ally with a particular
partner repeatedly over time. An example would be helpful to clarify its relationship with
the partner dispersion construct. Let’s assume ﬁrm X and ﬁrm Y both have nine R&D
alliances in their R&D alliance portfolio and the distributions of partners are as follows:

Firm X (A, A, A, B, B, B, C, C, C)

Firm Y (A, A, A, A, A, A, A, B, C)
Since both ﬁrms have three distinct partners out of nine alliances, their partner dispersion
level is equivalent. However, a closer look at the distributions of the three partners in

these two ﬁrms’ R&D alliance portfolio indicates that ﬁrm Y shows a much stronger

47

preference for one of its partners, A than X does. As such, the partner repeatedness of Y’s
portfolio is higher than that of X.

From a transaction cost perspective, repeated partnering offers partners
opportunities to better understand each other’s goals, long run motivation, and
capabilities. Repeatedness indicates intense and frequent interactions between two social
organizations. The frequent interactions between partners can thus nurture trust-based
governance mechanism, which would greatly reduce the monitoring costs (Goerzen
2007). The intensiﬁed social interaction is proven to be beneﬁcial in knowledge
acquisition in alliances (Ring and Van de Ven 1994; Yli-Renko et al. 2001) as the
increased trust level gives ﬁrm incentives to share knowledge and information with each
other (Dyer and Singh 1998; Zahra et al. 2000). The intense and frequent interaction
between repeated partners also enhances ﬁrm’s ability to identify pertinent knowledge
that are valuable to the ﬁrrn’s innovation effort (Cohen and Levinthal 1990; Lane and
Lubatkin 1998). Inter-ﬁrm knowledge sharing routines, which develop through the
speciﬁc alliance experiences, will lay a foundation for the absorptive capability that
enables alliance partners to recognize valuable knowledge and effectively transfer it
across ﬁrm boundaries.

Ring and Van de Ven (1994) note that ﬁrms enjoy greater intensity and frequency
of knowledge sharing with partners of high social interaction. The more intense social
interaction between partners, the more conﬁdent ﬁrms are with their partner’s intentions
and capabilities (Y li-Renko et al. 2001). A general beneﬁt of repeated partnering is that
the focal ﬁrm comes to know its partners better and thus would be able to cooperate with

each other better (Wuyts et al. 2004). Kogut and Zander (1996) proves that close and

48

intense social interaction help ﬁrms to overcome the tastiness of technological know-how
and thus greatly enhance knowledge transfer. Thus the repeated social interaction among
partners is beneﬁcial for assimilation of tacit, deeper knowledge from partners which
would increase a ﬁrm’s knowledge depth.

Secondly, through repeated partnering, dyadic partners may be induced to make
transaction speciﬁc investment that reduces transaction costs and thus increase the value
created (Dyer and Singh 1998). Yli-Renko, Autio,and Sapienze (2001) actually described
the whole process as self-enforcing. Intense social interactions give ﬁrms conﬁdence in
each other’s capability and behavioral certainty. This would intensify the ﬁequency and
depth of information exchange. A direct result of this process is the development of
relationship-speciﬁc common knowledge. This increased common knowledge would lead
to better absorptive capacity of the partners as the internalization of novel knowledge
depends on the existence of the prior related knowledge within the ﬁrm (Cohen and
Levinthal 1990). As both parties’ ability to absorb the other party’s knowledge increases,
they are more motivated to invest in knowledge sharing routines which would greatly
enhance knowledge learning for both parties.

Knowledge depth is also closely related to knowledge complexity. Repeated
partnering generates a speciﬁc advantage in transferring complex and tacit knowledge
across ﬁrm boundaries (Zucker et al. 2002). Frequent and repeated interaction facilitates
the transfer of the tacit knowledge (Hansen 1999) and generates a deeper understanding
of the newly identiﬁed and existing knowledge (Dewar and Dutton 1986) as repeated
interactions allow for the emergence of relationship speciﬁc heuristics (Uzzi 1997) and

shared mental models (Madhavan and Grover 1998). These heuristics and mental models

49

in turn facilitate the process of assirrrilating complex knowledge (Polanyi 1966; Wuyts et
al. 2004).

However, Hoang and Rothaermel (2005) argue that a portion of knowledge and
skills that accumulate based on repeated partnering over time may be partner speciﬁc. As
such, the positive effect of alliance experience on alliance performance is partner
speciﬁc. The notion of absorptive capability also lends support to the idea that the ability
of ﬁrms to learn form each other in an alliance depends on the similarity between the two
ﬁrm’s knowledge bases, organizational structure and dominant logic (Lane and Lubatkin
1998). It is argued in the literature that the more allying experiences two ﬁrms have, the
more likely they will develop a technological bases that are overlapping with each other
(Dussauge et al. 2000; Nakamura et al. 1996). Even though repeated partnering help ﬁrm
to learn ﬁ'om a particular partner, the knowledge range it is exposed to though, is limited
to what that particular partner possess. As such, knowledge breadth would be lirrrited.

As such, we hypothesize:

Hypothesis 5 .' The higher the repeated partnering in a ﬁrm ’s R&D alliance

portfolio, the lower its knowledge breadth.

Hypothesis 6: The higher the repeated partnering in a ﬁrm ’s R&D alliance

portfolio, the higher its knowledge depth.

3.2.4 Multiple Partner R&D Alliances and Knowledge Development

We deﬁne multiple partner R&D alliances as R&D alliances that involve more
than two partners. A ﬁrm’s R&D alliances portfolio will also differ in terms of how often
ﬁrms enter into alliances that involve multiple partners in each R&D alliance project.

Even though it is common that a ﬁrm will ally with only one other ﬁrm to achieve new

50

product development objectives, multiple partner alliances do occur in practices (Zeng
and Chen 2003).

The inﬂuence of a ﬁrm’s propensity to enter into multiple partner alliances on its
knowledge breadth development is evident. For any given size of a portfolio, ﬁrms with a
higher propensity to enter into multiple partner agreements are in effect allied with a lot
more partners than ﬁrms who prefer two-party alliances. This increased number of
partners offer better access to a broader range of knowledge and skills. A multi-partner
alliance, compared with a two-party alliance, allows ﬁrms to get access to more
diversiﬁed knowledge bases. As such, knowledge breadth of participating ﬁrm will be
greatly enhanced.

Even though a ﬁrm’s propensity to engage in multi-partner alliances have an
obvious positive effect on its knowledge breadth development, its effect a ﬁrrn’s
knowledge depth might be negative. Due to the unique characteristics of knowledge as a
resource, opportunistic behavior would be a major concern for ﬁrms to share knowledge
with others as they cannot effectively control the use of it by their partners once it is
transferred (Oxley 1997).

First of all, the increased number of partners in alliances by itself increases the
possibility that some one will act opportunistically. Due to the increased number of
players in the joint activity, ﬁrms’ screening ability of every partner’s credibility and
dependability will be imperfect. This concern would limit ﬁrm’s willingness to share any
sensitive, in depth information with the whole group. In addition, in a two-party alliance,
each player can effectively punish the other for opportunistic behaviors by choosing non-

cooperation in subsequent interactions. Thus, each player can attempt to shape the other’s

51

behavior and hence indirectly inﬂuence the. outcome of the alliance. This is however not
the case when multiple partners are brought together. In a project that involves multiple
partners, it is much harder to for any single player to shape the group dynamics
effectively (Zeng and Chen 2003). This causes a lack of controllability of other’s
behavior and the alliance outcome is more likely to increase the defective behavior.
When ﬁrms are not sure of the behaviors of the other ﬁrms which include but is not
limited to the application of the shared knowledge, ﬁrms may be reluctant to share any
sensitive, critical, and in depth knowledge with each other, especially those that are
critical for competitive purposes. As such, we hypothesize:

Hypothesis 7: The higher the propensity of a ﬁrm to enter into multiple partner

alliances, the higher its knowledge breadth.

Hypothesis 8: The higher the propensity of a ﬁrm to enter into multiple partner

alliances, the lower its knowledge depth.
3.2.5 Alliance Scope and Knowledge Development

We deﬁne R&D alliance scope as the number of technologies included in an
alliance agreement. When a ﬁrm wishes to undertake an R&D project for which it does
not currently possess all of the relevant technical knowledge, forming an alliance is
normally a solution. However, before an alliance is established, managers have to make
many critical decisions. One of these decisions is to deﬁne an appropriate scope of
activities included in the alliance agreement (Oxley and Sampson 2004).

Researchers have distinguished between vertical scope and horizontal scope of
alliances (Reuer et a1. 2002). Vertical scope refers to the extent partners combine multiple

and sequential functions of value chain activities within the alliance such as R&D,

52

manufacturing, and marketing. Horizontal cope of alliance refers to the size, complexity,
and technology variety that is covered in an alliance. Previous studies on alliance scope
have normally focused on the vertical scope of alliances due to the availability of
secondary data. However, as our focus is on the scope effect on knowledge accumulation,
we choose to focus on the horizontal scope of alliances as this normally refers to the
technology diversity involved, which has a direct impact on the knowledge a ﬁrm is
exposed to by his partners.

First of all, horizontal scope of alliances is a direct indication of the diversity of
knowledge that ﬁrms need to share with each other. For any given portfolio size, a ﬁrm’s
propensity to engage in multi-technology alliances will surely increase its exposure to a
broader span of knowledge as ﬁrms need to share their expertise in multiple areas to
achieve the common beneﬁt. This, as argued before, will increase the knowledge breadth
of participating ﬁrms.

On the other hand, when multiple technologies are included within one alliance
agreement, these technologies are normally related with each other. Because of the
contracting hazards associated with knowledge-based collaborations, transaction cost is a
major concerns for such endeavors (Oxley 1997). When multiple technologies are
involved in one agreement, the transaction cost in terms of monitoring partner behavior
becomes a major challenge. Unless there are synergistic values associated with the
multiple technologies speciﬁed in the agreement, it would be irrational to put two or
more unrelated technologies into one package for development purposes. As such, the
depth of knowledge would also increase as a result of the relatedness of the multiple

technologies.

53

Hypothesis 9: The higher the average scopes of R&D alliances in a ﬁrm ’s
alliance portfolio, the higher its knowledge breadth.
Hypothesis 10: The higher the average scopes of R&D alliances in a ﬁrm '3
alliance portfolio, the higher. its knowledge depth.

3.2.6 Knowledge Development and Innovation Output

The positive impact of a ﬁrm’s knowledge development on its innovation
outcome has been documented in the literature (Prabhu et al. 2005; Sorescu et al. 2003).
It is asserted that a fum’s ability to generate knowledge is a vital antecedents to its ability
to innovate (Prabhu et al. 2005). Newly developed knowledge allows ﬁrm to see existing
knowledge stores from new perspectives and the integration of new and old knowledge is
a major source of creativity of serving customer needs.

Prabhu, Chandy, and Ellis (2005) argues that both knowledge depth and
knowledge breadth contribute to a fum’s innovation success. The development of
knowledge depth allows ﬁrms to better identify and assimilate external knowledge and
turn them into innovative products and services (Cohen and Levinthal 1990). Firm with
low knowledge depth may suffer from technological lock out problem and fall short in
accurately assessing innovation potentials (Cohen and Levinthal 1989). F irrns with high
knowledge depth are better at evaluating knowledge and using it for innovation purposes
(Prabhu et al. 2005).

From a knowledge-based perspective, the breadth of knowledge is a key input
factor for innovation (Cohen and Levinthal 1990; Henderson 1994; Henderson and
Cockbum 1994). Knowledge breadth indicates the diversity of knowledge a ﬁrm

possesses. A wide array of knowledge sectors allow ﬁrms to integrate them in creative

54

ways to improve customer beneﬁts (Henderson and Cockbum 1994; Volberda 1996).
Broader knowledge base of a ﬁrm also offers greater ﬂexibility and adaptability to
turbulent environment as the availability of diverse knowledge safeguards the value of
the ﬁrm when technologies change rapidly (Prabhu et al. 2005). The bulk literature
strongly supports that idea that the broader a ﬁrm's knowledge, the greater its ability to
create innovations (Prabhu et al. 2005).

While the impact of knowledge depth and knowledge breadth on ﬁrm innovation
is well researched, we would like to make a contribution to the literature by theorizing the
differential impact of knowledge depth and knowledge breadth on radical and
incremental innovation. Due to the different natures of radical innovation and incremental
innovation, development of these two types of innovations may have very different
knowledge requirements (Sivadas and Dwyer 2000).

A further argument could be made in terms of the different impacts of knowledge
breadth and knowledge depth on radical innovation and incremental innovation
respectively. By deﬁnition, radical innovations are those that are built on new
technologies and often rely on the integration of a diversiﬁed pool of technologies
(Wuyts et al. 2004). As such, ﬁrms with a diversiﬁed pool of knowledge enjoy greater
advantage for the experimentation of new ideas and concepts. Because of the integrative
capability, which refers to ﬁrms’ ability to combine knowledge from different areas for
innovation purposes, the potential for integration increases exponentially with any
addition of new piece of information. As such, we would expect that the novelty of the

incoming information and the overall diversity of the knowledge inventory of a ﬁrm will

55

have a much greater impact on its radical innovation than the mere quantity of its
knowledge inventory.

As for incremental innovation which is more of modiﬁcation of existing products,
the quantity and relevance of incoming information may be more important than its
novelty (Wuyts et al. 2004). Incremental innovation is mostly based on an increased
understanding of the existing product. Sivadas and Dwyer (2000) also note that the
development of incremental innovation is strongly based on existing technologies. Firms
may arrive at incremental innovations without accessing novel knowledge or integrating
different technologies. What is important for incremental innovation from a learning
perspective is that the incoming information should be clearly associated with what is
already known. This could be translated into the idea that ﬁrms need to build a richer
understanding within a given knowledge domain to come up with incremental innovation.

Given the differences in the nature of the two kinds of innovation and their
different knowledge requirements, we hypothesize:

Hypothesis 1 I .' A ﬁrm ’s knowledge breadth has a greater impact on its radical
innovation than knowledge depth.

Hypothesis 12: A ﬁrm ’s knowledge depth has a greater impact on its incremental

innovation than knowledge breadth.

3.2.7 Innovation and Firm Financial Performance

It is well proven in the literature that innovation and new product development is
the major source for a ﬁrm’s competitive advantage in the market place and its ﬁnancial
performance (Geroski et al. 1993; Sorescu et al. 2003). Due to the increased customer

beneﬁt associated with a new product, ﬁrms enjoy higher return on their investment when

56

new product could be successfully commercialized (Chaney et al. 1991). F irrns manage

to build stocks of radical and incremental innovations over time to maintain competitive
in the market place. Higher levels of innovation enhance a ﬁrm's ﬁnancial performance

(Geroski et al. 1993).

Cho and Pucik (2005) have identiﬁed three streams of work that support a direct
relationship between innovation and ﬁrm performance. The ﬁrst is the positive
relationship between organizational innovation and ﬁrm performance. Examples are
given as Damanpour and Evan (1984) and Subrarnarrian and Nilakanta (1996). The
second stream is the positive relationship between product innovativeness and ﬁrm
performance. Examples include the studies by Kleinschmidt and Cooper (1991). The
third stream is on value innovation with an example of Kim and Mauborgne (1997). All
these studies offer empirical support for the positive relationship between innovation and
ﬁrm performance.

Hypothesis 1 1 .' A ﬁrm ’s stock of incremental innovation is positively related to its

ﬁnancial performance.

Hypothesis 12: A ﬁrm ’s stock of radical innovation is positively related to its

ﬁnancial performance.
3.2.8 Control Variables

In addition to the relationships discussed above, we control for other variables
that may affect a ﬁrm’s knowledge development, innovation outcome, and ﬁnancial
performance but are outside our theoretical focus. Speciﬁcally, we control for portfolio

size, R&D expenditure, and ﬁrm size for knowledge and innovation based dependent

57

variables. We will control for the number of radical and incremental innovation
introduced that year for ﬁnancial performance.

Portfolio size. When the dependent variables are knowledge and innovation, portfolio
size (i.e., the total number of R&D agreements in a portfolio) is found to be positively
related to both types of outcomes. Studies have shown that a larger portfolio offers more
exposure to different knowledge bases (Dewar and Dutton 1986) and the portfolio size’s
positive effect on innovation has also been documented (Powell et al. 1996; Shan et al.
1994; Wuyts et al. 2004).

R&D expenditures. We would also control for the R&D expenditure for knowledge and
innovation based outcomes. It is expected that the more resources devoted to research and
development, the greater fum’s knowledge base would be and the better their innovation
volume (Wuyts et al. 2004).

Firm size. We would also control for the size of the ﬁrm for all three types of dependent
variables (knowledge, innovation, and ﬁnancial performance). The effect of ﬁrm size on
knowledge and innovation development largely stems for the seminal work of
Schumpeter (1942). Different effects (positive, negative, and insigniﬁcant) were
documented in the literature (e. g., Ahuja and Lampert 2001; Chandy and Tellis 2000;
Miles et al. 2000). The effect of ﬁrm size on performance is based on the idea that large
have advantage in obtaining lower cost of capital while simultaneously lowering risk
(Chang and Thomas 1989; Goerzen 2007). Previous studies have suggested that the
number of employees, sales, and assets are all appropriate proxies for ﬁrm size (Harrison
et al. 1988; O'Sullivan and Abela 2007). We operationalize ﬁrm size as sales in our

study.

58

Number of innovations introduced. In testing the inﬂuence of ﬁrms’ stock of innovation
on its ﬁnancial performance, our conversation with practitioners suggest that we need to
control for the number of innovations introduced that year. When a new product is
introduced, abnormal expenditures may occur due to the change in the production,
distribution, and inventory processes. This is particularly true for radical innovation
which may require huge investment on the manufacturing facility. As this is expected to
negatively affect the bottom line of the ﬁrm, we would control this in our study.

To summarize, a conceptual model is developed which ﬁrst identiﬁes a ﬁrrn’s
R&D alliance portfolio characteristics and their inﬂuence on a frrm’s ability to develop
knowledge breadth and knowledge depth. Due to the different opportunities these
characteristics offer to the focal ﬁrm in terms of the diversity of knowledge it is exposed
to as well as the different level of contracting difﬁculties associated with each
characteristic, ﬁrrns’ ability to develop knowledge breadth and knowledge depth is either
enhanced or constrained. We further analyzed the differential impact of knowledge
breadth and knowledge depth on incremental and radical innovations. Finally, innovation
is hypothesized to be positively related to ﬁrm performance due to the increased
customer beneﬁts.
3.3 Research Methodology
3.3.1 Empirical Setting

We choose to empirically test our model in the pharmaceutical industry. In
particular, we examine the effect of pharmaceutical ﬁrms’ portfolio of R&D alliances on
their knowledge breadth and knowledge depth development which in turn inﬂuence their

innovation and ﬁnancial performance. There are several reasons we prefer the

59

pharmaceutical industry. First, pharmaceutical industry is a technology intensive industry
in which both knowledge development and irmovation effort play signiﬁcant roles in
establishing and sustaining ﬁrms’ competitive advantage (Wuyts et al. 2004). Second,
inter-ﬁrm c00peration in the pharmaceutical industry has been a popular business
practice since the 19803. Most pharmaceutical ﬁrms have entered into numerous
agreements and these portfolios of R&D agreements enjoy substantial variation in their
composition. As our model focuses on the different characteristics of R&D alliance
portfolio, this context would be ideal for testing our model (Wuyts et al. 2004).

Third, secondary data are available on all inter-ﬁrm agreements between
pharmaceutical ﬁrms in the United States since 1985 through various sources such as
Verispan, Bioscan, Recap, and IMS. These sources trace in great details of alliance
activities in the pharmaceutical industry. Finally, pharmaceutical industry has long been a
preferred context for empirically testing innovation and alliance related topics (e. g.,
Nicholls-Nixon and Woo 2003; Prabhu et al. 2005; Rothaerrnel and Deeds 2004; Sorescu
et al. 2003; Wuyts et al. 2004).

3.3.2 Sample

Our sample consists of 66 public pharmaceutical companies in the United Sates.
We limited our sample within public ﬁrms because we need to obtain ﬁnancial measures
as our dependent variable and control variables. Public ﬁrms are mandated to disclose
this information to their shareholders in their annual reports. We compiled data of R&D
alliance agreements, new patent approval, new drug approval, and ﬁnancial information

from 1991 to 2005. The starting point of 1991 is selected based on the fact that FDA did

60

not provide detailed information about newly approved drugs before then which would
prevent us from effectively classifying them into radical vs. incremental innovations.

The 66 pharmaceutical ﬁrms include most of the major players in the market such
as Pﬁzer, GlaxoSmithKlein, Eli Lilly, and Merck. A total of 3828 R&D alliances were
compiled into our data base for these 66 ﬁrms over the 15 year period. As expected, ﬁrms
demonstrated considerable variation in terms of their R&D alliance portfolio
characteristics such as size, repeatedness, dispersion etc; For example, some ﬁrms have
as many as 370 R&D alliances in their portfolio while some have zero. The average R&D

alliances portfolio size is 31. Table 2 provides a summary of the sample characteristics.

61

 

$.23 we. 3a 3.80.2 o 382:8 .5 93
Vm .N m ~ A oodm o soum>o§ 3258805
moo N3 9:. o 8:285 1863a
85 SN 3% 2: Sean S3329

Ga: 83 21.8 o £32m omBEoQ
moo mm; 8a _ 88m 83:...
w _ .o oi SA 2: 555 2&3:
:.o _ So a no Seemed
33 end oo; o mmoéoﬁomom
ta. oo. oo. _ oo. T $05.55 .m> 328203
a? 3.8 8.89: 8.3%; .8. _- 38:3: as 8am
8 : Em 0 35 £328
:o:a_>on @3655 :82 555qu 83552 maﬁa—SEE

 

85239220 «Es—am .8 FEES—6 N 935.

62

3.3.3 Data Sources

Four separate data sources provide information for our empirical testing: Recap,
US Patent and Trademark Ofﬁce/Delphion, FDA, and Compustat.

For the R&D alliance portfolio related constructs, we collect data from
Recombinant Capital Database. Recombinant Capital is a consulting ﬁrm that specializes
in pharmaceutical and biotechnology alliances based in the San Francisco Bay Area.
They use multiple sources such as trade literature, press releases, and annual reports to
compile a database of all the alliance activities of pharma/biotech companies. It also
provides detailed information about each identiﬁed alliance activity such as parties to the
agreement, nature of the agreement, development stage when alliances are formed, and
the technologies that the agreements cover. This database has been used in the literature
to construct alliance related variables (Wuyts et al. 2004).

For the knowledge-based constructs, we use information ﬁ'om US Patent and
Trademark Ofﬁce and Delphion. In the pharmaceutical industry, patent is an excellent
indicator of ﬁrms' technical knowledge (Prabhu et al. 2005; Rothaennel and Deeds 2004;
Sorescu et al. 2003; Wuyts et al. 2004). Patent ofﬁce also classiﬁes patents into different
classes and subclasses that indicate their ﬁeld of use. Therefore, it is possible to identify
the number of technical ﬁelds in which a ﬁrm has knowledge by studying the
classes/subclasses. Delphion is an excellent database that offers search ﬁrnctions. The
number of approved patents in each patent class and subclass for each ﬁrm in each year is
available from the Delphion database.

Third, we collect data on new drugs from the drug approval list of the Food and

Drug Administration (FDA) as our innovation based outcome constructs. FDA publishes

63

all the approved new drugs weekly since 1991. The publication provides additional useful
information about each drug, such as its therapeutic potential and chemical type. This
would help us to classify radical vs. incremental innovations. First, FDA categorizes all
new drugs according to their treatment potential and distinguishes between “standard
review” (therapeutic qualities are similar to those of an already marketed drug) and
“priority review” (Therapeutic qualities are signiﬁcantly superior to those drugs already
on the market) drugs. FDA also assigns a chemical type to each drug. Those that
incorporate a new technology is classiﬁed as Chenrical Type 1, otherwise it will be type
2, Type 3, etc. Following the literature, we would use both piece of information to
classify radical vs. incremental innovations.

Finally, we collect ﬁnancial performance (net income) as well as control variables
(ﬁrm size and R&D expenses) ﬁom the Compustat database.

Figure 2 is a graphical illustration of the data sources for our study.

64

 

 

H<PmDQECU («Gnu ZC—mhw—HEOHA—WD mﬁahdwn hdwUHM—
owmm ozomtom o

omoom 85:? o

8365me harm a 83555: 358083 c Sth ammo—Bog a 36535 Hogméﬁsz o
oﬂm SHE a 8:305: 323% o 53on 03226qu . wmocuoumomom 553m o

0885 82 o

 

 

 

 

 

 

56.69me ozomtom o

305:2 mam age; .

saga... mam 3822a .

 

 

wagon. Rev—ESQ .3“ 823m 3.5 "N ensur—

65

3.4 Measures
Horizontal alliances over vertical alliances (HOVA)

This construct measures a ﬁrm’s preference of allying with horizontal partners to
vertical partners. We classify an alliance as horizontal if the partner is another
pharmaceutical ﬁrm. We classify it as vertical if the partner is a biotech, university or
device ﬁrm. It is measured as the difference between the total number of horizontal R&D
alliances up to year t for ﬁrm i minus the total number of vertical R&D alliances up to
yeart for ﬁrm i then divided by the total alliances portfolio size for ﬁrm i up to year t.

Formally,

cum V cum
it

c m Hi! -
H 0 VA it“ = A cum
it

 

This construct will equal to 1 if the ﬁrm only allies with horizontal partners in its R&D

alliance portfolio. It will be —1 if it only allies with vertical partners.

Alliance Partner Dispersion (DISP)

This construct measures the distinct number of partners in a ﬁrm’s R&D alliance
portfolio for a given portfolio size. We measure it as the ratio of number of distinct
partners over the total number of R&D alliances.

Formally,

cum

DIS cum I)”
it = Aicum

This will equal to 1 when there are as many partners as there are alliances in the

portfolio, which means ﬁrms always ﬁnd new partners to work with. It will be close to

66

zero when there are many alliances but only a small number of distinct partners in the

portfolio.

Multiple partner alliance pr0pensity (M ULT)

This measures a ﬁrm’s preference to involve more than one partner in an R&D alliance
agreement. We ﬁrst count the total number of partners (including the repeated ones) and
then divide it by the total number of alliances.

Formally,

PT.”‘”' m j m.
MULTHM' = _—“_ Where PT” = 21’“, j

Asa". i=1
This measure will take a value of 1 if there are as many partners as there are alliances
which indicates that in each and every alliance in the portfolio there is only one partner to
the ﬁrm. However, the value will become signiﬁcantly bigger if the total number of
partners (including the repeated ones) is more than the number of alliances, which

indicates the ﬁrm has a preference to include more than one partners in each agreement.

Partner Repeatedness (RPEA)
This measures a ﬁrm’s preference to work with a subset of its partners repeatedly over

time. We operationalize it in the following way.

 

j
n. .
cum _ 2 _ ".1
RPEA, — 2 pM Where pm ..
i=1

Zinnia

The partner repeatedness index equals one when a ﬁrm allies with only a single partner

and it is close to zero when a ﬁrm spreads its R&D alliance activity over many partners.

67

Alliances Scope (SCOP)

This measures a ﬁrm’s preference to include multiple technologies within one alliance
agreement. We measure it by calculating the average number of technologies covered in
each alliance agreement in the portfolio. That is, we ﬁrst count the total number of
technologies (including the repeated ones) in the portfolio for a ﬁrm up to year t and then
divide it by the portfolio size up to year t.

Formally,

TC!!!"
SCO item = "cum
At:
This construct will take on a value of 1 when there are as many technologies (including
the repeated ones) as there are alliances, which indicate the ﬁrm, have a preference to

include only one technology in each alliance agreement. Otherwise, the value will be

greater.

Knowledge Breadth (KLBR)

Knowledge breadth has normally been measured by the number of classes in which a
ﬁrm holds patents. However, our sample consists of a single industry and as such most of
their patents fall into to classes 514 and 424. Both classes are deﬁned by USPTO as
“Drug, bio-affecting, body healing compositions”. When most of the patents are limited
to just two classes, the count of classes as a proxy for a ﬁrm’s knowledge breadth may
not be accurate (Prabhu et a1. 2005). As such, we decided to use the subclass under these
two classes as indicators of a ﬁrm’s knowledge breadth. As such, knowledge breadth is
measured as the number of patent subclasses under class 514 and 424 for each ﬁrm up to

yeart (Prabhu et al. 2005).

68

Knowledge Depth (KLDP)
Following the Operationalization approach used by Prabhu, et al., (2005), we measure
knowledge depth as the average number of approved patents per patent subclass for each

ﬁrm up to year t.

Radical Innovation (RINO)

We measure this construct as the total number of new radical drugs that are approved for
ﬁrm i in year t. A new drug’s radicalness is based on two factors: its therapeutical
potential and technology type. FDA assign a new drug a priority review status if they
conﬁrm that a new drug offers signiﬁcant better therapeutical beneﬁt to patients than
other drugs already on the market. If not, it will be a standard review status. FDA also
distinguishes chemical type used in the new drug. If it is involves an ingredient that has
never been used, it is labeled as chemical type 1. We classify a new approved drug as a
radical innovation if it is classiﬁed both priority review and chemical type 1. This
classiﬁcation is consistent with practices in the literature (Rothaerrnel and Deeds 2004;

Sorescu et al. 2003; Wuyts et al. 2004).

Incremental Innovation (IINO)
We measure this construct as the total number of new incremental drugs that are
approved for ﬁrm i in year t. All drugs that do not satisfy both radicalness conditions are

classiﬁed as incremental innovations (Rothaerrnel and Deeds 2004; Sorescu et a1. 2003;

Wuyts et al. 2004).

69

Radical Innovation Stock (RISD
We measure this construct as the total number of new radical drugs that are approved for

ﬁrm i up to year t.

Incremental Innovation Stock (115 T)
We measure this construct as the total number of new incremental drugs that are

approved for ﬁrm i up to year t.

Table 3 lists the measures of the constructs in the study and their corresponding data
sources.
A correlation matrix among the key constructs and control variables with their means and

standard deviations are presented in Table 4.

7O

 

$0:

8080080 0&8...“

a 8 88.88 080 .85 0808 08:08 >05 .m 8wa 08000: :3,
023 05. .08: x8>0 80800me 08 8 888m 080 08:08 >80
>05 mm _ 0» 0020 0a E? z .880 608090.— 08 000208 85 a 80%
0“ a: m 8.6 .8.“ 020.808 028:8 05 8 08888 888 05 mm ELKK

_u.~.

xi I a

23.0%“ I. SERROHDH—Ks
.x.

...
:8 V

.._

ﬁrmk

FADE

 

m<0md

808.88 888 8>0 85388 028:8

98m m: 00.88% 88m 8 80:3 08w 8 08—0 mm x 88 88.88 0888
8 >80 :0 00:8 8.5 8 80:3 080 08:00 80888080.— 8088a 05.
620.808 mm 8.8 8 085

:8 m0 00808000 0383880 05 8 038—08 _. 8.8 m0 00.888000

m0 8080808 08 08000808 ad 80.3. .28 8 _. 8.6 85 @058 8A
8.5 05 88 008: .8 8988 08 080808 03 J 80% 0“ a: _ 8.6 80m

. s
.s aﬁ. T:
i u . @0883

8:
a .2

7.4
.3.
kw
s

<mmm

 

@4va

.88 3 088 09 EB
03? 08 683850 88588 808.850 515 080800me 0: 80
08080 3038088 85 05 .2 ﬂ M0 0305 a 80 08» EB : a 80>
00 a: _ 8.5 “on 808:8 m0 8988 0288830 05 mm 5.? a 80%

3 a: 8 85 80.8 808.88 888% .«0 8:83: 038—888 05 mm 5.0.x

s?
Earn—AN

 

mmHQ

 

$0:

808.88 8808.8: 8% 00:08.88 0.85
a 88088 H 8 T 800382 02.? 8 08: 53 z .H 80% 8 a: _ 88m
80m 002858 m0 838:: 0383830 05 8 shw a 80» 8 a: m 85

.5.“ 002828 8083 m0 898:: 0388880 05 mm 53$ .8 80% 9

Q: ~ gm how woogS—w ~SﬂONtO£ MO HODESH— 0>Um_=§o 05 mm Ea”:

2

50V
.58»: I 23”.:

<>Om

 

 

000.83 8.5

 

0:80:3—

 

08802

 

8:80:00

 

 

000.80m «:5 can 80583—880520QO 30:50:00 m 038,—.

71

828888 8808088 ”OZ:
”8088888 8088 ”022 ”8808 08023080— 89158 88808 08008508. ”~8qu ”0800.0. 008:8 “800m ”00888 88.88 08888

 

 

 

 

 

 

 

 

 

 

 

 

”8.1582 ”00088080808 88.88 Emma ”8088808 88.88 89mm ”00888 8088> 88 008838 880888 .8 80808.88 2505
888
880.58 880% 8 88 020.888 0.8.8 8 8 0008828 Q88 .8 80.888 888. 080 020.888
00888888
8888800 08008 82 8.88888
88.8800 008m 080 888
08880880
8888800 Q88
_ 88
888 ”083 808080
880» 8 888888 <98 30308 888880 .32 OZ:
«58 8030008 88 m 88 .8 88.8 8808088 308 8 8888 888 088. ”888088 8088088088.
8 ”088 808880
8.80% 8 8>08888 3080.. 88088 .888 0258
<Q8 <98 8030008 88 8 8.8 .8 8:88 8088 B08 .8 88888 888 088. ”888088 8088080888.
8 80% 8 8: 8.8 880
8088890.?53 88 08—085 8088 88 88088 8088888 .8 88888 0888 088. 894M
.8? 888 3m 9.88 8888 8 808 8 88 08088
80.8889088m3 8088888 0.88 8080 88 80880 888—0880 8088 .8 8882 8mg
.8me 08 :85 88> 08 683880
.30888008 080 88888 880 8080088 b0>0 .2 8 .8 088>
8 0088 :85 z .8 80» 08 88 m 8.8 88 020.888 08 8 8080 8080808 5&8 800m
8<0mm 08 888085 0080888008 .8 88888 888 08 8 0888 SE 23mm
80.8% 8.5 088053— 08802 88.50800

 

 

 

 

 

20.8ch m 038,—.

72

888 0808808 ”Nmna 8288888808088 ”OZ:
828888 808088 ”028% 8808 080083088 ”mm-U8 ”888088 08008888 ”Mag "08008 008888 ”800m ”008888 88.88 28888
H 8.152 80080080808 88.88 “888% 8280808 88.88 ”89D ”008888 8088> 88 008888 8808.808 .80 800808.888 H<>Omv
80>0_ mod 8 8808888 2 8288800 ._. 808— 86 8 8808880 8 82880800 .1.

 

88.388 08.2.... @288 80:83 VmN mm. ma. Sim: mm. wﬁ 8m. 3. Eu. 8.0
3.8m: wodm $8.58 888$: 28 mo. Nod 3.3. mm; 08.8 cm. a. co. 8802
8 .308. .188. :88. *...Nm. if. .358. .850. cc. :3. 3.3.- ....L.N.- *._.E.mEOOZHHmZ .3
8 .180. ion. iam. .Ibm. ..:..Nm. imw. 1.2. 1:. .398.- ..szV 3N8. Nmmm .2
8 .283. Slum. 1.88. 1.9m. 18m. .18 8. :3. tom: ...*mm.- 8.8:. Comm .2
8 3.8. mo. .32. ......mm. Sr 5.- 8.. 8.. *8. mmimdﬁ
8 ._:_.mm. ..:L 8. 3.3.. No. :3. SEN.- ..ivmr .32. OZ: .8
8 .182. tom. 80. .32. .33.- ..iomr we. OE .w
8 3.3. 80. co.- .12.- ...*mm.- .13. 8O; .8
8 .288. .33. $.08.- ..cwwvr 1.3. mm; .o
8 38:. tom: .3; 8.- .388. 800m .m
8 **NN.- 3&8.- Nor 8.52 .v
8 .183. S.- <mmm .m
8 co.- 89D .N
8 8503.8
2 N8 : o8 8 w 8. o m .8 m m 8

 

80888800 80v— 80 8.582 828—0800 8 0—888.

73

3.5 Analysis and Findings
3.5.1 Model Speciﬁcation

A close examination of our data set suggests that there are several characteristics
of the data that may require specialrconsideration to obtain robust results.

First of all, we collected data on a ﬁrm-year basis and obtained a panel dataset (66
ﬁrms across 15 years of time). For econometric analysis of panel data, assuming the
observations are independently distributed over time is thus unfounded for our dataset
(Wooldridge 2003). Our model speciﬁcations need to account for the heterogeneity
caused by the unobserved ﬁrm-speciﬁc effects as unobserved factors other than those
explicitly addressed in our model could also bear some inﬂuence on the dependent
variables (Prabhu et al. 2005). In order to truthfully assess the inﬂuence of our
hypothesized variables on the dependent variables, we need to control for these
unobserved effects. Secondly, some of the dependent variable such as knowledge
breadth, radical innovation, and incremental innovation are measured as count variable
(nonnegative and integers). Ordinary least squares is thus inappropriate for count data as
the distribution of the count variables may not follow normal distribution patterns
(Sorescu et al. 2003).

To successfully account for the heterogeneity and count variable problems, we
formulate a Poisson Random Effects model when dependent variable is count variable
and a simple random effects model when it is not. Using a random effects model over a
ﬁxed effects model is consistent with practices in the literature concerning knowledge

and innovation related topics (Prabhu et al. 2005; Sorescu et a1. 2003).

74

Speciﬁcally, to test the inﬂuence of R&D alliance portfolios properties on

knowledge development, we use the following model speciﬁcation:

KLBR. = A. + ADISB, + ﬂzRPE/t, + AMUL 7:, + @5601: +
ﬂSHO VA" + ,B6SALES,, + ,B7R & D, + ,BgPFSZ, + 4’, + vi,

KLDPit = 160 +ﬂIDISBt +ﬂ2RPEAit +ﬂ3MUL72't +ﬂ4SCOPit +
,BSHOVA, + ,BéSALES, + ,67R & D, + ,68PFSZ,., + 4,. + 0,,

Where 4, is unobserved ﬁrm-speciﬁc effect

0,, is the remaining error term.

To test for knowledge development impact on innovation outcomes, we estimate

the following model:
RING“ = ,60 + ,6,KLBR,, + ,BZKLDPi, + ,63SALESR + A)? & D, + 4". + 0,,
IINOi, = ,60 + ,BIKLBR” + ﬂzKLDP“ + ,BSSALES, + ,64R & D, + {i + 0,,
Where 4, is unobserved ﬁrm-speciﬁc effect

U“ is the remaining error term.

To test for ﬁrm’s innovation impact on ﬁnancial performance, we estimate the

following model:

NETINCOME, = A, + ,6,R1N0,, + ﬂ,11N0,, + ,63RIST” + ,6,IIST,.,
+ ,BSSALES,, + 4". + u,

Where {i is unobserved ﬁrm—speciﬁc effect

0,, is the remaining error term.

3.5.2 Statistical Results

3.5.2.1 R&D Alliance Portfolio Characteristics and Knowledge Development

75

We estimated two models to show the impact of ﬁrms’ R&D alliance portfolio
characteristics’ effect on their knowledge breath and knowledge depth development. It is
important to note that while knowledge breadth is a count variable which is thus
estimated as a Poisson Random Effects model. The knowledge depth though is not count
variable and is estimated using a simple random effects model.

Table 5 presents the statistical results of the two models. H1 argues that a ﬁrm’s
preference for horizontal partners over vertical alliances has a negative impact on its
knowledge breadth development. The coefﬁcient of horizontal/vertical is negative and
signiﬁcant ([3 = -0.08, p<0.1). Therefore, H1 is supported. H2 suggests that a ﬁnn’s
preference for horizontal partners over vertical partners has a positive impact on its
knowledge depth development. However, the coefﬁcient is not statistically signiﬁcant.
Therefore, H2 is not supported

The coefﬁcient of alliance portfolio dispersion on knowledge depth is negative
and signiﬁcant (B = -2.81, p<0.01). This is a contrary ﬁnding to H3 in which we
hypothesized a positive relationship. However, H4 is supported as the coefﬁcient on
knowledge depth is negative and signiﬁcant (B = -1.75, p<0.01) indicating dispersion of
the portfolio partners negatively inﬂuences its knowledge depth development.

Partner repeatedness is hypothesized to have a negative impact on knowledge
breadth (H5) and a positive effect on knowledge depth (H6). The result supports H5 as
the coefﬁcient is negative and signiﬁcant ([3 = -1.87, p<0.01). We ﬁnd, a signiﬁcant and

negative coefﬁcient ([5 = -O.41, p< 0.01) for knowledge depth though which contradicts

the hypothesized relationship in H6.

76

Results support both H7 and H8. H7 hypothesized a positive relationship between
multiple partner alliances and knowledge breadth. The coefﬁcient is positive and
signiﬁcant (B = 0.39, p<0.01). Thus, H7 is supported. H8 is also supported with a
negative and signiﬁcant coefﬁcient (B = -0.55, p<0. 1) indicating a negative relationship
between multiple partner preferences and knowledge depth deve10pment.

The result also supports both H9 and H10. H9 hypothesized a positive
relationship between alliance scope and knowledge breadth. The coefﬁcient for this
relationship is positive and signiﬁcant (B = 0.42, p<0.01). H9 hypothesized a positive
relationship between alliance scope and knowledge depth. The coefﬁcient for this
relationship is also positive and signiﬁcant (B = 0.23, p<0.05).

Overall, the result largely supports the general idea that alliance portfolio

conﬁgurations have signiﬁcant impact on ﬁrms’ knowledge development.

3.5.2.2 Knowledge Development and Innovation Output

Again, we estimated two models to test the impact of knowledge breadth and
knowledge depth on two types of innovations. Both dependent variables are count
variables and thus both were estimated as a Poisson Random Effects model.

Table 6 presents the result of the analysis. The results generally support H11 and
H12. H11 hypothesized knowledge breadth will have a greater impact on radical
innovation than knowledge depth does. The result support H11 as the coefﬁcient of

knowledge breadth on radical innovation is positive and signiﬁcant (B = 0.01 , p<0.05)

while the coefﬁcient of knowledge depth is not signiﬁcant (B =-0.02, p=96).
H12. hypothesized that knowledge depth will have a greater impact on incremental

innovation than knowledge breadth does. The coefﬁcient on knowledge depth is positive

77

and signiﬁcant (B = 0.36, p<0.05). However, we found a negative and signiﬁcant

coefﬁcient for knowledge breadth on incremental innovation (B = -0.00, p<0.01).

Table 5 Alliance PortfolioCharacteristics on Knowledge Development

 

 

 

 

Independent Variables Knowledge Breadth Knowledge Depth
Intercept 6.02*** 3.60***
Partner Repeatedness -1.87*** -0.41***
HorizontaWertical -0.08* -0.10
Alliances Scope O.42*** 0.23**
Multi-Partner 0.39*** -0.55*
Dispersion -2.81*** -1.75***
Firm Size -0.00 0.00
Portfolio Size 0.00*** 0.01***
R&D Expenditure -0.00 -0.00
Log Likelihood -2956.05
Overall R2 0.28
Wald X2 8173.76*** 892.79***
Unstandardized coefﬁcients are shown.
*** p<0.01
** p<0.05
*p<OJ

78

Table 6 Knowledge Development and Innovation Outcome

 

 

 

 

 

 

 

 

Independent Variables Radical Innovation Incremental Innovation
Intercept -2.25*** 0.44
Knowledge Breadth 0.01" -0.00***
Knowledge Depth -0.02 0.36“
Firm Size -0.00** 0.00
R&D Expenditure -0.00 -0.00
Log Likelihood -193.73 -767.87
Wald x2 16.67*** 68.11***
Unstandardized coefﬁcients are shown.
*** p<0.01

** p<0.05

Table 7 Innovation and Firm Financial Performance
Independent Variables Net Income
Intercept 85.20
Radical Innovation Stock 3909*
Incremental Innovation Stock 4465*“
Firm Size 0.04***
Radical Innovation Launch -319.81***
Incremental Innovation Launch -4.20
Overall R2 0.66
Wald )0 1055.31***
Unstandardized coefﬁcients are shown.
*** p<0.01
* p<0.1

79

:1

3.5.2.3 Innovation and Firm Financial Performance

We estimated a simple random effects model to test the relationship between
radical innovation stock and incremental innovation stock on ﬁrm ﬁnancial performance.
Table 7 presents the statistical results.

As hypothesized in H13 and H14, both radical innovation stock and incremental
innovation stock have a positive impact on ﬁrm performance. H13 was supported as the
coefﬁcient of radical innovation stock on performance is positive and signiﬁcant (B
=39.09, p<0.1). The coefﬁcient of incremental innovation stock on performance is also
positive and signiﬁcant (B =44.65, p<0.01) indicating a positive relationship between a
ﬁrm’s incremental innovation stock on its performance.

We controlled for the number of radical and incremental innovations for that year
as we would expect the introduction of new drugs would incur abnormal expenses in
manufacturing and distribution, which would negatively impact the bottom line. Our
results indicate that introducing radical innovation will signiﬁcantly inﬂuence the bottom
line of a ﬁrm with a negative and signiﬁcant coefﬁcient of B =-3l9.81 (p<0.01).

Table 8 provides a summary of our hypotheses testing results.

80

Table 8 Summary of Hypotheses Testing Results

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Hypotheses IV DV Hypothesized Result
Relationship
H1 HOVA KLBR - Supported
H2 HOVA KLDP + N.S.
H3 DISP KLBR + Contrary
Finding
H4 DISP KLDP - Supported
H5 RPEA KLBR - Supported
H6 RPEA KLDP + Contrary
Finding
H7 MULT KLBR + Supported
H8 MULT KLDP - Supported
H9 SCOP KLBR + Supported
H10 SCOP KLDP + Supported
H1 1 KLBR>KLDP RINO + Supported
H12 KLDP>KLBR IINO + Supported
H13 RIST NET INCOME + Supported
H14 IIST NET INCOME + Supported

 

(HOVA: preferences of horizontal alliance over vertical alliance; DISP: partner
dispersion; RPEA: partner repeatedness; MULT: multiple partner alliance; SCOP:
alliance scope; KLBR: knowledge breadth; KLDP: knowledge depth; RINO: radical
innovation; IINO: incremental innovation; PF SZ: portfolio size)

81

3.6 Discussion

Our ﬁndings support our central argument that the conﬁguration of R&D alliance
portfolio has signiﬁcant impacts on a ﬁrm’s knowledge development. We found that
those managers’ preferences for different types of agreements and partners inﬂuence its
ability to develop knowledge inventory. We also found that knowledge depth and
knowledge depth play different roles in ﬁrms’ radical and incremental new product
development.

Next, I would like to offer some potential explanations to the unexpected
signiﬁcant effects of portfolio dispersion on lmowledge breadth and partner repeatedness
on knowledge depth development. An extremely dispersed portfolio suffers from the
lack of trust and common knowledge base for learning to occur. Even though partners
possess complementary skills, they will not be able to assimilate it if they are not willing
to share with each other. Even if they are willing to share, the lack of common knowledge
base as a requirement for learning may also affect how well they can assimilate
knowledge form each other. As such, extreme alliance portfolio dispersion may result in
a negative impact on a ﬁrm’s knowledge breadth.

The negative impact of highly repeated partnering could probably be due to the
limitation of knowledge stores of partners can share with each other. Even though the
social relatedness and trust level is high and ﬁrms are thus willing to share with each
other what they know, the sharing is limited to what they know. As the repeated
partnering gets overly redundant, partners may ﬁnd it difﬁcult to ﬁnd anything new to

offer to their partners as the proprietary information from each other is depleted.

82

The negative inﬂuence of knowledge breadth on incremental innovation is yet
another surprising ﬁnding. One explanation could be offered that when ﬁrms overly
expand its knowledge domains, management will lose focus on any particular knowledge
sector. Information overﬂow and knowledge over-inventory may thus limit
management’s ability to identify new opportunities for innovation possibilities.

3.6.1 Theoretical Contribution

Our study is a timely contribution to the call for more research on alliance issues
from a portfolio approach (Bamford and Ernst 2002; Hofﬁnann 2007; Wuyts et al. 2004).
Our framework greatly enhances our understanding of the conﬁguration effects of ﬁrms’
R&D alliance portfolio on its knowledge and innovation inﬂuence. Wuyts, Dutta and
Stremersch (2004) have pointed out the importance of examining the alliances from a
portfolio perspective. They identiﬁed two related descriptors: technological diversity and
repeated partnering. Our framework builds on this work and advances our understanding
by incorporating a more complete array of agreement characteristics and empirically
demonstrated their inﬂuences on ﬁrms’ knowledge development and innovation output.

Our study also makes a major theoretical contribution by incorporating
knowledge into the framework. Previous studies normally have made a direct linkage
between alliances and innovation (e. g., Rothaerrnel and Deeds 2004) or directly with the
ﬁnancial performance (6. g., Goerzen 2007; Goerzen and Bearnish 2005). Even though
we did not formally test the mediation function of knowledge and innovation on ﬁrm
ﬁnancial performance, the differential impact of a particular R&D alliance portfolio
characteristic on knowledge breadth and knowledge depth as well as the differential

direct inﬂuence of knowledge breadth and knowledge depth on radical innovation and

83

incremental innovation indicate that the relationships could be much more complicated
than what we had assumed in the past studies.

Past studies in establishing a linkage between alliances portfolio descriptors to
economic performance has normally acknowledged that there could be different theories
that would lead to quite opposite conclusions even for the same descriptor. For example,
Goerzen and Beamish (2005) used both network theory and transaction cost analysis in
analyzing the effect of network diversity on ﬁrm performance. These two theories
predicted two contrary conclusions. Even though their ﬁndings in the end support the
transaction cost analysis prediction which is quite contrary to some other ﬁndings in the
literature, it is a difﬁcult conclusion to make that the network theory has limited
application in alliance management given the vast literature in support of it.

In another study of Goerzen (2007) in which partner repeatedness is the focal
characteristic, network theory and transaction cost analysis are both used in making
opposite predictions. The ﬁndings in the study supported network theory predictions.
Again, in view of the mass literature on the transaction cost analysis, coming to a
conclusion that transaction cost analysis has a limited application is not completely
convincing. Considering that network theory normally predicts along the line of resources
and knowledge heterogeneity and transaction along the line of relational embeddedness,
our ﬁndings suggest that there is actually a ﬁne balance that ﬁrms need to keep in terms
of the beneﬁt of these two dimensions. Going extremes in either high dispersion or
overly repeated partnering could be detrimental on a ﬁrm’s knowledge breadth or depth

development.

84

We also make a theoretical contribution to the innovation literature by
demonstrating the differential impact of knowledge breadth and knowledge depth on
innovation output. Our ﬁndings are complementary to that of Prabhu et al. (2005) who
identiﬁed the differential impact of different sources of knowledge on a ﬁrm’s innovation
capability: internal knowledge and external knowledge. Our ﬁndings mostly examined
the dimensions of knowledge and their differential impact on radical and incremental
innovations. Precious research has mostly focused on the differential ﬁnancial
implications of radical and incremental innovation (e. g., Ahuja 2000b; Dewar and
Dutton 1986; Sorescu et al. 2003; e. g., Wuyts et al. 2004). Our study demonstrates that
there are different knowledge dimension requirements in order to come up with these two
types of innovations. While incremental innovation is most directly inﬂuenced by
knowledge depth, radical innovation is only directly related to a ﬁrm’s knowledge depth.
As such, a theoretical argument can be made that research on alliances’ impact on
innovation need to clearly differentiate two types of knowledge as well as two types of
innovations in the future.

One caution needs to be made in terms of the interpretation of the knowledge
dimensions and different types of innovation. It would be incorrect to state that
knowledge depth is not important for radical innovation or knowledge breadth is not
important for incremental innovation. As we only tested for the main effects of the
knowledge dimensions, we can only conclude that developing knowledge depth without
developing breadth will not lead to radical innovation. The same can be said about the
incremental innovation. We can only conclude that developing knowledge breadth

without developing knowledge depth will not lead to incremental innovation outcome.

85

There could be, however, interaction effects between these two dimensions of knowledge
that would be beneﬁcial to both types of innovations. As that is beyond the scope of our
study, it offers a great opportunity for future research.

3.6.2 Managerial Contribution .

First of all, alliance portfolio conﬁguration has a large impact on how well ﬁrms
can build up their knowledge reservoir. From partner selection, alliance structure, and
alliance content, ﬁrms can strategically formulate and conﬁgure their alliance portfolios
to achieve knowledge development objectives. If the strategic goal of a ﬁrm is to expand
its knowledge breadth, that is to expand on the number of domains with which they have
skills, ﬁrms can ﬁrst of all increase their partnership with vertical partners. Vertical
partners bring the beneﬁt of complementary rather than similar skills and technologies.
Knowledge sharing among vertical partners is also enhanced by the fact that they are not
directly competing with each and ﬁrm will be less concerned about the opportunistic
behaviors when they disclose information to their partners. The combined effect of fresh
information base and the willingness to share would certainly help them expand their
knowledge breadth more efﬁciently.

From an alliance structure perspective, they could also increase their knowledge
breadth by incorporating multiple partners in each individual alliance. Including multiple
partners in the agreement is equivalent to setting up multiple alliances simultaneously
from a learning perspective. Firms are exposed to the knowledge stores of all the
partners. The effect of simultaneous access to different sources is more than an additive

game. Firms could potentially combine knowledge they learned from different partners

86

and put them in new perspective and integrate them in creative ways which will result in
completely new insights.

Firms also need to restrain from showing extreme preference for a subset of
partners, as this will be an inhibitor of knowledge breadth development. Repeatedly
partnering with the same partner limits ﬁrms’ exposure to new and fresh information.
Their knowledge is thus limited to the extent what that particular partner has to offer. As
the technology develops rapidly in the market place, over preference of a partner may
create a lock in problem for the focal ﬁrm as they continuously invest in the partner
speciﬁc capabilities.

Changing the content of the alliance agreements could also expand ﬁrms’
knowledge breadth. Instead of focusing only on one technology in the agreement, alliance
partners can incorporate multiple related technologies to increase the span of knowledge
sharing. These related technologies would foster a better learning opportunity among
partners due to the increased volume and breadth of learning.

For ﬁrms whose strategic goal is to develop its knowledge depth, certain
strategies could also be formulated from this study. First of all, they need to limit their
portfolio dispersion. This means that always looking for new partners to work with may
not be a good idea for knowledge depth development. Knowledge development needs
two critical conditions, ﬁrms’ willingness to share and ﬁrms’ ability to learn. Both
conditions require the nurturing of social capital between partners. Due to the contracting
hazard problem associated with knowledge transfer, trust between partners is a
determinant of whether partners will be willing to disclose and share sensitive and critical

information with each other. On the other hand, ﬁrms need to build some common

87

knowledge in order to facilitate the learning process. Always looking for new partners to
work with indicates that the every time an alliance is formed, the knowledge bases
between partners may be completely different. This would create the problem of inability
to learn, as partners do not share the bases for learning.

They should also try to limit their use of multiple partner agreements. Due to the
increased potential of opportunistic behaviors, ﬁrms in this kind of alliances may only
share superﬁcial or public knowledge with each other but would be reluctant to share
critical and sensitive information. This would work against the learning objective of the
focal ﬁrm as mostly what partners share with each other are what they already know.
Knowledge depth thus cannot be developed efﬁciently.

Lastly, they could also work on the content of the alliances agreements by
increasing the number of technologies partners collaborate on. As discussed before, the
relatedness of these technologies within each agreement should have a facilitating impact
on the understanding of knowledge elements. In addition, due to the high relatedness of
multiple technologies, their combination or integration will result in new understanding
as well.

Our ﬁndings also suggest that there is something ﬁrms can do to achieve both
knowledge breadth and knowledge depth development simultaneously. The strategy
would be to increase the alliances scope as it is found to be positively related with both
knowledge-related dependent variables. All the other strategies favor the development of
one knowledge dimension but inhibit the other.

Our ﬁndings also shed on some light on knowledge development strategies for

innovation purposes. This study provides insights in terms of what dimension of

88

knowledge ﬁrms need to focus on depending on the innovation type they are after. For
ﬁrms that emphasize radical innovations as this type of innovation normally has a better
marketing position compared with incremental innovations, their strategic focus should
be on obtaining newer and fresher information. They need to develop their ability to
integrate knowledge they learned from different sources to come up with innovative
products. On the other hand, when a ﬁrm’s strategic goal is to come up with
incrementally new products (mature industry), their focus of learning should be on
knowledge depth development. Try to have a deeper and better understanding within the
established technology domain is proven to be critical for incremental innovations.

To summarize, our ﬁndings offer a complete framework on how to achieve
innovation objectives through successful knowledge management. And successful
knowledge management comes ﬁom careful conﬁguration of R&D alliance portfolio.
Management needs to use a back loop style of thinking when using ﬁndings from this
study. They should ﬁrst identify what their innovation objectives are as radical
innovation may not be the optimal route for all ﬁrms and all industries. F inns with
limited resources and with industry constraints may consider incremental innovation as
the strategic priority. Once the type of innovation is identiﬁed, our study should guide
them in terms of the appropriate knowledge development strategy that should help them
to achieve their goals. Again, ﬁndings in our study would provide suggestions on how
they should conﬁgure their alliance portfolio to secure the necessary knowledge
elements.

3.6.3 Limitations and Future Research

89

Our study has several limitations that need to be noted. First of all, the
generalizability of our study needs to be considered as we only sampled publicly traded
pharmaceutical ﬁrms. Even though we only attempt to generalize the ﬁndings to
technology intensive industries such as telecommunication and electronics, the
applicability of the ﬁndings in other industries needs to be carefully evaluated. This is
particularly a concern when we think about the stringent government regulations imposed
on the innovation measure that we adopted in the study.

The second limitation is that causal inferences of our ﬁndings are limited, as we
did not take time lags in our model. This is largely due to the nature of our data. We have
a panel data set that includes count variables and potentially lagged count dependent
variables. Econometric techniques in dealing with all three issues (panel data, count
dependent variables, lagged effect) involved here are still in their infancy (Prabhu et al.
2005). Even though the generalized method of moment (GMM) has the potential to
account for all the unique characteristics, we did not use it in testing our model because
our subj ect/time ratio is considerable low compared with what is required for the
application of GMM. As such, we could only make inferences in terms of the correlation
not causality.

Our framework only depicts a direct effect model. The knowledge depth and
knowledge breadth dimensions offers future research opportunities in terms of
understanding how knowledge breadth and knowledge depth interact to impact radical
and incremental innovations. A ﬁrm’s knowledge depth would be helpful in choosing the
target with the most promising knowledge, absorbing the knowledge more rapidly, and

exploiting it to come up with innovative products (Prabhu et al. 2005).

90

There could also be some moderating factors that worth further studies. Even
though the discussed hypotheses are based on the assumption of knowledge heterogeneity
across ﬁrms, the breadth and depth of knowledge that each ﬁrm is exposed to only
represent the potential of the knowledge learning. In order to realize the knowledge
beneﬁts from these portfolio characteristics, ﬁrms need to ensure that knowledge is
effectively transferred across the ﬁrm boundaries. Future studies could examine the
mechanisms that would facilitate the knowledge transfer process.

By deﬁnition, inter-ﬁrm R&D alliances involve two or more ﬁrms combining
resources for some agreed upon activities. Partners have fewer incentives to make such
investments ex ante if they might be subject to ex post opportunism in the distribution of
resulting surpluses (Oxley 1997). Equity participation reduces the possibility of
opportunistic behavior by aligning the incentives of the partners. Their mutual interest in
the success of collaboration reduces problems of moral hazards often associated with
cost-plus scheme. Opportunism by an equity partner is penalized through the reductions
in the value of its equity holding. As a result, ﬁrms have less concern for their knowledge
sharing behaviors. In addition, in certain equity based R&D alliances such as joint
ventures, the environment provides much better opportunities for personal contacts,
teaching, participation, and communications among partners which are necessary for the
acquiring difﬁcult-to-articulate knowledge (Polanyi 1966). As such, the effect of equity
R&D alliances in a ﬁrm’s portfolio needs to be further addressed.

The relationship between knowledge breadth and knowledge depth on radical and
incremental innovation also could be moderated by a ﬁrm’s ability to utilize the

knowledge. It would be reasonable to argue that a ﬁrm with a superior capability in

91

integrating knowledge may realize a higher rate of innovation output than ﬁrms who
lacks this capability. As such understanding the ﬁrm level capability on both types of

innovation and ﬁnancial performance offers another venue for ﬁiture research.

92

CHAPTER 4
ESSAY TWO: NON-ADDITIVITY EFFECT OF R&D ALLIANCE PORTFOLIO
SIZE ON INNOVATION: AN INTEGRATION CAPABILITY PERSPECTIVE
4.1 Introduction
New product development capability plays a critical role in a ﬁrm’s growth and
survival in the market place and ﬁrms need to constantly introduce innovative and better
products that meet market needs over time (Dougherty and Hardy 1996; Nerkar and
Roberts 2004; Penrose 1959). However, most ﬁrms are ﬁnding it increasingly difﬁcult to
develop new products purely internally (Sivadas and Dwyer 2000; Wind and Mahajan
1997). First of all, new product development is a highly demanding endeavor in terms of
both ﬁnancial and human capital resources. Firms need to make huge investment
commitment upfront without the guarantee that a new product will be developed.
Secondly, the risk associated with new products extends beyond the development stage.
Only half of the introduced new products will survive the market test during
commercialization (Page 1993; Schmidt and Calantone 2002; Zirger and Maidique 1990).
However, the most important reason ﬁrms can not take on the new product
development all by themselves may have something to do with the changing competitive
environment in the market place. In order to develop a truly innovative product, ﬁrms
need to pool knowledge from different areas(Cohen and Eliashberg 1997; Kogut and
Zander 1992; Prabhu et al. 2005). Increasing complexity in new product development and
fast advancing technology mandates ﬁrms to look beyond their boundaries to access
knowledge and technology for their new product development purposes (Sood and Tellis

2005; Wuyts et al. 2004).

93

As a result, inter-ﬁrm collaborations on R&D activities become a common
practice for ﬁrms to overcome the ﬁnancial constraints, share the inherent risk associated
with new product development, speed up the NPD process, and to get access to the most
advanced technologies not available within a ﬁrm (Sivadas and Dwyer 2000; Wuyts et al.
2004). R&D alliances are deﬁned as voluntary arrangements between ﬁrms to exchange
and share the knowledge as well as resources with the intent to develop a new product
(Gulati 1998).

Firms seldom rely on a single R&D alliance to achieve their innovation goals.
Due to the knowledge heterogeneity across ﬁrms in the market, they have to form
alliances with multiple partners either simultaneously or sequentially to establish and
maintain their competitive positions (Hoffmann 2007). Organizations enter into dozens of
R&D alliances as organizations evolve into loosely knit organizations (Bamford and
Ernst 2002). A conservative estimate would be 30 alliances for most of them and many
have more than 100 (Bamford and Ernst 2002).

Academic studies of R&D alliances have largely been through the lens of
knowledge-based view, an offspring of the classic resource based-view of the ﬁrm.
Studies following this tradition have drawn from works by Penrose (1959) and Teece and
Pisano (1992) and maintained that the inimitable ﬁrm heterogeneity, or the possession of
unique “knowledge”, “competences”, or “capabilities” may be an important source of
sustainable competitive advantage. Past research in this area has discussed the features of
alliance partners that facilitate the ﬂow of knowledge across ﬁrms (e. g., Cohen and
Levinthal 1990). Others examined the motivation differences in ﬁrms’ decision of

entering alliances (e. g., Grant and Baden-Fuller 2004). All these studies addressed

94

alliance issues from an isolation perspective by taking the unit of analysis at the project
level.

Scholars have also realized that understanding the R&D alliance portfolio issues
may be of greater value to business practitioners (Hoffrnann 2007). Firms need to
successfully conﬁgure their alliance portfolio in order to enjoy a “sustainable”
competitive advantage. As such, scholars have studied portfolio related topics from
different perspectives. These topics cover drivers of portfolio design (Hoffrnann 2007),
portfolio’s inﬂuence on innovation and proﬁtability (Wuyts et a1. 2004), and the impact
of alliance experience on subsequent alliance success (Hoang and Rothaerme12005;
Rothaerrnel and Deeds 2004).

Despite the rich literature on the R&D alliance portfolio related issues, ranging
from partner selection to alliance management, there is still one key question seeking a
satisfactory answer: the nature of the relationship between a ﬁrm’s R&D alliance
portfolio size and its innovation output. Existing studies on this question have offered
different ﬁndings. Some studies have found a positive liner relationship between the
number of R&D alliances of a ﬁrm and its innovation output (e. g., Baum et al. 2000;
Rothaermel and Deeds 2004; e. g., Shan et al. 1994). Other studies though have argued a
non-linear relationship between the size of a ﬁrm’s R&D alliance portfolio and its
innovation outcome (Rothaennel and Deeds 2004). As such, it is very inconclusive as to
the true functional from of the impact of the number of R&D alliances on innovation.
One research question that remains is then: how does ﬁrm’s learning capability and
combinative capability change the marginal returns on innovatiOn by adding one more

alliance into a ﬁrm’s R&D alliance portfolio? In this study, we draw on the literature of

95

organizational learning and combinative capability and argue that the relationship
between the size of a ﬁrm’s R&D alliance portfolio and its innovation output
demonstrates diminishing returns at low to moderate level and increasing marginal
returns at moderate to high level. The relationship is thus inverse S-shaped.

The following of the manuscript is organized as follows: we will ﬁrst provide a
brief review of the literature on alliance portfolio size and innovation. We then present
our theoretical arguments and develop hypotheses. We then provide an overview of the
research design including the sample, data sources, and measures. This will be followed
by statistical analysis and results. We wrap'this essay up with a discussion of the results.
4.2 Received Wisdom on R&D Alliance Portfolio size and Innovation.

R&D alliances provide ﬁrms with access to novel information and technological
know-how which can be an important input factor in new product development (Shan et
al. 1994). Studies addressing the relationship between the number of R&D alliances and
ﬁrm’s innovation output have all agreed that there should be a positive inﬂuence of the
number of a ﬁrm’s R&D alliances on its new product development success (Baum et al.
2000; Rothaennel and Deeds 2004). However, there are different views regarding the
actual functional form between the number of R&D alliances and new product
development.

Most studies on R&D alliances and innovation have postulated a positive linear
relationship between the number of R&D alliances a ﬁrm has and its new product
development. This is evident in several studies where number of alliances is used as
either a direct predictor or control variable in proposed models. One caution of word need

to be stated though that all these studies are merely interested in proving there is a

96

positive relationship between R&D alliances and innovation outcome. Their research
questions are not speciﬁcally addressing the ﬁmctional form regarding the impact. The
reason why we classify them in to the liner positive relationship category is because of
the statistical treatment they have used on the number of R&D alliances in testing their
models. Therefore, they are analyzing the relationship in a more general approach.

Two studies in support of the linear positive relationship are conducted in the
startups context. Startups, due to their newness to the industry, face considerable hazards
in securing their position in the market (Baum et al. 2000). Alliances with established
ﬁrms offer the beneﬁt of accessing to knowledge, obtaining critical resource, establishing
stability, and offering legitimacy in the industry (Baum et al. 2000; Shan et al. 1994).
Using data from BoiScan with a sample of 85 US biopharmaceutical ﬁrms, Shan, Walker
and Kogut (1994) found that a startup’s cumulative cooperative ties positively inﬂuences
their innovation output. The ﬁnding is supported by the Baum, et a1. (2000) study in
which the innovation output of the ﬁrm is positively related to the number of alliance a
ﬁrm establishes at founding.

Rothaerrnel and Deeds (2004) further classiﬁed R&D alliances into exploration
and exploitation alliances based on motivation differences. Exploration alliances are
motivated to discover new and novel technology and information. Exploitation, on the
other hand, aims to beneﬁt from existing technologies or know how. They postulate that
there are actually a linkage between these two type of R&D alliances as exploration
based R&D alliances motivate subsequent exploitation based R&D alliances. They also
further classiﬁed innovation into product in development and product on market. The

overall framework is that exploration positively inﬂuences product in development,

97

which positively impacts exploitation-based alliances, and this will positively impact
product on the market. Their framework is tested in the pharmaceutical industry and the
positive linear relationship between alliances and innovation obtained statistical support.

Due to the prevalence of the implicit positive linear relationship between number
of alliances and innovation output, studies that address alliances and innovation issues
normally control for the size of alliance portfolio in a linear fashion (e. g., Wuyts et al.
2004)

However, other studies have argued different functional forms between the
number of R&D alliances and innovation. Deeds and Hill (2004) argue that the ﬁmctional
form should be an inverted U-shape. They argue that even though the number of alliances
will have a positive impact on innovation output, anticipating a constant return from each
individual alliance may not be realistic. In fact, they argue that the marginal returns of
adding alliances to the whole portfolio should demonstrate diminishing and even negative
returns. The rationale is that not all alliances contribute equally to the focal ﬁrm. Firms
choose to ally with the most promising partners ﬁrst and the later partners will normally
contribute much less than previous partners. Thus the marginal return is expected to be
diminishing as the number of alliances grows. In addition, extremely large number of
R&D alliances also increases the burden of the management. Bounded rationality, the
limitation of management capability in managing an extended alliance portfolio will lead
to negative marginal returns. The inverted U-shaped relationship is tested in the
biotechnology industry and was statistically supported.

To summarize, there exists two formulations in the literature of the relationship

between the number of R&D alliances and innovation output: positive linear and inverted

98

U-shaped. Even though both forms were supported statistically, we would like to propose
an advanced functional form of the relationship and empirically compare it against both
existing forms in the literature. We would argue that even though the marginal retm'n of
alliances will decrease at the low to moderate level, it will actually increase after a critical
point where ﬁrms would have accumulated enough knowledge mass that would enable
the application for the combinative capability to increase the innovation output
exponentially. The functional form of relationship between the number of R&D alliances
and innovation output is inversed S-shaped as shown in Figure 3. We next present the
theoretical foundation and hypotheses development.

4.3 Theoretical Foundation and Hypotheses Development

4.3.1 Diminishing Returns with Limited Number of R&D Alliances

While we agree with previous studies that the total number of R&D alliances will
in general have a positive impact on a ﬁrm’s innovation output, we concur with Deeds
and Hill (2004) that the relationship is non linear. Several reasons may contribute to the
decreasing marginal returns of R&D alliances at low to moderate level.

First of all, all alliances partners are not created equal. Neither is their
contribution to the focal ﬁrm in new product development. Firms tend to ally with the
most promising partners ﬁrst (Hoang and Rothaermel 2005; Rothaerrnel and Deeds
2004). As such, diminishing returns on newer alliances become a natural result due to the
fact that the later partners will only make less contributions compared to earlier partners
as the complementary resources they possess will be relatively minor to the needs of the

focal ﬁrm (Rothaerrnel and Deeds 2004). Selecting the next best partner becomes

99

particularly challenging when the foundation of the collaboration is the tacit knowledge,
which is exactly the case for R&D alliances (Hoang and Rothaerrnel 2005).

Secondly, ﬁrms With limited alliance experiences also suffer from management
inability to effectively screen partners and monitor alliance performance (Rothaerrnel and
Deeds 2004). Due to the long durations of most R&D alliances which often last for years,
ﬁrms have to enter into multiple R&D alliances simultaneously. As the number of
alliances grows, managers with limited alliance experience may ﬁnd it increasingly
difﬁcult to monitor its partner’s behaviors and keep up its own learning actions (Hoang
and Rothaennel 2005). This will give a higher probability of partner opportunistic
behaviors and inefﬁcient learning from external sources. As such, with the bounded
rationality, ﬁrms may not be able to beneﬁt as much from their later alliances as they can
from the early few alliances.

Third, management’s inexperience in alliance management may also contribute to
the decreasing marginal return in terms of partner selection. F inns with limited alliance
experience may often enter into alliances with the expectation that the partners will
contribute complementary resources to their new product development efforts, only to
ﬁnd out ex post that the resources from the partner is a poor match for their new product
development goals (Rothaerrnel and Deeds 2004).

Due to the inherent risks associated with R&D alliances, particularly in terms of
potential partner opportunistic behaviors, ﬁrms may prefer to ally with partners that share
certain attributes with their successful prior partners. This in turn will contribute to the

decreasing marginal return, as there will be overlap of the complementary resources that

100

the new and the prior partners share with the focal ﬁrm. As such, the novelty and
complementarities of the resources of the new alliances will be discounted.

Even if ﬁrms in their early alliances stage venture into completely new partners,
the marginal return may still be decreasing. Firms then need to go beyond their comfort
zone and go through a process of dealing with unfamiliar entities. This process will
increase the odds of adverse selection in that ﬁrms may end up allying with partners with
inferior properties. Given the limited experience the focal ﬁrm possess in dealing with
alliance partners, the focal ﬁrm may draw on what it has learned from the prior limited
experiences, the strategy may not be appropriate for the current partner given its newness
in terms of resources and managerial styles (Sampson 2005). However, establishing new
processes and routines that would accommodate the current partner will increase the
managerial information-processing demand due to the lack of trust and familiarity
between partners (Goerzen and Bearnish 2005).

Finally, a major motivation of forming R&D alliances is to access new and novel
information. With limited knowledge base, any newly assimilated information may not
be readily mapped into the technological network within the ﬁrm and thus does not
directly contribute to development efforts the focal ﬁrm is actively engaging in. As such,
forming new alliances in the early stage may provide a venue for knowledge accessing
but may not be able to materialize the know how in face of limited internal knowledge
base (Macher and Boemer 2006).

As such, we hypothesize:

101

Hypothesis 1: The positive eﬂect of number of a ﬁrm ’s alliances on its innovation
output exhibits decreasing marginal returns when the alliance
numbers are low to moderate.

4.3.2 Increasing Returns with Expanded Number of R&D Alliances

Two major arguments support the idea of increasing marginal returns of R&D
alliances on innovation output once ﬁrms have accumulated certain amount of alliance
management experiences and knowledge base: the learning effect of alliance
management and the combinative capability.

Successful management of alliances has been regarded as a critical capability that
establishes competitive advantage for ﬁrms (Dyer and Singh 1998). This capability is
developed through the accumulation of repeated experiences and active learning over
time by dealing with a diversiﬁed portfolio of alliance partners (Hoang and Rothaerrnel
2005; Kale and Singh 2000).

Alliance management is not an easy task. Partners in an alliance often times have
competing interests and expectations (Sampson 2005). They may also have divergent
managerial styles, communication mechanisms, and business routines. These differences
make successful coordination across ﬁrm boundaries a daunting task. What ﬁrms can do
to overcome those barriers is to establish what scholars call “alliance capability” (Kale et
al. 2002).

Alliance capability is a direct result of participation in diverse alliance activities.
Experiences in managing alliance play a critical role in developing this capability (Hoang

and Rothaerrnel 2005; Sampson 2005). However, performance does not improve

102

automatically as the experiences accumulate. It is rather an outcome of the deliberate
learning that accompanies the alliance experiences (Pisano et al. 2001).

Organizational learning on alliances management occurs in an iterative fashion
(Hoang and Rothaerrnel 2005). When ﬁrms repeatedly participate in alliance activities,
they would accumulate some fundamental understanding of the coordination mechanisms
in dealing with partners and learn about the potential outcomes of various alliance
management practices and routines (Sampson 2005). These inferred learning is then
coded, stored, and disseminated within the organization that can be effectively activated
when certain stimuli is present in future alliance management activities (Hoang and
Rothaerrnel 2005).

As such, extensive experiences in alliance management help ﬁrms gain insights to
identify effective processes for exchanging information with their partners and to manage
complex activities with highly uncertain outcomes (Sampson 2005). In addition, ﬁrms
also gain hands-on experiences on how to disseminate the knowledge or know how they
learned from their partner within the organization by actually engaging in all these
activities (Sampson 2005). When all these experiences, understandings, inferences, and
insights are coded and disseminated across the organization, alliance capability has
emerged in the ﬁrm.

The alliance capability is said to have a positive impact on current and future
alliance outcome (Hoang and Rothaerrnel 2005; Kale et al. 2002; Sampson 2005).
Alliances capability enables ﬁrms to establish effective manual, databases, diagnostic
tools, and simulation systems to aid in the assessment of current alliance performance,

selection of future partners, and evaluation of alternative actions (Hoang and Rothaerrnel

103

2005). As such, a ﬁrm’s ample experiences would lead to increased marginal returns on
the number of alliances by securing higher beneﬁts ﬁom future partners.

The second major reason for increasing marginal returns once a ﬁrm has
established a critical mass of knowledge and know how is the combinative capability.
Henderson and Cockbum (1994) identify two types of “competences” that may serve as
the foundation of a ﬁrm’s enduring competitive advantage: component competence and
architectural competence. Component knowledge refers to the local capabilities and
knowledge that are fundamental to day-to-day problem solving and architectural
knowledge refers to the ﬁrm’s ability to use this knowledge by integrating them
effectively and developing fresh component competence. While both competences are
important elements for a ﬁrm to remain competitive in the market, the architectural
knowledge is more fundamental as they are more heterogeneously distributed within an
industry, impossible to buy or sell in the available factor market, and difﬁcult to replicate
as suggested by the resource based view (Barney 1991).

One key assumption of the knowledge-based view is that ﬁrms are heterogeneous
in terms of knowledge they possess. F irms seldom stick with one alliance but normally
enter into different alliances depending on the demand of the task at hand. In fact, ﬁrms
may simultaneously enter into different alliances with different partners. A larger number
of alliances bring the beneﬁt of more diversiﬁed knowledge and skills to the exposure of
the focal ﬁrm.

There are two views in terms of the motives of ﬁrms’ entering strategic alliances:
Knowledge accessing and knowledge acquiring. These two concepts correspond to two

conceptually distinct dimensions of knowledge management in the knowledge-based

104

literature. First, ﬁrms are interested in increasing its stock of knowledge and this is
referred by March (1991) as exploration and Spender (1992) as knowledge generation.
On the other hand, ﬁrms also engage in activities that deploy existing knowledge to
create value and this is referred by March (1991) as exploitation and Spender (1992) as
knowledge application.

In relation to R&D strategic alliances, knowledge generation points to alliances as
vehicles of learning in which each partner uses the alliance to transfer and absorb the
partner’s knowledge and skills. On the other hand, the knowledge application points to
alliances as vehicles to simply access the knowledge base of the partners in order to
exploit complementarities but with the intention to maintain its distinctive base of
specialized knowledge (Grant and Baden-Fuller 2004). Even though there is a clear
distinction conceptually about these two types of knowledge sharing, the accepted
wisdom is that learning occurs in all alliance activities. The only difference is the degree
or level of learning that occurs. For example, Inkpen (1998) observes that knowledge
acquiring occurs in all alliances while some are more aggressive than others. Grant and
Baden-Fuller (2004) also notes that learning occurs in all alliances and some alliances are
more motivated primarily by the desire to acquire partner’s knowledge. Hamel (1991)
explicitly stated that the acquisition of new knowledge is often the motivation for
establishing inter-organizational collaborations.

In this manuscript, we take on the idea that learning occurs in all strategic R&D
alliances. Even if the primary motive of some alliances is exploitation-orientated,
learning occurs due to the fact that ﬁrms need to reveal to a certain level of their

knowledge base to each other in order to accomplish the common goal. Knowledge

105

acquiring (learning) implies internalization of alliance partner’s expertise, skills and
processes. The internalization process leads to the fact that the focal knowledge becomes
an integrative part of the knowledge base of the ﬁrm. Once a ﬁrm has internalized the
knowledge and skills acquired ﬁ'om its partner, it often enjoys the freedom in terms of the
application of it (Oxley 1997).

Prior studies have suggested that knowledge diversity has a positive performance
effect given that diversity can enhance the breadth of perspectives, cognitive resources
and overall problem solving capabilities (Goerzen and Bearnish 2005; Hargadon and
Sutton 1997; Koka and Prescott 2002). Due to the fact that ﬁrms not only access
knowledge through their alliance activities but also acquire knowledge from their
partners (Mowery et al. 1996), the effect of obtaining one new piece of information or
technological skill on a ﬁrm’s innovation output may be more than additive. Once ﬁrms
internalize their partner’s knowledge, they could combine the newly learnt knowledge
with their existing knowledge to come up with innovation that is not possible with either
one alone.

The architectural competence of an organization allows it to make use of its
component competencies: to integrate them together in new and ﬂexible ways and to
develop new architectural and component competencies as they are required. Once a new
knowledge element is acquired from an alliance partner, the ﬁrm has at its discretion to
the use of it at its own will. Knowledge, unlike other resources that are physical in nature,
does not diminish its value by being employed somewhere else. That is, the value of a
piece of metal to task A is going to be lessened if it were to be used in task B. However,

knowledge does not suffer from this consumption deﬁciency. Kogut and Zander (1992)

106

argues that the central competitive dimension of what ﬁrms know need to know is to
create and transfer knowledge efﬁciently within an organizational context. They further
argue that knowledge of a ﬁrm is competitively consequential; leaming cannot be
characterized as independent of the current capabilities and knowledge base. When a ﬁrm
enters into an R&D alliance, the knowledge that it absorbs from partner may be
component in nature. However, this knowledge will not necessarily be locked in any
particular area where it is initially applied. Instead, knowledge ﬂows in an organization
and its applicability and value is realized by combining it with the existing knowledge
base of the ﬁrm.

Knowledge of alliance partner is primarily exposed to individuals that engage in
the daily communication with the partners. The ability of individual to transfer their
knowledge from one domain to another and to combine other knowledge in more
productive ways makes humans distinct from other resources (Penrose 1959). Innovation,
as an outcome of the new learning is a product of the ﬁrm’s combinative capability by
applying its newly acquired knowledge to its existing knowledge (Kogut and Zander
1992). Innovation is thus viewed as the new combination of existing knowledge and
incremental learning.

Due to the ﬁrm’s possession of the combinative capability, we can thus expect
that the marginal return to the diversity of a ﬁrm’s R&D alliances could be increasing
once ﬁrms have accumulated a critical mass of knowledge to apply and integrate the
newly assimilated knowledge. That is to say, a ﬁrm’s capability to come up with
innovative product or services displays super additive characteristic. In a standard

analysis of production, economies of scope are present when the costs of conducting two

107

or more activities jointly are lower than if they are conducted separately. This occurs
most obviously when these activities can share inputs at no additional cost, which is an
important aspect of knowledge based innovation (Henderson and Cockbum 1994). The
public goods aspect of knowledge means that knowledge capital acquired in one alliance
process may be utilized as a productive input to other related programs, at little or no
additional cost to the ﬁrm.

An illustration of this super additive nature of knowledge accumulation and
innovation is that when a ﬁrm is forming its 101 alliances, the effect of this alliance is not
simply an additive of another possible innovation to the ﬁrm. Instead, the ﬁrm may
combine the knowledge that it acquires from its 101 partner with knowledge that it
obtained with its ﬁrst 100 partners and the possibility of the innovation outcome clearly
goes beyond the simple innovation outcome of the focal activity of the 101 alliance itself.

Beneﬁts of diversity may also arise if discoveries made in one program stimulate
the output of another through cross fertilization of ideas and other forms of knowledge
spillovers. Henderson and Cockbum (1994) identiﬁes the fact that in pharmaceutical
research, several important central nervous systems therapies were discovered as the
result of the search for drugs in the cardiovascular system.

To illustrate how the combinative capability is established within a ﬁrm,
Hargadon and Sutton (1997) give an example of a California-based product development
company, IDEO, consciously attempts to leverage the heterogeneity of knowledge in the
generation of new ideas. At its brainstorming sessions, the ﬁrm not only assembles teams
with diverse knowledge but the team is encouraged to recall, recontextualize, and to

recombine elements of seemingly unrelated knowledge in the idea generation process.

108

Rodan and Galunic (2004) further argue that knowledge diversity is useful for the
implementation of the new ideas in innovation. It may help managers build a sound
causal understanding of the relationships between the elements in the complex system
(McGrath et a1. 1996). However, we expect the combinative effect will only take place
when there exists a critical mass of knowledge and know how within the ﬁrm which
allows for easy identiﬁcation of relatedness to integrate the newly assimilated knowledge.
Otherwise, ﬁrms may ﬁnd it difﬁcult to establish linkage and integration across
completely novel information with its current knowledge structure as the integration
possibility is limited by the knowledge base the ﬁrm.

To conclude the above discussion, we could expect an increasing marginal return
to a ﬁrm’s moderate to high number of alliances. Thus we hypothesize:

Hypothesis 1 .' The positive eﬂect of number of a firm ’3 alliances on its innovation
output exhibits increasing marginal returns when the alliance
numbers are moderate to high.

To summarize, we hypothesize that when ﬁrms only have limited number of
alliances, the marginal return will be diminishing due to the limited experiences ﬁrms has
in managing alliances and lack of knowledge mass for the combinative capability to be
effective. However, once a ﬁrm has passed a critical point in terms of the nrunber of
alliances it has, the marginal return will be increasing for the following two reason: 1) the
increased capability in managing alliances due to learning effect which result in more
knowledge intake by entering into a new alliance; 2) the knowledge mass that already
resides within the ﬁrm offers opportunity for the ﬁrm to combine the new knowledge or

know how with its existing knowledge to come up with novel innovation. As such, the

109

overall functional form of the relationship between the number of alliances and
innovation output will demonstrate an inverse S-shape as depicted in Figure 3.
4.3.3 Control Variables

In addition to the relationships‘discussed above, we control for other variables
that may affect a ﬁrm’s innovation output that are outside our theoretical focus.
Speciﬁcally, we control for R&D expenditure and ﬁrm size.
R&D expenditures. We would control for the R&D expenditure for knowledge and
innovation based outcomes. It is expected that the more resources devoted to research and
development, the greater ﬁrm’s knowledge base would be and the better their innovation
volume (Wuyts et al. 2004).
Firm size. We would also control for the size of the ﬁrm for all three types of dependent
variables (knowledge, innovation, ﬁnancial performance). The effect of ﬁrm size on
knowledge and innovation development largely stems for the seminal work of
Schumpeter (1942). Different effects (positive, negative, and insigniﬁcant) were
documented in the literature (e. g., Ahuja and Lampert 2001; Chandy and Tellis 2000;
Miles et al. 2000). The effect of ﬁrm size on performance is based on the idea that large
have advantage in obtaining lower cost of capital while simultaneously lowering risk
(Chang and Thomas 1989; Goerzen 2007). Previous studies have suggested that the
number of employees, sales, and assets are all appropriate proxies for ﬁrm size (Harrison
et al. 1988; O'Sullivan and Abela 2007). We operationalize ﬁrm size as sales in our

study.

110

Figure 3: Relationship between Number of Alliances and Innovation Output

Innovation Output

‘l

 

A
7

Number of Alliances

111

4.4 Methodology
4.4.1 Empirical Setting

We choose to empirically test our hypotheses in the pharmaceutical industry.
There are several reasons we prefer the pharmaceutical industry. First, the pharmaceutical
industry is a technology intensive industry in which a ﬁrm’s innovation output plays a
signiﬁcant role in establishing and sustaining a ﬁrm’s competitive advantage (Wuyts et
al. 2004). Second, inter—ﬁrm cooperation in the pharmaceutical industry has been a
popular business practice since the 19803. Most pharmaceutical ﬁrms have entered into
numerous agreements. As our focus is on the relationship between the number of R&D
alliances and innovation output, the large number of alliances in the industry is important,
especially considering we hypothesized different effects on low and high number of
alliances on innovation. This would greatly enhance the testability of our model (Wuyts
et al. 2004).

Third, secondary data are available on all inter-ﬁrm agreements between
pharmaceutical ﬁrms in the United States since 1985 through various sources such as
Verispan, Bioscan, Recap, and IMS. These sources trace in great details of the alliance
activities in the pharmaceutical industry. Finally, pharmaceutical industry has long been a
preferred context for empirically testing innovation and alliance related topics (e. g.,
Nicholls-Nixon and Woo 2003; Prabhu et al. 2005; Rothaerrnel and Deeds 2004; Sorescu
et al. 2003; Wuyts et al. 2004).

4.4.2 Sample
Our sample consists of 66 public pharmaceutical companies in the United Sates.

We limited our sample to public ﬁrms because we need to obtain ﬁnancial measure such

112

as sales as our control variable. Public ﬁrms are mandated to disclose this information to
their shareholder in their annual reports. We compiled data of R&D alliance agreements,
new drug approval, and ﬁnancial information from 1991 to 2005. The starting point of
1991 is selected based on the fact thatFDA did not provide detailed information about
newly approved drugs before then.

The 66 pharmaceutical ﬁrms include most of the major players in the market such
as Pﬁzer, GlaxoSmithKlein, Eli Lilly, and Merck. A total of 3828 R&D alliances were
compiled into our data base for these 66 ﬁrms over the 15 year period. As expected, ﬁrms
demonstrated considerable variation in terms of their R&D alliance portfolio size. Some
ﬁrms have as many as 370 R&D alliances in their portfolio while some have zero. The
average R&D alliances portfolio size is 31.

4.4.3 Data Sources

Three separate data sources provided information for our empirical testing:
Recap, FDA, and Compustat.

For the number of R&D alliances, we collect data ﬁom Recombinant Capital
Database. Recombinant Capital is a consulting ﬁrm that specializes in pharmaceutical
and biotechnology alliances based in the San Francisco Bay Area. They use multiple
sources such as trade literature, press releases, and annual reports to compile a database
of all the alliances activities of pharma/biotech companies. It also provides detailed
information about each identiﬁed alliance activity such as parties to the agreement, the
nature of the agreement, development stage when alliance is formed, and the technologies

that the agreement covers. This database has been used in the literature to construct

alliance related variables (Wuyts et al. 2004).

113

We collected data on new drugs from the drug approval list of the Food and Drug
Administration (FDA) as our innovation output. FDA provides all the approved new
drugs weekly since 1991.

Finally, we collected control variables (ﬁrm size and R&D expenses) from the
Compustat database.

4.5 Analysis and Findings

4.5.] Model Specification

A close examination of our data set suggests that there are several characteristics
of the data that may require special consideration to obtain robust results.

First of all, we collected data on a ﬁrm-year basis and obtained a panel data (66
ﬁrms across 15 years of time). For econometric analysis of panel data, assuming the
observations are independently distributed over time is unfounded (Wooldridge 2003).
Our model speciﬁcations need to account for the heterogeneity caused by the unobserved
ﬁrm-speciﬁc effects. Unobserved factors other than those explicitly addressed in our
model could also bear some inﬂuence on the dependent variables (Prabhu et al. 2005). In
order to truthfully assess the inﬂuence of our hypothesized variables on the dependent
variable, we need to control for these unobserved effects. Secondly, the dependent
variable, number of innovation is a count variable (nonnegative and integers). Ordinary
least squares is thus inappropriate for count data (Sorescu et a1. 2003).

To successfully account for the heterogeneity and count variable problems, we
formulate a Poisson Random Effects model. Using a random effect model over a ﬁxed
effects model is consistent with practices in the literature concerning alliances and

innovation related topics (Prabhu et al. 2005; Sorescu et al. 2003).

114

As such, we test the hypotheses with the following cubic random effect Poisson

regression model.

INNO VA TION, = ,60 + ,8, ALLIANCES i, + ,6, ALLIANCES; + ,83ALLIANCES3
+ ,B4SALES,, + ,BSR & D, + {i + 0,,

Where 4’, is unobserved ﬁrm-speciﬁc effect

0,, is the remaining error term

To control for the multicollinearity problem associated with a cubic regression
model, we used orthogonal polynomial variables as predictor variables (Homburg et al.
2005). Orthogonal polynomial variables are linear combinations of the simple
polynomials and are pair-wise uncorrelated which completely eliminates the problem of
multicollinearity.

4.5.2 Statistical Results

In order to compare our model with existing models in the literature, we estimated
four different models. We present the statistical results in Table 9. Model 1 is the base
model in which only control variables are included. Model 2 is the prevalent linear
relationship model in which the original number of alliances is entered as a predicator.
Model 3 is the inverted U-shape model with the square term added. Model 4 is our
proposed inverted s-shaped model with the cubic term added as a predictor.

We used model 4 to test our hypotheses. The coefﬁcient for the cubic term is
statistically signiﬁcant (B = 0.09, p<0.01). More importantly, it is positive and
signiﬁcant. This supports our model which indicates that the marginal returns decreases
at the beginning but increases at the end showing an inverse s-shaped relationship.

Therefore, our model is supported.

115

In order to compare our model with proposed models in the literature, we
conducted Wald tests on the signiﬁcance of each added variable. All the Wald Chi
squares are statistically signiﬁcant. This proves that the inclusion of the cubic term
signiﬁcantly improves the model ﬁt which supports the advantage and adequacy of our
proposed model. We also compared the ﬁt of the models using Akaike’s Information
Criterion (AIC) of model evaluation (Akaike 1974). The result support the cubic model
over other speciﬁcations as the corresponding AIC value (2897.32) is smaller than that of
all the other three models. We also compared the three models using the Schwartz
Bayesian Information Criterion (Schwarz 1978). The BIC statistics of the four models
also leads to the same conclusion that the cubic model is superior to the other three
models.

Overall, the ﬁndings support our proposed model (both hypotheses). The function
is concave for low to moderate number of alliances and convex for moderate to high
number of alliances. There is an inﬂection point where the function changes from

concave t0 CODVOX.

116

8.9a ._.
~9on “.2.”
.8505 08 356508 woamucwucﬂmcD

 

 

 

 

3.3% wm.wm¢m 3.3% 3.5% Em
2.3% 2; am Sammm 852$ 02
.13: 5.0an 3.38m 3 Nx e35
1%.me 1.322 $523 33.20 mg as?
38.0 See 805550 a
tad. 12.0- essay. 808:2 co a
3.86 3.86 :23 8&5. co a
18.0. o: #8.? 18¢ 05:20me as?
8.0- 33 RS- 23 9% ES
.553 two; ream ......o~.~ 50223

v .352 m .232 N 6.82 ~ .352 833...; Eminence:—

 

2530 5555::— EE moon—«=3. Guam me .395. Z a «Sun.

117

4.6 Discussion
4.6.1 Contribution

Our results strongly support our hypothesized inverse S-Shaped relationship
between a ﬁrm’s number of alliances and its innovation output. This means that when the
alliances numbers are low, the innovation output impact demonstrates the sub-additivity
due to the inexperience in manager’s ability to manage alliances in an effective way and
the limited knowledge base ﬁrms can draw on to integrate the newly acquired knowledge.
However, ﬁrms will start to enjoy an increasing marginal return once they have
established a critical mass of knowledge to have a base for the occurrence of knowledge
integration as well as a better capability in managing alliances more successﬁrlly.

This ﬁnding is consistent with the general conclusions ﬁom previous studies in
that the number of alliances a ﬁrm has is positively related to its innovation output
(Kotabe et al. 1996; Rothaerrnel and Deeds 2004; Shan et al. 1994). Our ﬁndings is also
consistent with the non linearity conclusion drawn by Deeds and Hill (2004). However,
our ﬁndings contribute to the literature by theorizing that the true functional form of the
relationship in not linear, not an inverse U-shaped but in fact follows an inverse S-Shape.

Theoretically, we motivated our study by the combinative capability of ﬁrms. Our
research thus goes beyond the constraints of the simple knowledge accessing foundation
and establishes that ﬁrms actually integrate what they learn from their partners with
existing knowledge base to increase the innovation output. By proving that there are
actually both concave and convex relationships in a ﬁrm’s alliance portfolio building
process in terms of innovation implication, our model provides a basis for ﬁrms to ﬁnd

out the optimality in their alliance activities.

118

Our study supports the idea of developing two important capabilities as a
foundation for a ﬁrm’s competitive advantage: alliances management capability and
combinative capability. In order to shorten the path of the early diminishing marginal
returns, ﬁrms need to increase their learning pace in terms of how to better manage their
alliances and increase their alliance success. Even though an alliance experience is an
important factor in determining the learning process of alliance management, it does not
necessarily need to be a passive learning.

There are several strategies ﬁrms can employ to speed up the learning process that
will help them to reach the increasing marginal retum stage early. First of all, ﬁrms need
to formalize their alliance management learning process. They need to establish a
learning culture in their alliance management process in that the learning not only
involves the learning of technological capabilities from partners, but also involve the
learning of the alliance management itself. This involves the deliberate efforts to
articulate, codify, share, and internalize alliances management know how within the ﬁrm
(Kale and Singh 2007).

Firms need to encourage the sharing of individually held alliance management
know how by establishing practices such as formal and regular debrieﬁng of alliance
manager, establishing internal reports and presentations, or even keep a simple logbook
of all the alliances related activities, actions, and outcomes. Managers need to actively
participate in the process of codifying these tacit knowledge into manual and tools to
have a clear understanding of what works, what does not work, and why. Experiences
from individual alliance cases also need to be shared among managers through informal

mechanism such as daily conversations and discussions or formal mechanisms such as

119

alliance committee or task forces (Kale and Singh 2007). A culture that facilitates the
learning would speed up the capability development of the ﬁrms in terms of the
management of alliance. This would help ﬁrms to ride over the diminishing marginal
returns faster than competitors and realize the increasing marginal returns in an earlier
stage.

In order to increase the learning effectiveness of alliance management,
organizations also need to make organizational changes that accommodate such a
learning culture. Kale, Dyer and Singh (2002) argue that experiences by itself may be a
necessary condition for alliance success but not a sufficient one. They suggest one way
ﬁrms can effectively capture, integrate, and disseminate the alliance-management know
how is to create a separate, dedicated alliance management function. This unit will
coordinate all the alliances related activities and generate higher returns from alliances.
Such a function will facilitate a systematic and routine implementation of the steps
involved in the leaming process (Kale et al. 2002).

A dedicated alliance management function not only becomes the focal point for
codifying, sharing, and internalizing the tacit know how of alliance management, it also
increase the external visibility of the ﬁrm as it would actively interact with potential
alliance partners. As a result, adverse selection problems will be mitigated, as the
increased attractiveness of the focal ﬁrm will lead to a larger pool of alliance candidates.
In addition, such a dedicated also establishes legitimacy to reach and access resources
across division within an organization to support alliances related decision. The function

also motivates ﬁrms to systematically evaluate its alliance performance and create

120

metrics to this evaluation end which will both enhance the learning beneﬁt of the alliance
management (Kale et al. 2002).

Another important capability ﬁrms need to develop is the combinative capability
in integrating knowledge sections. This is an important organizational principle which
deals with the successful structure and coordination of individual and group knowledge
within the ﬁrm (Kale et al. 2002). A ﬁrm’s unique capability to deploy and integrate its
existing and newly assimilated knowledge is fundamental to its long term success and is
difﬁcult t replicate by its competitors (Henderson and Cockbum 1994). In fact, the close
linkage between knowledge accessing and knowledge integrating make the combinative
capability highly desirable to organizations. Henderson and Cockbum (1994) comment
that two forms of integrative competence (combinative capability) are particular
important: the ability to access knowledge from outside the boundary of the organization
and the capability to integrate knowledge ﬂexibly across divisions and units within the
organization.

In order to foster the development of such a capability, ﬁrms need to carefully
coordinate the activities and knowledge sharing across different divisions. Grant (1996)
views the organizational capability as an outcome of knowledge integration and proposed
strategies that would facilitate the emergence of such a capability. Firms need to establish
a centrally located decision maker. Centrally coordinated activities and product designs
will have a positive impact on its product success in the market (Clark and F ujimoto
1990)

The above discussion confers to the notion in the literature that organizational

capability is the foundation for ﬁrms’ long-term competitive advantage. The two

121

capabilities we identiﬁed in our manuscript helps ﬁrms to cut down the concave curve of
alliances’ impact and boost the convex stage of the alliances’ impact on innovation.
Organizations need to establish both an organizational culture and corresponding
organizational structure and process to realize the beneﬁt.
4.6.2 Limitations and Future Research

Our study has several limitations that need to be noted. First of all, the
generalizability of our study needs to be considered as we only sampled publicly traded
pharmaceutical ﬁrms. Even though we only attempt to generalize the ﬁndings to the
technology intensive industries such as telecommunication and electronics, the
applicability of the ﬁndings in other industries needs to be carefully evaluated. This is
particularly a concern when we think about the stringent government regulations imposed
on the innovation measure that we adopted in the study.

The second limitation is that we could not make strong causal inferences, as we
did not take time lag in our model. This is largely due to the nature of our data. We have
a panel data set that includes count variables and a potentially lagged dependent variable.
Econometric techniques in dealing with all three issues involved here (panel data, count
dependent variable, lagged impact) are still in their infancy (Prabhu et al. 2005). Even
though the generalized method of moment has the potential to account for all the unique
characteristics, we did not use it in testing our model because our subject/time ratio is
relatively low compared with what is required for the application of GM. As such, we
could only make inferences in terms of the correlation not causality.

As for future research, we believe that literature could greatly beneﬁt

examinations of certain moderating factors between the number of alliances and

122

innovation output relationship. For example, ﬁrms differ in their capability to combine
knowledge in its possession. Combinative capability develops in an organization as the
ﬁrm learns to manage its knowledge ﬂow more effectively. As the combinative
capabilities differ across ﬁrms, we may expect a moderating effect of the combinative
capability in the proposed inverse s-shape relationship between the number of alliances
and innovation outcome. It is not difﬁcult to see that ﬁrms that have a superior capability
to combine their knowledge elements will obtain more innovation outcomes than those
with lesser capabilities. The logic follows that ﬁrms will be at a better position in
identifying new opportrmities and have better chances of discovering new products by
combining their knowledge due to effective ﬂow knowledge within the organization. As
such, at the same level of the number of alliances, these ﬁrms will result in more
innovation outcomes than others.

A ﬁrm’s product portfolio may also moderate this relationship. The most widely
cited source of synergy in organizations with multiple businesses is the resources
relatedness among business units. Resource relatedness refers to the presence of similar
activities and shared resources across business units of the ﬁrm. Researchers building on
the RBV of diversiﬁcation posit that sharing of strategic resources among business units
creates cross business resource based synergy (Feinberg and Gupta 2004). As such, we
also expect that the inverse S-shaped relationship to be moderated by the product scope a
ﬁrm competing in. The reasons are the following.

First, an organization with multiple business units are more likely to have the
experiences and expertise in effectively moving around human resource with diverse

knowledge stores to better identify new opportunities (Feinberg and Gupta 2004). The

123

reason is that more business units offer greater opportunity and greater experience at
building communication linkages among the various units. These intraﬁnn linkages are
likely to serve as an asset in utilizing the knowledge accumulated by any new R&D
alliances in a particular business unitto the whole organization (F einberg and Gupta
2004)

Second, a more direct reason for the moderating effect of scope of business units
lies in the fact that a ﬁrm with more business units simply has more places within the
organization that may have the capacity to utilize the newly acquired knowledge ﬁ'om its
alliance partner. In addition, these businesses units are likely to have developed the
ability to absorb knowledge from outside through prior interactions with other business
units. As such, we would expect that, holding constant the number of alliances, a ﬁrm
with more business units or product scope would probably come up with more
innovations than a ﬁrm that has a narrower business unit scope. These potential

moderators offer great future research opportunities

124

REFERENCES

Afuah, Allan (2000), "How Much do Your Co-opetitors' Capabilities Matter in the Face
of Technological Change?" Strategic Management Journal, 21 (3), 387.

Ahuja, Gautam (2000a), "Collaboration Networks, Structural Holes, and Innovation: A
Longitudinal Study," Administrative Science Quarterly, 45 (3), 425-55.

---- (2000b), "The Duality of Collaboration: Inducements And Opportunities In The
Formation Of Interﬁnn Linkages," Strategic Management Journal, 21 (3), 317.

Ahuja, Gautam and Curba Morris Lampert (2001), "Entrepreneurship in The Large
Corporation: A Longitudinal Study of How Established Firms Create Breakthrough
Inventions," Strategic Management Journal, 22 (6/7), 521.

Akaike, H. (1974), "A New Look at the Statistical Model Identiﬁcation," in IEEE Trans.
Autom. Control Vol. 19.

Anand, Bharat N. and Tarun Khanna (2000), "Do Firms Learn To Create Value: The
Case Of Alliances," Strategic Management Journal, 21 (3), 295.

Andaleeb, Syed Saad (1992), "The Trust Concept: Research Issues for Channel of
Distribution," Research in Marketing, 11, 1.

Anderson, James C. and James A. Narus (1990), "A Model of Distributor Firm and
Manufacturer F inn Working Partnerships," Journal of Marketing, 54 (1), 42-58.

Arrow, K.J. (1971), Essays in the Theory of Risk Bearing. Chicago: Markham.

Atuahene-Gima, Kwaku (2005), "Resolving the Capability-Rigidity Paradox in New
Product Innovation," Journal of Marketing, 69 (4), 61-83.

Bamford, James and David Ernst (2002), "Managing an Alliance Portfolio," The
McKinsey Quarterly, 3, 29-39.

Barkema, Harry G., John H. J. Bell, and Johannes M. Pennings (1996), "Foreign Entry,
Cultural Barriers, And Learning," Strategic Management Journal, 17 (2), 151-66.

Barney, J. (1991), "Firm Resources and Sustained Competitive Advantage," in Journal of
Management Vol. 17.

Barney, J. B. (2001), "Resource-Based Theories of Competitive Advantage: A Ten-Year
Perspective on the Resource-Based View," in Journal of Management Vol. 27.

125

Baum, Joel A. C., Tony Calabrese, and Brian S. Silvennan (2000), "Don't Go It Alone:
Alliance Network Composition And Startups' Performance In Canadian Biotechnology,"
Strategic Management Journal, 21 (3), 267.

Bettis, R. A. and C. K. Prahalad (1995), "The Dominant Logic: Retrospective and
Extension," Strategic Management Journal, 16 (1), 5-14.

Booz, Allen and Hamilton (1982) in New Product Development in the 19803 B2 - New
Product Development in the 19805. New York: Booz, Allen & Hamilton.

Burt, Ronald (1992), Structural Holes. Cambridge, MA: Harvard University Press.

Chandy, Raj esh K., J aideep C. Prabhu, and Kersi D. Antia (2003), "What Will the Future
Bring? Dominance, Technology Expectations, and Radical Innovation," Journal of
Marketing, 67 (3), 1-18.

Chandy, Raj esh K. and Gerard J. Tellis (2000), "The Incumbent's Curse? Incurnbency,
Size, and Radical Product Innovation," Journal of Marketing, 64(3), 1-17.

---- (1998), "Organizing for Radical Product Innovation: The Overlooked Role of
Willingness to Cannibalize," Journal of Marketing Research (JMR), 35 (4), 474-87.

Chaney, Paul K., Timothy M. Devinney, and Russell S. Winer (1991), "The Impact of
New Product Introductions on the Market Value of F irms," Journal of Business, 64 (4),
573-610.

Chang, Y. and H. Thomas (1989), "The Impact Of Diversiﬁcation Strategy On Risk-
Return Performance," in Strategic Management Journal Vol. 10.

Cho, Hee-Jae and Vladimir Pucik (2005), "Relationship between Innovativeness, Quality,
Growth, Proﬁtability, and Market Value," Strategic Management Journal, 26, 555-75.

Chung, Seungwha, Harbir Singh, and Lee Kyungmook (2000), "Complementarity, Status
Similarity and Social Capital as Drivers of Alliance Formation," Strategic Management
Journal, 21 (1), 1.

Clark, Kim B. and Takahiro Fujimoto (1990), "The Power of Product Integrity. ,"
Harvard Business Review, 68 (6), 107-18.

Cohen, Morris A. and J ehoshua Eliashberg (1997), "An Anatomy of a Decision-Support
System for Developing and Launching Line Extensions," Journal of Marketing Research
(JMR), 34(1), 117-29.

Cohen, Wesley M. and Daniel A. Levinthal (1990), "Absorptive Capacity: A New
Perspective on Learning and Innovation," Administrative Science Quarterly, 35 (1), 128-
52.

126

---- (1989), "innovation and Learning: The Two Faces of R&D," Ecomonic Journal 99
(397), 569-96.

Coleman, James S. (1993), "The Rational Reconstruction of Society," American
Sociological Review, 58 (1), 1-15.

Cooper, Lee G. (2000), "Strategic Marketing Planning for Radically New Products,"
Journal of Marketing, 64(1), 1-16.

Damanpour, F. and W. M. Evan (1984), "Organizational Innovation and Performancezthe
Problem of "Organizational Lag"," Administrative Science Quarterly, 29(3), 392—409.

Das, T. K. and Teng Bing-Sheng (2000), "A Resource-Based Theory of Strategic
Alliances," Journal of Management, 26 (1), 31-62.

Dewar, Robert D. and Jane E. Dutton (1986), "The Adoption of Radical and Incremental
Innovations: An Empirical Analysis," Management Science, 32 (11), 1422-33.

Dougherty, Deborah and Cynthia Hardy (1996), "Sustained Product Innovation in Large,
Mature Organizations," Academy of Management Journal, 39 (5), 1120.

Doz, Y. L. (1998), Alliance AdvantagezThe Art of Creating Value through Partnering.
Boston: harvard Business School Press.

Doz, Yves L. (1996), "The Evolution of Cooperation in Strategic Alliances: Initial
Conditions or Learning Processes?," Strategic Management Journal, 17 (7), 55-83.

Dussauge, Pierre, Bernard Garrette, and Will Mitchell (2000), "Learning from Competing
Partners: Outcomes and Durations of Scale and Link Alliances in Europe," Strategic
Management Journal, 21 (2), 99.

Dyer, Jeffery H. and Harbir Singh (1998), "The Relational View: Cooperative Strategy
and Sources of Interorganizational Competitive Advantage," Academy of Management
Review, 23 (4), 660-79.

Dyer, Jeffrey H. (1997), "Effective Interﬁrm Collaboration: How Firms Minimize
Transaction Costs And Maximize Transaction Value," Strategic Management Journal, 18
(7), 535-56.

Dyer, Jeffrey H. and Kentaro Nobeoka (2000), "Creating And Managing A High-
Perforrnance Knowledge—Sharing Network: The Toyota Case," Strategic Management
Journal, 21 (3), 345.

127

Eisenhardt, Kathleen M. and Claudia Bird Schoonhoven (1996), "Resource-based View
of Strategic Alliance Formation: Strategic and Social Effects in Entrepreneurial F inns,"
Organization Science, 7 (2), 136.

F einberg, Susan E. and Anil K. Gupta (2004), "Knowledge Spillovers and the
Assignment of R&D Responsibility to Foreign Subsidiaries," Strategic Management
Journal, 25 (8), 823-45.

Freeman, Linton C. (1979), "Centrality in Social Networks: Conceptual Clariﬁcation,"
Social Networks, 1, 215-39.

---- (1992), "The Sociological Concept of ’Group': An Empirical Test of Two Models,"
American Journal of Sociology, 98, 152-66.

Geroski, Paul, Steve Machin, and John Van Reenen (1993), "The Proﬁtability of
Innovating Firms," RAND Journal of Economics, 24 (2), 198-212.

Glaister, Keith W. and Peter J. Buckley (1996), "Strategic Motives For International
Alliance Formation," Journal of Management Studies, 33 (3), 301-32.

Goerzen, Anthony (2007), "Alliance Networks and F irm Performance: The Impact of
Repeated Partnerships," Strategic Management Journal, 28 (5), 487-509.

Goerzen, Anthony and Paul W. Bearnish (2005), "The Effect of Alliance Network
Diversity on Multinational Enterprise Performance," Strategic Management Journal, 26
(4), 333-54.

Grant, Robert M. (1997), "The Knowledge-based View of the Firm: Implications for
Management Practice," Long Range Planning, 30 (3), 450-54.

---- (1996), "Toward A Knowledge-Based Theory Of The Firm," Strategic Management
Journal, 17, 109-22.

Grant, Robert M. and Charles Baden-Fuller (1995), "A Knowledge-Based Theory Of
Inter-Firm Collaboration," Academy of Management Proceedings, 17-21.

---- (2004), "A Knowledge Accessing Theory of Strategic Alliances," Journal of
Management Studies, 41 (1), 61-84.

Gulati, Ranjay (1998), "Alliances and Networks," Strategic Management Journal, 19 (4),
293.

---- (1995), "Does Familiarity Breed Trust? The Implications of Repeated Ties for
Contractual Choice in Alliances," Academy of Management Journal, 38 (1), 85.

128

---- (1999), "Network Location and Learning: The Inﬂuence of Network Resources and
Firm Capabilities " Strategic Management Journal, 20 (5), 397.

Gulati, Ranj ay and Martin Gargiulo (1999), "Where do Interorganizational Networks
Come From?," American Journal of Sociology, 104 (5), 1439-93.

Gulati, Ranj ay, Nitin Nohria, and Akbar Zaheer (2000), "Guest Editors' Introduction To
The Special Issue: Strategic Networks," in Strategic Management Journal. 21 ed.

Hamel, Gary (1991), "Competition For Competence And Inter-Partner Learning Within
International Strategic Alliances
" Strategic Management Journal, 12(4), 83-103.

Hansen, Morton T. (1999), "The Search-Transfer Problem: The Role of Weak Ties in
Sharing Knowledge Across Organization Subunits," Administrative Science Quarterly,
44(1), 82-111.

Hargadon, Andrew and Robert I. Sutton (1997), "Technology Brokering and Innovation
in a Product Development Firm," Administrative Science Quarterly, 42 (4), 716-49.

Harrison, J. Richard, David L. Torres, and Sal Kukalis (1988), "The Changing of the
Guard: Turnover and Structural Change in the Top-Management Positions,"
Administrative Science Quarterly, 33 (2), 211.

Harrison, Jeffrey 8., Michael A. Hitt, Robert E. Hoskisson, and R. Duane Ireland (2001),
"Resource Complementarity in Business Combinations: Extending the Logic to
Organizational Alliances," Journal of Management, 27 (6), 679.

---- (1991), "Synergies and Post-Acquisition Performance: Differences versus Similarities
in Resource Allocations," Journal of Management, 17 (l), 173.

Hemphill, Thomas and Nicholas Vonortas (2003), "Strategic Research Partnerships: A
Managerial Perspective," Technology Analysis & Strategic Management, 15 (2), 255.

Henderson, Rebecca (1994), "Managing Innovation in the Information Age," Harvard
Business Review, 72 (3), 100-05.

Henderson, Rebecca and Iain Cockbum (1994), "Measuring Competence? Exploring ﬁrm
Effects in Pharmaceutical Research," Strategic Management Journal, 15 (8), 63-84.

Hitt, Michael A., M. Tina Dacin, Edward Levitas, J ean-Luc Arregle, and Anca Borza
(2000), "Partner Selection In Emerging And Developed Market Contexts: Resource-
Based And Organizational Learning Perspectives," Academy of Management Journal, 43
(3), 449-67..

129

Hitt, Michael A., R. Duane Ireland, S. Michael Camp, and Donald L. Sexton (2001),
"Guest Editors' Introduction To The Special Issue Strategic Entrepreneurship:
Entrepreneurial," Strategic Management Journal, 22 (6/7), 479.

Hoang, Ha and Frank T. Rothaerrnel (2005), "The Effect of General and Partner-Speciﬁc
Alliance Experience on Joint R&D Project Performance," Academy of Management
Journal, 48 (2), 332-45.

Hofﬁnann, Werner H. (2007), "Strategies for Managing a Portfolio of Alliances,"
Strategic Management Journal, in press.

Homburg, Christian, Nicole Koschate, and Wayne D. Hoyer (2005), "Do Satisﬁed
Customers Really Pay More? A Study of the Relationship Between Customer
Satisfaction and Willingness to Pay," Journal of Marketing, 69 (2), 84-96.

Inkpen, Andrew C. and Mary M. Crossan (1995), "Believing Is Seeing: Joint Ventures
And Organization Learning," Journal of Management Studies, 32 (5), 595-618.

Inkpen, Andrew C. and Adva Dinur (1998), "Knowledge Management Processes and
International Joint Ventures," Organization Science, 9 (4), 454.

Ireland, R. Duane, Michael A. Hitt, and Deepa Vaidyanath (2002), "Alliance
Management as a Source of Competitive Advantage," Journal of Management, 28 (3),
413-46.

Kale, Prashant, Jeffrey H. Dyer, and Harbir Singh (2002), "Alliance Capability, Stock
Market Response, and Long Term Alliance Success: The Role of the Alliance Function,"
Strategic Management Journal, 23 (8), 747.

Kale, Prashant and Harbir Singh (2007), "Building Firm Capabilities through Learning:
the Role of the Alliance Learning Process in Alliance Capability and F irm-Level Alliance
success," Strategic Management Journal, in press.

---- (2000), "Learning And Protection Of Proprietary Assets In Strategic Alliances:
Building Relational Capital," Strategic Management Journal, 21 (3), 217.

Khanna, Tarun (1998), "The Scope of Alliances," Organization Science, 9 (3), 340-55.
Kharma, Tarun, Ranjay Gulati, and Nitin Nohria (1998), "The dynamics of learning
alliances: Competition, cooperation, and relative scope," Strategic Management Journal,

19(3), 1193.

Kim, W. C. and R. Mauborgne (1997), "Value Innovation: the Strategic Logic of High
Growth," Harvard Business Review, 75 (1), 102-12.

130

Kleinschmidt, E. J. and R. G. Cooper (1991), "The Impact of Product Innovativeness on
Performance," Journal of Product Innovation Management, 8 (4), 240-51.

Kogut, Bruce and Udo Zander (1992), "Knowledge of the Firm, Combinative
Capabilities, and the Replication of Technology," Organization Science, 3 (3), 383-97.

Kogut, Bruce and Udo Znader (1996), "What Firms Do? Coordinationa, Identity, and
Learning," Organization Science, 7, 502-18.

Koka, Balaj i R. and John E. Prescott (2002), "Strategic Alliances as Social Capital: A
Multidimensional View," Strategic Management Journal, 23 (9), 795.

Kor, Y. and J. T. Mahoney (2004), "Edith Penrose's (1959) contributions to the resource-
based view of strategic management," in Journal of Management Studies Vol. 41.

Kotabe, Masaaki, Arvind Sahay, and Preet S. Aulakh (1996), "Emerging role of
technology licensing in the development," Journal of Marketing, 60 (1), 73.

Kraatz, Matthew S. (1998), "Learning by Association? Interorganizational Networks and
Adaptation to Environmental Change," Academy of Management Journal, 41 (6), 621-43.

Lam, Alice (1997), "Embedded F irms, Embedded Knowledge: Problems of Collaboration
and Knowledge Transfer in Global Cooperative Ventures," Organization Studies (Walter
de Gruyter GmbH & Co. KC), 18 (6), 973.

Lampel, Joseph and Jamal Shamsie (2000), "Probing the Unobtrusive Link: Dominant
Logic and the Design of joint Ventures at General Electric," Strategic Management
Journal, 21 (5), 593.

Lane, P. J. and M. Lubatkin (1998), "Relative Absorptive Capacity and
Interorganizational Learning," Strategic Management Journal, 19 (5), 461-77.

Lazzarini, Sergio G. (2007), "The Impact of Membership in Competing Alliance
Constellations: Evidence on the Operational PErformance of Global Airlines," Strategic
Management Journal, 28, 345-67.

Lorenzoni, Gianni and Andrea Lipparini (1999), "The Leveraging of Interﬁrm
Relationships as a Distinctive Organizational Capability: A longitudinial Study,"
Strategic Management Journal, 20 (4), 317.

Lyles, Marjorie A. and Jane E. Salk (2007), "Knowledge acquisition from foreign parents

in international joint ventures: an empirical examination in the Hungarian context,"
Journal of International Business Studies, 38 (1), 3-18.

131

Macher, Jeffrey T. and Christopher S. Boemer (2006), "Experience and Scale and Scope
Economies: Trade-offs and Performance in Development," Strategic Management
Journal, 27 (9), 845-65.

Madhavan, Ravindranath and Rajiv Grover (1998), "From Embedded Knowledge to
Embodied Knowledge: New Product Development as Knowledge Management," Journal
of Marketing, 62 (4), 1-12.

Madhok, Anoop and Stephen B. Talhnan (1998), "Resources, Transactions and Rents:
Managing Value Through Interﬁrm Collaborative Relationships," Organization Science,
9 (3), 326.

Makhija, Mona V. and Usha Ganesh (1997), "The Relationship Between Control and
Partner Leaming-related Joint Ventures," Organization Science, 8 (5), 508.

March, James G. (1991), "Exploration And Exploitation In Organizational Learning,"
Organization Science, 2 (1), 71.

McGee, Jeffrey B, Michael J. Dowling, and William L. Megginson (1995), "Cooperative
Strategy And New Venture Performance: The Role Of Business Strategy And
Management Experience," Strategic Management Journal, 16 (7), 565-80.

McGrath, Rita Gunther, Ming-Hone Tsai, S. Venkataraman, and I. C. MacMillan (1996),
"Innovation, Competitive Advantage and Rent: A Model and Test," Management
Science, 42 (3), 389.

Miles, Morgan P., Jeffrey G. Covin, and Michael B. Heeley (2000), "The Relationship
Between Environmental Dynamism And Small Firm Structure, Strategy, And
Performance," Journal of Marketing Theory & Practice, 8 (2), 63.

Mohr, J akki J ., Robert J. Fisher, and John R. Nevin (1996), "Collaborative
communication in interﬁrrn relationships: Moderating effects of integration and," Journal
of Marketing, 60(3), 103.

Mowery, David C., Joanne E. Oxley, and Brian S. Silverrnan (1996), "Strategic Alliances
And Interfrrrn Knowledge Transfer," Strategic Management Journal, 17, 77-91.

Nakamura, M. , J. M. Shaver, and B. Yeung (1996), "An Empirical Investigation of joint
Venture Dynamics: Evidence from US-Japan Joint Ventures," International Journal of
Industrial Organization, 14, 521-41.

Nerkar, Am] and Peter W. Roberts (2004), "Technological And Product-Market

Experience And The Success Of New Product Introductions In The Pharmaceutical
Industry," Strategic Management Journal, 25 (8/9), 779-99.

132

Nicholls-Nixon, Charlene L. and Carolyn Y. Woo (2003), "Technology Sourcing And
Output Of Established Firms In A Regime Of Encompassing Technological Change,"
Strategic Management Journal, 24 (7), 651.

Nonaka, Ikujiro (1994), "A Dynamic Theory of Organizational Knowledge Creation,"
Organization Science, 5 (1), 14.

O'Sullivan, Don and Andrew V. Abela (2007), "Marketing Performance Measurement
Ability and F inn Performance," Journal of Marketing, 71 (2), 79-93.

Oxley, Joanne E. (1997), "Appropriability Hazards and Governance in Strategic
Alliances: A Transaction Cost Approach," The Journal of Law, Economics, &
Organization, 13 (2), 387-409.

Oxley, Joanne E. and Rachelle C. Sampson (2004), "The Scope And Governance Of
International R&D Alliances," Strategic Management Journal, 25 (8/9), 723-49.

Page, Albert L. (1993), "Assessing New Product Development Practices and
Performance: Establishing Crucial Norms," Journal of Product Innovation Management,
10(4), 273-90.

Parkhe, Arvind (1991), "Interﬁrm Diversity, Organizational Learnings, And Longevity In
Global Strategic Alliances," Journal of International Business Studies, 22 (4), 579-601.

---- (1993), "Strategic alliance structuring: A game theoretic and transaction cost
examination of interﬁnn," Academy of Management Journal, 36 (4), 794.

Pelled, L. H., K. M. Eisenhardt, and K. R. Xin (1999), "Exploring the Black Box: An
Analysis of Work Group Diversity, Conﬂict, and Performance," Administrative Science
Quarterly, 44(1), 1-28.

Penrose, E. T. (1959) in The Theory of the Growth of the F inn B2 - The Theory of the
Growth of the Firm. Oxford: Blackwell.

Pisano, Gary P., Richard M. J. Bohmer, and Amy C. Edmondson (2001), "Organizational
Differences in Rates of Learning: Evidence from the Adoption of Minimally Invasive
Cardiac Surgery," Management Science, 47 (6), 752.

Polanyi, Michael (1966), The Tacit Dimension. New York: Anchor Day Books.

Powell, Walter W., Kenneth W. Koput, and Laurel Smith-Doerr (1996),
"Interorganizational Collaboration and the Locus of Innovation: Networks of Learning in
Biotechnology," Administrative Science Quarterly, 41 (1), 116-45.

133

Prabhu, J aideep C., Rajesh K. Chandy, and Mark E. Ellis (2005), "The Impact of
Acquisitions on Innovation: Poison Pill, Placebo, or Tonic?," Journal of Marketing, 69
(1), 114-30.

Reid, Douglas, David Bussiere, and Kathleen Greenaway (2001), "Alliance Formation
Issues for Knowledge-based Enterprises," International Journal of Management Reviews,
3 (1), 79.

Reuer, Jeffrey J. and Aﬁica Arino (2007), "Strategic Alliance Contracts: Dimensions and
Determinants of Contractual Complexity," Strategic Management Journal, 28 (3), 313-30.

Reuer, Jeffrey J. and Roberto Ragozzino (2006), "Agency hazards and alliance
portfolios," Strategic Management Journal, 27 (1), 27-43.

Reuer, Jeffrey J ., M. 20110, and Harbir Singh (2002), "Post-forrnation Dynamics in
Strategic Alliances," Strategic Management Journal, 23 (2), 135-51.

Rindﬂeisch, Aric and Jan B. Heide (1997), "Transaction cost analysis: Past, present, and
future applications," Journal of Marketing, 61 (4), 30.

Rindﬂeisch, Aric and Christine Moorman (2001), "The Acquisition and Utilization of
Information in New Product Alliances: A Strength-of-Ties Perspective," Journal of
Marketing, 65 (2), 1-18.

---- (2003), "Interﬁrm Cooperation and Customer Orientation," Journal of Marketing
Research (JMR), 40(4), 421-36.

Ring, P. S. and A. H. Van de Ven (1994), "Developmental Processes of Cooperative
Interorganizational Relationships " Academy of Management Review, 19, 90-118.

Rodan, Simon and Charles Galrmic (2004), "More Than Network Structure: How
Knowledge Heterogeneity Inﬂuences Managerial Performance and Innovativeness,"
Strategic Management Journal, 25, 541-62.

Rothaerrnel, Frank T. and David L. Deeds (2004), "Exploration And Exploitation
Alliances In Biotechnology: A System Of New Product Development," Strategic
Management Journal, 25 (3), 201-21.

Rothaennel, Frank T., Michael A. Hitt, and Lloyd A. Jobe (2006), "Balancing VErtical
Integration and Strategic Outsourcing: Effects on Product Protfolio, Product Success, and
Firm Performance," Strategic Management Journal, 27, 1033-56.

Sampson, Rachelle C. (2005), "Experience Effects and Collaborative Returns in R&D
Alliances," Strategic Management Journal, 26 (11), 1009-31.

134

Santoro, Michael D. and Joseph P. McGill (2005), "The Effect of Uncertainty and Asset
Co-specialization on Governance in Biotechnology Alliances," Strategic Management
Journal, 26 (13), 1261-69.

Schmidt, Jeffrey, B. and Roger Calantone, J. (2002), "Escalation of Commitment During
New Product Development," Journal of the Academy of Marketing Science, 30 (2), 103.

Schumpeter, Joseph (1942), CapitaliSm, Socialism, and Democracy. New York Harper.

Schwarz, G. (1978), "Estimating the dimension of a model," in The Annals of Statistics
Vol. 6.

Shan, Weij an, Gordon Walker, and Bruce Kogut (1994), "Interﬁrm Cooperation and
Startup Innovation in the Biotechnology Industry," Strategic Management Journal, 15,
387-94.

Simonin, Bernard L. (1999), "Ambiguity and the process of knowledge transfer in
strategic alliances," Strategic Management Journal, 20(7), 595-623.

---- (1997), "The importance of collaborative know-how: An empirical test of the
learning organization," Academy of Management Journal, 40 (5), 1150.

Sivadas, Eugene and F. Robert Dwyer (2000), "An Examination of Organizational
Factors Inﬂuencing New Product Success in Internal and Alliance-Based Processes,"
Journal of Marketing, 64 (1), 31-49.

Slowinski, Gene (2001), "Becoming the Partner of Choice," Pharmaceutical Executive,
21 (12), 52-56.

Sood, Ashish and Gerard J. Tellis (2005), "Technological Evolution and Radical
Innovation," Journal of Marketing, 69 (3), 152-68.

Sorescu, Alina B., Raj esh K. Chandy, and Jaideep C. Prabhu (2003), "Sources and
Financial Consequences of Radical Innovation: Insights from Pharmaceuticals," Journal
of Marketing, 67 (4), 82-102.

Spender, J. C. (1992), "Limits to Learning from the West: How Western Management
Advice May Prove Limited in Eastern Europe," International Executive, 34 (5), 389-413.

Stuart, Toby E. (2000), "Interorganizational Alliances And The Performance Of F inns: A
Study Of Growth And Innovation," Strategic Management Journal, 21 (8), 791.

Subramanian, A. and S. Nilakanta (1996), "Organizational Innovativeness: Exploring the

Relationship between Organizational Determinants of Innovation, Type of Irmovations,
and Measures of Organizational Performance," Omega, 24 (6), 631-47.

135

Teece, David J. (1992), "Foreign Investment and Technological Development in Silicon
Valley," California Management Review, 34 (2), 88.

Tiwana, Amrit and Mark Keil (2007), "Does Peripheral Knowledge Complement
Control? An Empirical Test in Technology Outsourcing Alliances," Strategic
Management Journal, 28 (6), 623-34.

Tsai, Wenpin and Sumantra Ghoshal (1998), "Social Capital And Value Creation: The
Role of Intraﬁrrn Networks," Academy of Management Journal, 41 (4), 464-76.

Uzzi, Brian (1997), "Social Structure and Competition in Interﬁrm Network: The
Paradox of Embeddedness," Administrative Science Quarterly, 42 (1), 35-67.

Vives, X. (1990), "Trade Association Disclosure Rules, Incentives To Share Information,
And Welfare," in RAND Journal of Economics Vol. 21.

Volberda, Henk W. (1996), "Toward the Flexible Form: How to Remain Vital in
Hypercompetitive Environments," Organization Science, 7 (J uly/August), 359-74.

von Hippel, E. (1987), "Cooperation Between Rivals: Informal Know-how Trading," in
Research Policy Vol. 16.

Walker, Gordon, Bruce Kogut, and Shan Wijian (1997), "Social Capital, Structural Holes
and the Formation of an Industry Network," Organization Science, 8 (2), 109.

Wemerfelt, Birger (1984), "A Resource-based View of the F inn," Strategic Management
Journal, 5 (2), 171-80.

White, Steven and Steven Siu—Yun Lui (2005), "Distinguishing Costs of Cooperation and
Control in Alliances," Strategic Management Journal, 26 (10), 913-32.

Wiklund, Johan and Dean Shepherd (2003), "Knowledge-Based Resources,
Entrepreneurial Orientation, And The Performance Of Small And Medium-Sized
Businesses," Strategic Management Journal, 24 (13), 1307-14.

Williamson, Oliver E. (1975) in Markets and Hierarchies: Analysis and Antitrust
Implications B2 — Markets and Hierarchies: Analysis and Antitrust Implications. New
York: The Free Press.

---- (1991), "Comparative Economic Organization: The Analysis of Discrete Structural
Alternatives," Administrative Science Quarterly, 36 (2), 269-96.

---- (1994), "Transaction Cost Economics and Organization Theory," in The Handbook of

Economic Sociology B2 - The Handbook of Economic Sociology, N. J. Smelser and R.
Swedberg (Eds). Princeton, NJ: Princeton University Press.

136

---- (1998), "Transaction Cost Economics: How It Works: Where It Is Headed," De
Economist (0013-063X), 146 (l), 23.

Wind, Jerry and Vij ay Mahajan (1997), "Issues and Opportunities in New Product
Development: An Introduction to the Special Issue," in Journal of Marketing Research

(JMR). 34 ed.
Wooldridge, Jeffery M. (2003), Introductory Econometrics. Mason, OH: South-Westem.

Wuyts, Stefan, Stefan Stremersch, and Shantanu Dutta (2004), "Portfolios of Interﬁrm
Agreements in Technology-Intensive Markets: Consequences for Innovation and
Proﬁtability," Journal of Marketing, 68 (2), 88-100.

Yli-Renko, Helena, Erkko Autio, and. Harry J. Sapienza (2001), "Social Capital,
Knowledge Acquisition, And Knowledge Exploitation In Young Technology-Based
F irms," Strategic Management Journal, 22 (6/7), 587.

 

Young-Ybarra, Candace and Margarethe Wiersema (1999), "Strategic Flexibility in
Information Technology Alliances: The Inﬂuence of Transaction Cost Economics and
Social Exchange Theory," Organization Science, 10(4), 439.

Zahra, Shaker A., R. Duane Ireland, and Michael A. Hitt (2000), "International
Expansion By New Venture Firms: International Diversity, Mode Of Market Entry,
Technological Learning, And Performance," Academy of Management Journal, 43 (5),
925-50.

Zeng, Ming and Xiao-Ping Chen (2003), "Achieving Cooperation In Multiparty
Alliances: A Social Dilemma Approach To Partnership Management," Academy of
Management Review, 28 (4), 587-605.

Zirger, Billie Jo and Modesto A. Maidique (1990), "A Model of New Product
Development: An Empirical Test," Management Science, 36 (7), 867-83.

Zucker, Lynne G., Michael R. Darby, and Jeff S. Armstrong (2002), "Commercializing
Knowledge: University Science, Knowledge Capture, and F inn Performance n
Biotechnology," Management Science, 48 (1), 138-53.

137

L

llljlljll

0

       

ulrlrrjrrjrrrr