University of Southern California Dissertations and Theses

The role of capabilities in new alliance creation and performance: a study of the biotechnology industry

(USC Thesis Other)

The role of capabilities in new alliance creation and performance: a study of the biotechnology industry

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

THE ROLE OF CAPABILITIES IN NEW ALLIANCE CREATION AND
PERFORMANCE: A STUDY OF THE BIOTECHNOLOGY INDUSTRY

by
Kimberlie J. Stephens

_____________________________________________________________________

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(COMMUNICATION)

August 2009

Copyright 2009 Kimberlie J. Stephens

ii
Acknowledgements
This dissertation would not have been possible without the help and support of many
colleagues, friends and family members. I would specifically like to thank my advisor,
Janet Fulk and my committee members Peter Monge and Kyle Mayer for valuable
feedback on early drafts of this dissertation. I would also like to thank several friends
for their endless encouragement and cheerleading in the process, including Jennifer
Heckman, Becca Wriedt, Dawn DenHartog and many, many others along the way. I
could not have done this without the help of my family who have taught me the value
of learning and perseverance; specifically, my parents, Dave and Ronda; my
grandparents, Donald and Elizabeth; and my siblings, Shannon, Mary and Matthew.
Finally, a very special thanks goes to the one without whom I would surely have given
up, my husband, Michael.
iii
Table of Contents
Acknowledgements ii
List of Tables iv
List of Figures v
Abstract vi
Chapter 1: Introduction 1
Chapter 2: Literature Review 6
Chapter 3: The Biotechnology Industry 31
Chapter 4: Methods and Analysis 43
Chapter 5: Results 57
Chapter 6: Discussion and Conclusion 80
References 96
Appendix A 107

iv
List of Tables
Table 2.1: Hypothesis Summary 30
Table 4.1: Deal Type Definitions and Categorizations 47
Table 4.2: Variable Summary 53
Table 5.1: DealType * Year Crosstabulation 57
Table 5.2: Descriptive Statistics for Full Multinomial Logistic 59
Regression Model

Table 5.3: Descriptive Statistics for Multinomial Logistic 60
Regression for Alliance between Past Partners

Table 5.4: Estimates (log-odds and odds) from selected multinomial 62
logistic regression models

Table 5.5: Estimates (log-odds and odds) from selected multinomial 68
logistic regression models for those alliances where partner
experience is one or more

Table 5.6: Event Study Results 70
Table 5.7: Summary of Excess Returns Across Event Windows 70
Table 5.8a: Descriptive Statistics for Ordinary Least Squares Models 72
Table 5.8b: Cross-sectional analysis of alliance performance 74
Table 5.9a: Descriptive Statistics for Ordinary Least Squares Models 77
for Deals where Firms were Partners in the Past

Table 5.9b: Cross-sectional analysis of alliance performance for alliances 78
where firms were partners in the past

Table 5.10: Hypothesis results summary 79

v
List of Figures

Figure 3.1: New Biotech Drug and Vaccine Approvals/ 34
New Indication Approvals by Year

Figure 3.2: Biotechnology Industry Sales and Revenue by Year 35
Figure 3.3: Biotechnology and Pharmaceutical Stock Indexes 37
vi
Abstract
Strategic alliances, or interorganizational relationships, are a prevalent strategy
used to achieve organizational goals. Along with the increasing prevalence of
alliances has come a heightened need to understand the dynamics of these
relationships, as approximately half of the alliances entered fail (Kale, Dyer & Singh,
2002). This research, therefore, seeks to expand on our current understanding of
alliance capability development by exploring the effects of specific types of alliance
and technological experience as well as their interactions on new alliance formation
and performance. Using the resource based view of the firm as a theoretical
framework, this research finds that firms are significantly more likely to enter into
alliance types, and with specific alliance partners, with which they have past
experience. The RBV notion that firm resources will be utilized in strategic decisions
is supported by this finding. Experience in a particular product area was also found to
increase the likelihood of firms entering into non-R&D type relationships suggesting
the use of alliance relationships to capitalize on or seek to exploit product specific
resources in the alliance context. Alliance experience in general was found to
positively influence the performance of alliance relationships, while specific types of
alliance experience were only found to influence performance when entering into that
type of relationship. In this way, firms seem to develop distinct capabilities in specific
alliance types that are not necessarily easily transferred from one relationship type to
another.

1
Chapter 1: Introduction
Strategic alliances, or interorganizational relationships, are a prevalent strategy
used to achieve organizational goals. Many scholars have attributed the alliance’s rise
in popularity to the increase in competition due to globalization (Castells, 2000), and
to the necessity of gaining access to resources required to create competitive
advantage (Teece, 1986). Several benefits are thought to accrue to firms that enter
alliances, such as opportunities for organizational learning (Child, 2001; Simonin,
1999), uncertainty reduction (Podolny, 1994; Provan, 1982), organizational legitimacy
or status (Chung, Singh & Lee, 2000; Li & Berta, 2002) and interorganizational
interaction experience (Gulati, 1995a; 1998). Along with the increasing prevalence of
alliances has come a heightened need to understand the dynamics of these
relationships, as approximately half of the alliances entered fail (Kale, Dyer & Singh,
2002). It is important to note that this number may be inflated due to the fact that
some relationships counted as failures may actually be termination of successfully
completed relationships (Kogut, 1989; Duysters, Kok & Vaandrager, 1999).
Regardless of this fact, however, many scholars agree that most alliance failures are
the result of unintentional problems (Kale et al, 2002). As such, research efforts have
begun to focus on exploring what factors may contribute to an alliance capability
(Kale & Singh, 2007), and how relationships evolve with the goal of understanding
how firms can learn to consistently and successfully execute alliance relationships
(Doz, 1996; Ariño & de la Torre, 1998). Much of the research in this area, however,
looks either at how capabilities relate to the decision to enter into new relationships
(Wang & Zajac, 2007) or how capabilities relate to alliance performance (Anand &
2
Khanna, 2000; Koh & Venkatraman, 1991; Park, Mezias & Song, 2004). This
research addresses both of these aspects of alliance capabilities and seeks to expand on
our current understanding in the area of alliance capability development by exploring
the effects of specific types of alliance and technological experience as well as their
interactions on new alliance formation and performance. As a result, the following
research questions will be explored:
RQ1: What types of firm level capabilities are predictive of the types of
new relationships that firms choose to enter?
RQ2: What types of firm level capabilities have a bearing on the success of
new relationships?
Firm level capabilities refer to those competencies that a specific firm may
have as a result of the ability to execute routines to effectively utilize resources within
the firm (Nelson & Winter, 1982). This notion of differential resource utilization is
the foundation of the resource based view of the firm (Penrose, 1959; Barney, 1991).
One way in which capabilities can be measured is through the extent of experience a
firm has in a specific area, where additional experience allows a firm to learn how to
use necessary resources in an increasingly efficient manner (Wang & Zajac, 2007;
Anand & Khanna, 2000; Rothaermel & Deeds, 2006). This research will test the
assertions of the resource-based view and explore how firm level alliance capability
and technological capability influence the types of new relationships that firms choose
to enter. Specifically alliance capability will be explored in terms of firm-level
experience with different alliance types, including license agreements, non-R&D
agreements, and R&D agreements, as well as the experience a firm has with partner
3
firms. The technological capability will be explored in terms of a firm’s experience
with developing products in general, as well as experience with developing products in
a specific disease or indication category.
The resource-based view of the firm forms the theoretical foundation for this
research (Penrose, 1959; Wernerfelt, 1984; Wang & Zajac, 2007). If firms realize that
resources can form the basis for competitive advantage (Teece, Pisano & Shuen, 1997),
then it follows that utilization of those resources, in this case alliance and
technological capabilities, would be important when firms are considering entering
new relationships (Wang & Zajac, 2007; Mahoney & Pandian, 1992). As a result, the
resource-based view informs the hypotheses development and the results shed light on
the reliability of the resource-based view in terms of how experience influences
alliance decision making and performance outcomes. By looking at how capabilities
relate to the tendency to form new relationships as well as the subsequent performance
of those relationships, this research provides insight into the way in which those
internal to a firm, as well as those external to the firm may view potential firm
resources.
This research finds that firms are significantly more likely to enter into alliance
types, and with specific alliance partners, with which they have past experience. The
RBV notion that firm resources will be utilized in strategic decisions is supported by
this finding. Although product development experience did not have any impact on
the likelihood of a firm choosing a particular alliance type, the technological capability,
product area experience, was found to increase the likelihood of firms entering into
non-R&D type relationships suggesting the use of alliance relationships to capitalize
4
on or seek to exploit those resources in the alliance context. Alliance experience in
general was found to positively influence the performance of alliance relationships,
while specific types of alliance experience were only found to influence performance
when entering into that type of relationship. In this way, firms seem to develop
distinct capabilities in specific alliance types that are not necessarily easily transferred
from one relationship type to another. Contrary to expectations, partner experience in
general had a negative impact on wealth creation over the event period. This finding,
although somewhat unexpected lends credence to the idea that firms gain legitimacy
and credibility from early relationships as the ability to attract known partners
provides information to the market on an otherwise little known firm (Nicholson,
Danzon & McCullough, 2005; Stuart, Hoang & Hybels, 1999). Especially in the
context of new relationships with repeat partners, the second or third relationship does
not provide the information benefits and might therefore appear to negatively impact
performance. It is unlikely that this effect is related to the firm’s actual alliance
capabilities, but the negative performance effects may be due to the conflation of
information on a firm’s commercial viability with the potential benefits inherent in
any new alliance relationship.
This research supports not only the idea that firms make strategic decisions
based on existing capabilities, but that those capabilities are also often related to the
final performance of the specific business process. In addition, these findings
demonstrate the ways in which distinct capabilities can form around specific
experiences, as well as how those different capabilities may interact to influence the
ultimate outcome of a strategic decision. In this way recent scholarship on the
5
resource-based view asserting the importance of bundling capabilities to address
certain business processes (Ray, et al. 2004) and the idea that a wide range of factors
may play a role in the successful development of a single capability (Helfat, 2000) is
supported.
The following chapter will review the existing literature on learning and
capability development in the context of alliance relationships and lay out several
testable hypotheses. The third chapter outlines the industry in which these ideas will
be tested. The fourth chapter articulates the data, methods and analysis used to test the
hypotheses. The fifth chapter provides the results and the sixth chapter presents a
discussion of the findings. The final chapter articulates conclusions for theory and
practice and implications for expanding on the work presented here.

6
Chapter 2: Literature Review

Research on interorganizational alliances has been developed in many areas,
including what motivates firms to enter into alliances (Chung, Singh & Lee, 2000;
Simonin, 1999; Teece, 1986), how firms tend to choose alliance partners (Gulati,
1995a; Podolny, 1993) and what types of contracts govern relationships with different
characteristics (Crocker & Masten, 1988; Kalnins & Mayer, 2004; Joskow, 1987).
One area that is less well understood, however, is what experience capabilities lead a
firm to choose one type of alliance relationship over another. In an effort to explore
this area the following section will review the literature on the resource based view of
the firm (RBV) and its relationship to alliance decisions. Using several different types
of firm level experience as measures of firm level capabilities, several hypotheses are
developed to examine how firms determine alliance type decisions and whether
experience translates into increased performance as measured by multiplying the
market capitalization of a firm by its abnormal stock returns. Abnormal stock returns
measure the difference between stock returns upon announcement of a specific event
and the average variation in daily stock price over a given period prior to the event. In
other words, it determines whether the earnings (or losses) on the day of a particular
event are greater than the normal daily earnings (or losses) for that company.
The following section begins with a review of the literature on the resource-
based view of the firm and its connection to the alliance literature. This work is used
in conjunction with areas found to be important in prior alliance research to develop
hypotheses in the areas of alliance capabilities, including relational and partner
capabilities, and technological capabilities. Research in each of these areas will be
7
reviewed and hypotheses proposed to address how the capability contributes to new
relationship decisions. An additional section will explore how each capability type
translates into relationship performance.
Resource Based View
Research on the resource based view of the firm is based on early work by
Penrose (1959) who asserted that a “firm may achieve rents not because it has better
resources, but a firm’s distinctive competence involves making better use of its
resources” (p. 54). In this way, firms are differentially able to perform similar tasks
due to effective resource utilization. More recent work has expanded on the idea of
the resource based view by articulating what types of things qualify as firm resources
(Wernerfelt, 1984; Barney, 1991) and how resources can be used to generate rents
(Mahoney & Pandian, 1992). Wernerfelt (1984) included anything that could be
considered a strength or weakness of a firm in his definition of resources, such as
brand-names, technology and skilled personnel. Barney (1991) used Daft’s definition
of resources, which includes “all assets, capabilities, organizational processes, firm
attributes, information, knowledge, etc. controlled by a firm that enable the firm to
conceive of and implement strategies that improve its efficiency and effectiveness
(Daft, 1983)” (p. 101). With this conception, he attempted to identify whether or not
certain resources or resource combinations could provide firms with sustained
competitive advantage (Barney, 1991). Barney (1991) outlined four criteria necessary
for resources to provide sustained competitive advantages, including that they be rare,
imperfectly replicable by other firms, valuable and nonsubstitutable. In this way, the
resource-based view asserts that when a firm utilizes its resources or capabilities in
8
strategic decisions returns will be maximized (Mahoney & Pandian, 1992). When
applying this notion to the alliance context, it resonates with Zajac and Olsen’s (1993)
assertion that firms seek to maximize value in addition to minimizing costs in
transactions. Therefore, it can be expected that firms will seek to grow, or expand into
areas where it is perceived that they have resources which will produce competitive
advantage and thus will be able to successfully earn abnormal returns (Penrose, 1959).
Capabilities can come from effective use of many types of resources. Barney
(1991) discusses the way in which firm reputation may be considered a resource and
thus a source of competitive advantage. A reputation can be rare if only a few
competing firms have such a positive reputation and it can be imperfectly imitable if a
series of previous events have contributed to this reputation. An example might be
Johnson and Johnson’s (J&J) reputation, which is often partially attributed to its
treatment of the Tylenol crisis in 1982 (Mitchell, 1989). When several people died
after taking Tylenol, J&J’s president immediately recalled all Tylenol on the market.
It turned out that the deaths were the result of tampering, not due to any direct fault of
Johnson & Johnson. This decision was seen as highly respectable and has been lauded
as the model for handling organizational crises. It has also contributed to their
reputation as a responsible company. It would be difficult even for J&J to reproduce
the conditions which contributed to the development of this reputational resource.
Routines and organizational processes can also be capabilities (Zollo & Winter,
2002). Zollo and Winter (2002) describe how firms can develop capabilities through
routines. They identify learning mechanisms, including experience accumulation,
knowledge articulation and knowledge codification which contribute to capability
9
development. The idea that experience accumulated over time can result in
organizational routines stems from Cohen and Bacadayan (1994) and Nelson and
Winter’s (1982) work which argues that firms create routines that are stored as
procedural memory when repeated over time. Argote (1999) has also found that
organizational learning arises from repeated execution of similar tasks. For instance,
in the information technology industry reliability of electronic components is
important. A firm may have a valuable resource if it has developed a routine for
testing its components to ensure quality and reliability. This resource may take some
time to develop as the firm uses trial and error to find the most effective and efficient
way to test components. Once this routine has been established, it may however, serve
as a valuable resource for that firm.
More recent work in the area of the RBV has attempted to empirically validate
the resource-based view determining whether certain resources are in fact valuable,
rare, inimitable or non-substitutable (see Newbert, 2007 and Armstrong & Shimizu,
2007 for reviews). There has also been a lively debate surrounding whether or not the
theory is tautological, suggesting that something can be identified as a resource only if
it is valuable, and the theory is upheld when resources are found to be valuable (Priem
& Butler, 2001a; Barney, 2001; Priem & Butler, 2001b; Makadok, 2001). Although
this debate does, perhaps rightfully, question the value of Barney’s (1991) four
characteristics of what constitutes a resource, there has been another emerging theme
in this research which proposes a potentially more valuable and productive way to
think about the resource-based view. This conversation emphasizes the broader idea
of “competitive heterogeneity” (Hoopes, Madsen & Walker, 2003), where firms have
10
varying degrees of competitive competence. Also intertwined with this notion is
the role that evolution or development of capabilities plays in the resource based view.
Helfat (2000) asserts that capabilities are constantly evolving and that different factors
may play an important role in competitiveness at different times, but in reality the
factors are inextricably linked. Evolving capabilities have been referred to as dynamic
capabilities (Eisenhardt & Martin, 2000; Helfat & Peteraf, 2003). Barney, Wright and
Ketchen (2001) simply define dynamic capabilities as “capabilities that are dynamic”
(p. 630). The literature on the resource-based view has also been criticized for the
simplistic ways in which resources are measured (Priem & Butler, 2001a) which has
precipitated several articles addressing methodology. Rouse and Daellenbach (1999)
suggest fieldwork or survey methodology to get at the complexity inherent in
capabilities, but in general issue a call for recognition in methodological choices of the
complexity of capabilities (Eisenhardt & Martin, 2000; Ray et al, 2004) due to the
potential ways in which knowledge and routines can interact to create new knowledge
(Zander & Kogut, 1995). This study attempts to take a modest step toward capturing
capability complexity by incorporating several different types of capability data to
understand their collective impact on strategic alliance formation and performance.
The RBV has also been applied to a range of issues specifically in the area of
strategic alliances (Combs & Ketchen, 1999; Madhok & Tallman, 1998; Eisenhardt &
Schoonhoven, 1996; Arya & Lin, 2007). Capabilities have been used in a variety of
ways in this context. For instance, Mayer and Salomon (2006) explored how
governance is determined in light of technological and partner capabilities in the
information technology sector. They found that firms with strong technological
11
capabilities were more likely to subcontract in the presence of hold-up contractual
hazards. Hold-up situations arise when a transaction requires investment in assets that
have little value outside the transaction and thus provide an opportunity for a partner
to act opportunistically to earn excessive rents. In this way, the resources that a firm
possesses influence the strategic decisions that a firm makes in the context of
governance choice. Wang and Zajac (2007) looked at how relational and partner
capabilities influence a firm’s decision to ally or to acquire. They found that dyads
with more collective alliance experience were more likely to engage in future alliance
relationships as well as acquisitions. This finding supports the idea that alliance
experience builds the alliance capability by providing firms with more general
knowledge on how to deal with other firms as well as with alliance relationships.
As can be seen, there are a variety of capabilities that a firm could choose to
capitalize on when making strategic decisions. As a result, it is unlikely that
measurement of any set of capabilities will be able to fully explain a firm’s strategic
decisions and resulting performance. This study has relied on existing research to
identify several different capabilities that have been found to be critical in the alliance
formation process: alliance capabilities and technological capabilities. In following
with prior research, the extent of experience with a given process or routine will be
used to characterize firm level capabilities (Mayer & Salomon, 2006; Rothaermel &
Deeds, 2006; Wang & Zajac; 2007; Park, Mezias & Song, 2004). The research
reviewed below aims to emphasize the importance of the two capabilities and the
expected ways in which both alliance and technological capabilities interact to predict
strategic alliance decisions and outcomes.
12
Alliance capability
In this study three types of alliances are explored, including the license
agreement, the research and development (R&D) agreement and the non-R&D
agreement. License agreements include those alliances that grant another entity the
rights to sell, market or distribute a product. R&D agreements in this study include
agreements where research, collaboration and development take place among the
partners. Finally, non-R&D agreements are those relationships that are related to the
distribution, sales and marketing of products and technology. In this way it could be
said that the distinction between R&D agreements and non-R&D agreements is that
R&D relationships are characterized by product development, whereas non-R&D
relationships are characterized by bringing the product to market.
Each of these alliance types requires differing routines, interaction and
interdependence (Contractor & Lorange, 1988; Gulati, 1995a). Over time firms learn
to use prior experiences to obtain skills to interact with and manage other firms within
the relationship (Cohen & Levinthal, 1990; Lorenzoni & Lipparini, 1999). In this way,
general alliance experience alone may not provide enough of a picture of the types of
experiences that may assist a firm in choosing the most successful alliance type as
demonstrated by Kale et al’s (2002) research. It is possible that certain types of
alliance experience employ different routines, causing only certain routines to mature
in one alliance type, while others mature with experience in another alliance type. For
example, research has shown that more interdependent relationships, such as research
and development agreements or joint ventures, are more uncertain, and as a result
firms experience more unexpected events due to the difficulty in fully planning for or
13
articulating all contingencies in a relationship (Arino & Torre, 1998). In this way,
experience with different types of alliances may result in distinct capabilities. In the
instance of more interactive relationships a firm may learn how to manage internal
resources to facilitate integration with the partner firm. For more arms-length license
agreements, a firm may learn which types of situations warrant more detailed
contractual terms or which terms are key to successful license relationships. Therefore,
it is expected that firms will develop capabilities for different types of interdependence
and as a result choose to capitalize on those capabilities by entering new relationships
of the same type.
H1a: Firms with more experience with license agreements will be more likely
to enter into new license agreements than other types of agreements
H1b: Firms with more experience with non-R&D agreements will be more
likely to enter into new non-R&D agreements than other types of agreements
H1c: Firms with more experience with R&D agreements will be more likely to
enter into new R&D agreements than other types of agreements
In addition to thinking of capability development as a way to develop
specialized knowledge or experience in a specific type of relationship, it is also
possible that firms will seek to expand their range of capabilities when choosing new
relationship types. Powell et al (2005) found that diversity, or the tendency to enter
into different types of relationships was the dominant attachment logic of a large
network of biotechnology firms. The network literature has also found that diversity
of ties can be beneficial. Burt (1995) found that actors that connected different,
unconnected groups of actors were more likely to have access to new ideas. Uzzi
14
(1997) found that individuals that only sought new relationships with the same
group of partners often became cut off from changes in the external environment.
Extending these findings to the area of alliance capability development, experience
with a new type of relationship may provide the firm with new information about
other firms, about the environment, or about how contracts for other relationship types
are constructed. In short, there may be learning benefits to increasing diversity of ties
a firm has in its portfolio. In this way, it can be expected that diversity will play a
significant role in the relationship selection process. Which type of relationship
selected will depend on a firm’s existing portfolio of experience and therefore the type
is not explicitly hypothesized in this study.
H2: Firms will be more likely to choose relationship types that increase
diversity than those that decrease diversity
Product Development Capability
One of the common motivations articulated in the research for entering into
alliance relationships is access to resources. Teece (1986) notes that additional
resources may be necessary for firms to capitalize on their own innovations or
competencies. In this way, an alliance represents an opportunity to utilize resources to
maximize firm value. Van de ven (1976) highlights resource exchange as one of the
primary dimensions of interorganizational relationships, and suggests that an internal
need for resources motivates organizations to seek such relations. In a study of
technology alliances, Hagedoorn (1993) found that technology resource
complementarity and market access are both major motivations for firms to enter into
technology alliances. He also found that these resources are often sought with the
15
hope of reducing the innovation time span and bringing products to market more
quickly. Resource complementarity also plays a significant role in predicting choice
of alliance partners in several different industries (Chung et al. 2000; Gulati, 1995b).
Resource complementarity is measured in these cases by looking at the niche in which
firms operate within an industry. If firms operate in non-overlapping niches, then their
relationships are considered to include resource complementarity in that one niche has
a certain set of resources not available to other niches. Chung et al. (2000) studied
alliance formation in the investment banking industry, an industry known for its
emphasis on firm status. Chung et al. (2000) pointed out that in industries like the I-
banking industry niches can be conceived from multiple perspectives; for this reason
they used multiple variables to operationalize niches, including whether the firms are
issuers and investors in a particular area and whether the bank has strength in a
particular geographic location.
One common goal in any firm is the successful launch of new products or
services. As Teece (1986) notes, it is often necessary to seek out partners with
complementary capabilities in order to accomplish this goal. In this way, the number
of products that a firm has on the market represents a capability of the firm to
capitalize on its resources, whether those resources include product development,
product marketing or product manufacturing. In the biotechnology industry, for
instance, the number of products a firm has represents the firm’s competence in taking
a product through the FDA clinical trials, conducting the research to find a viable
product formula or locating a firm with suitable complementary skills to make this
happen (Deeds & Hill, 1996; Rotheramel & Deeds, 2006). The total number of
16
products that a biotechnology company has in production is an indicator of the
amount of experience the firm has with the routines necessary to create a product.
Since the time to market for a typical product in this industry is 8-15 years, the number
of products, rather than the time it takes to bring a product to market is a more
appropriate measure of product experience (Rotheramel & Deeds, 2006). Deeds and
Hill (1996) find that as a firm enters more alliance relationships their product
development increases; however, eventually it reaches a point of diminishing returns.
The rationale behind this finding is that using alliances to gain access to
complementary resources is not a risk free venture and performance of the
relationships may suffer as a firm enters into a greater number of relationships and is
unable to adequately invest necessary resources into partner selection or relationship
management (Deeds & Hill, 1996). In this way, more relationships are better, but only
to a point. When examining the implications that this finding might have for the types
of relationships a firm will enter, while optimizing their product development, it
stands to reason that a firm will be able to handle more low-interaction relationships
and fewer high interaction relationships. Research and development relationships
require a high amount of interaction (Contractor & Lorrange, 1988), and since R&D in
the biotechnology industry is a long-term process, it is more likely that firms that are
able to produce more products will have a higher likelihood of entering into more non-
R&D and license agreements than more interaction intensive R&D agreements.

17
Therefore, when predicting how product experience might influence alliance type
decisions, it is expected that:
H3: Firms with more product experience are more likely to enter into non-
R&D and license agreements than R&D agreements
Product Area Capability
In addition to accessing resources, firms often seek out alliances for the
purposes of learning (Khanna, Gulati & Nohria, 1998; Child, 2001; Simonin, 1999).
For instance, Kogut (1988) concludes through a review of empirical work on joint
ventures that knowledge acquisition is one of the primary motivations for entering
such relationships. The issue of knowledge and how it operates within the firm has
been the subject of much theorizing (Grant, 1996; Spender, 1996). Tsoukas (1996)
conceptualized the firm as a distributed knowledge system, asserting that knowledge is
local and thus distributed throughout the firm. A challenge to the firm in this situation
is that no one individual holds all the knowledge necessary for the firm to capitalize on
all its resources, nor does one person know exactly what all the firms members know
(Tsoukas, 1996). In this way, utilizing knowledge held within the firm is a difficult
task. Others have similarly written about the challenges involved in transferring
knowledge and its relationship to firm level capabilities. Kogut and Zander (1992)
assert that knowledge can be a competitive advantage provided it is not easily imitated
or transferred. Knowledge transfer occurs through new combinations of existing
knowledge and within the context of social interactions. Through these processes of
recombination and interaction new capabilities can be developed within the firm.
Characteristics of knowledge, such as codifiability, can also influence the extent to
18
which firms are able to effectively spread capabilities across firms, but also
contribute to the ease with which such capabilities can be imitated by other firms
(Zander & Kogut, 1996). Knowledge based theories of the firm view the firm as a
dynamic, evolving system of knowledge (Spender, 1996) where individuals hold
knowledge and where organizational structures are responsible for integrating and
applying that knowledge (Grant, 1996). This perspective provides a framework
through which to understand interorganizational relationships in two different ways.
One way is as a means to learn or acquire new knowledge. A second way is as an
opportunity to capitalize on acquired knowledge.
Empirical research looks at alliances in both these ways. First, researchers
have examined what types of processes facilitate learning and the development of
capabilities. One motivation for a firm to enter into a joint venture (JV) relationship is
to learn from a partner. Inkpen and Dinur (1998) identified four processes which are
critical for parent-JV learning. These processes included technology sharing, JV-
parent interaction, personnel transfers and strategic integration. Technology sharing
refers to both firms agreeing to share manufacturing or product technology in some
capacity, whether through meetings with plant managers and manufacturing directors
or borrowing engineers to help analyze a firm process. JV-parent interactions allow
knowledge sharing between the JV and the parent companies to extend the learning
happening through the venture to the sponsoring firm. Personnel transfers are simply
the sharing of experts across organizations, which can have knowledge spillover
effects. Finally strategic integration provides opportunities for each entity to learn
from the other. In the instance of a parent and JV, an example of a strategic link could
19
be when a parent firm uses the JV to produce a product which provides
opportunities for knowledge sharing and future cooperation (Inkpen & Dinur, 1998).
For instance, if a large pharmaceutical company is hoping to move into more
biologically-based products, it may form a joint venture with a biotechnology
company and send several key scientists to do rotations in the joint venture labs to
learn more about biologically-based pharmaceutical research. Over time, these
scientists will gain an expertise in an area that the pharmaceutical company may
capitalize on by using the knowledge to create biologically-based products in-house.
A second way to look at the role of learning in alliance relationships is how
existing capabilities might best be utilized in new relationships. Mowery, Oxley and
Silverman (1996) explore how knowledge in interorganizational relationships impacts
firms’ technological capabilities. They find that equity joint ventures are more
effective in transferring complex technical knowledge than more arms-length
relationships, such as licensing agreements. This suggests that certain types of
relationships provide a more effective means to capitalize on existing knowledge than
other relationship types. They also find evidence supporting the idea of absorptive
capacity. Absorptive capacity refers to the extent to which a firm is able to learn to
apply knowledge or information (Lane & Lubatkin, 1998). Absorptive capacity is
often measured by looking at the extent of cross-patenting to indicate the extent to
which a firm was able to utilize knowledge from the partner firm. Specifically,
Mowery et al (1996) find that a firm’s prior knowledge in an area, as measured by
their patent portfolio, influences the extent to which the firm is able to acquire
knowledge in that area from a partner.
20
The research reviewed above provides some insight into the types of alliance
relationships that firms seek out when attempting to expand their knowledge of a
specific area. It also suggests that there may be two different strategies for expanding
capabilities involved in the decision to enter new alliances. These two strategies
might be referred to as exploration and exploitation (March, 1991). The exploration
method suggests that experience in a specific product area provides firms with the
specific knowledge necessary to share, apply and develop routines to create new
specialized knowledge (Cohen & Levinthal, 1990). In this way, it may be expected
that firms will enter into new exploratory relationships, or research and development
relationships in areas where they have existing competencies. Conversely, the idea of
exploitation may suggest that firms will seek to utilize the capabilities they have in a
specific area when selecting new alliance relationships. When applying this idea to
the types of experience in a product area, the exploitation strategy suggests that firms
that have had successful products in the area of heart surgery, for instance, will have
developed competencies in that area. They will likely have a sales force that is
knowledgeable in the area, contacts with specialists in the area, and relationships with
other downstream audiences necessary to make more similar products successful.
Therefore, getting access to new similar products is a means for capitalizing on these
capabilities. License and non-research and development agreements, such as
marketing or distribution agreements will both provide access to new products.
Therefore, two competing hypotheses are proposed, one that asserts the exploration
strategy, the expansion of knowledge in a specific product area, and one that asserts
the exploitation strategy, or the utilization of knowledge in a specific area.
21
H4a: Firms with greater experience in a product area will be more likely to
enter into R&D agreements than non-R&D and license agreements.
H4b: Firms with greater experience in a product area will be more likely to
enter into non-R&D and license agreements than R&D agreements in that
same product area.
Partner Capability
Inspired by Granovetter’s (1985) classic work on embeddedness, a large body
of alliance research examines the ways in which social relations influence economic
exchange (Uzzi, 1996; 1997; Gulati, 1995; Kogut & Zander, 1996). Such work
explores partner selection, governance in relations and the role of trust in alliance
relationships. It is possible that experience with past partners may include a set of
partner specific routines which constitute a unique type of capability, and thus also
influence a firm to repeatedly select the same partner to execute new relationships.
The discussion of relational embeddedness at the firm level often centers on
the concept of trust. Trust is a difficult issue to understand in this context due to the
conflation of trust as an organization level construct as opposed to an interpersonal
level construct (Zaheer, McEvily & Perrone, 1998). Organizational level trust is
defined as the trust that those within the organizations have for the partner
organization (Zaheer, et al 1998). Trust has been found to accrue incrementally as two
firms gain more experience with each other. This is suggested in Gulati’s (1995a)
findings in that the larger the number of prior ties between two firms the more likely
these firms are to form a new non-equity based relationship. Non-equity relationships
are thought of as relationships that require more trust due to the uneven exchange of
22
resources, or equity invested in the relationship. In this way, trust increases with
experience and reduces transaction costs associated with repeat relationships because
equity agreements are more costly to develop than non-equity relationships (Gulati,
1995a). In a study conducted in the insurance industry, Zaheer and Venkatraman
(1995) found that when firms have higher levels of quasi-integration, meaning they
depend heavily on each other for business or necessary resources, they are more likely
to have trusting relationships. These trusting relationships are also found to positively
correlate with greater degrees of joint action, or integrated organizational processes
(Zaheer & Venkatraman, 1995). Board interlocks characterized by trusting
relationships have also been found to increase the likelihood that alliance relationships
will form between firms which the trusted board members represent (Gulati &
Westphal, 1999). In this way, trust can be seen as a competitive advantage in that it
can work to reduce transaction costs. Barney and Hansen (1994) note, however, that
in order for trust to act as a competitive advantage firms need to appropriately match
governance with the type of trust present in each exchange. They identify three forms
of trust, strong form trust, semi-strong form trust and weak form trust. Strong form
trust refers to those relationships where the “trust emerges in the face of significant
exchange vulnerabilities, independent of whether or not elaborate social and economic
governance mechanisms exist, because opportunistic behavior would violate values,
principles and standards of behavior that have been internalized by parties to an
exchange” (Barney & Hansen, 1994, p. 179). Weak form trust on the other hand
exists in exchanges were no vulnerabilities are present. Semi-strong form trust is
created when governance is erected to protect against any possible vulnerabilities.
23
Research has also shown that when two firms work together over a series of
relationships that they learn in the process to work together more effectively, as
evidenced in the increasing specificity of contractual provisions (Mayer & Argyris,
2004). In this way, a partner capability may also reflect the establishment of routines
to discover the best ways to define roles and clarify expectations in relationships.
Similarly, partner specific knowledge has been found to increase the likelihood of
future alliances and acquisitions (Wang & Zajac, 2007). Therefore, if firms develop a
capability to work with certain partners on certain types of relationships, and firm
capabilities do in fact drive strategic decisions, then it can be expected that:
H5: Firms with experience with a partner will be more likely to enter into
alliance types with which they have experience with that partner compared to
other alliance types
Performance
In the previous section, it was argued that firms will utilize a variety of
capabilities when making decisions to enter into new strategic alliances. It is hoped
that this variety will begin to capture the complexity characterized in descriptions of
capabilities in more recent resource-based literature (Eisenhardt & Martin, 2000;
Barney, 2001). This section seeks to test the performance aspect of the resource-based
view. The ultimate goal in understanding firm level capabilities is what types of
capabilities produce sustained competitive advantage, clearly a different measure than
performance, as it indicates an enduring advantage over time (Rouse & Daellenbach,
1999). Performance, however, represents an important first step toward gaining
sustained competitive advantage. Ray, Barney and Muhanna (2004) suggest that
24
looking at the performance of specific business processes is favorable to using an
overall firm level measure of performance, such as return on assets. This study,
therefore, will hypothesize relationship level performance based on each of the
existing firm level capabilities introduced above.
Alliance Capability and Alliance Performance
Work in the area of alliance capabilities has focused on how firms can improve
management of individual alliance relationships to improve success rates. Kale, Dyer
and Singh (2002) found that alliance experience and a dedicated alliance function, a
person or team responsible for managing alliances, improve the long term success and
result in positive abnormal stock market returns on announcement of the relationship.
Simonin (1999) asserts that a firm must actively translate alliance experience into
know-how before it can be utilized and contribute to a collaborative capability. Kale
and Singh (2007) also find that the process of sharing and codification of alliance
experience contributes to development of an alliance capability. Zollo and colleagues
look at how routines can be developed to make alliance outcomes more consistently
positive. They find that general alliance experience is not directly related to increased
success and attribute this to difficulty assessing performance in a way that allows for
adjustment of routines and to the ambiguity and uncertainty associated with each new
alliance (Zollo, Reuer & Singh, 2002). Since this study also tries to capture a more
complex picture of firm capabilities by also measuring alliance type and technological
capabilities, it is possible that when controlling for these other types of experience that
general alliance experience will positively influence performance of new alliance
relationships.
25
H6: Firms with more general alliance experience will have stronger alliance
performance than firms with less general alliance experience
In addition to general alliance experience, it is also possible that specific types
of experience will provide firms with increased performance when entering new
relationships of that same type. Anand and Khanna (2000) examined the learning
effects of license agreements and joint ventures and found that the cumulative
experience with joint ventures improves stock market reaction to announcement of a
new joint venture, whereas the same does not hold true for license agreements. This
study includes a different set of alliance type experience variables for which
performance has yet to be tested.
H7a: For firms that are entering into license agreements, more experience with
license agreements will result in stronger alliance performance than firms with
less license agreement experience
H7b: For firms that are entering into non-R&D agreements, more experience
with non-R&D agreements will result in stronger alliance performance than
firms with less non-R&D agreement experience
H7c: For firms that are entering into R&D agreements, more experience with
R&D agreements will result in stronger alliance performance than firms with
less R&D agreement experience

26
Another way in which alliance experience may be captured is by the
diversity of experience a firm has in its portfolio. Diversity refers to the distribution
of experience across the three alliance types, and therefore, greater portfolio diversity
may yield increased performance.
H7d: More alliance type diversity in a firm’s experience portfolio will result
in stronger alliance performance
Product Development Capabilities, Product Area Capabilities and Alliance
Performance
As mentioned in the discussion of product development capabilities above, a
common motivation for firms entering into alliance relationships is to gain access to
new technologies and speed the innovation process (Hagedoorn, 1993) to ultimately
produce new products. Product development experience has been used as an outcome
measure in the biotechnology industry on the grounds that it demonstrates a firm’s
ability to employ the proper routines to bring products through the development phase
and ultimately to the market (Rothermel & Deeds, 2006). This notion is also
supported by Cohen & Levinthal’s (1990) idea of absorptive capacity. Previous
experience developing products provides firms a capacity to absorb, or draw upon a
store of prior related knowledge in making new products, thus ultimately increasing
the likelihood of being able to produce more new products. It is unlikely that this
capability alone will necessarily provide firms with increased alliance performance as
a firm may have produced all products in its product portfolio independent of any
relationships with other firms. It is, however, possible that the ability of a firm to
develop products is partially a reflection of its ability to successfully select and enter
27
opportune alliance relationships that produce successful products. Therefore,
interacting the product development capability with an alliance capability may
increase the overall performance of alliance relationships.
H8: The interaction of alliance capability and product development capability
will have a positive impact on a firm’s alliance performance
In a similar way, a firm’s capability in a specific product area may be due to an
isolated expertise with a specific technology, or perhaps even due to the one discovery,
or one highly effective scientific expert in a certain area. As a result, looking at
product area capability alone may suggest a proficiency in one area of expertise that
was developed without the participation of other firms in the form of alliance
relationships. Over time, the likelihood that a firm will need to look to external firms
for access to new ideas or skills will increase. Therefore, consideration of the firm’s
alliance experience in addition to its product area experience may improve the overall
performance of the alliance.
H9: The interaction of alliance capability and product area capability will have
a positive impact on a firm’s alliance performance
Partner Capability and Alliance Performance
In seeking out the competitive advantage or abnormal returns available in
trustworthy relationships, firms often seek out repeated relationships or links with the
same set of firms resulting in relational networks among that set of firms (Gulati,
1998; Gulati & Gargiulo, 1999). These networks provide member firms with
information on other firms within the network regarding trustworthiness as well as
capabilities (Gulati, 1995a; Gulati, Nohria & Zaheer, 2000). The resulting structure of
28
these networks, in turn works to influence the future decisions of individual firms
(Gulati, 1998; Sydow & Windeler, 1998). Gulati and Gargiulo (1999) have found that
interdependence among actors within a network, as well as their relational and
structural embeddedness influence partner selection. Relational embeddedness refers
to the history of direct interaction that shapes the set of actions available to individuals
within the relationship (Gulati, 1998). Structural embeddedness moves beyond the
immediate direct ties captured in relational embeddedness to emphasize how the
structural positions of the partners in a network of relationships provide access to
information about other organizations in the network through indirect ties (Gulati,
1998). Such structural constraints can produce highly embedded firms, who rely
exclusively on a close knit group of ties to carry out transactions (Uzzi, 1996). Uzzi
(1996) found, in a study of New York apparel companies, that embedded ties, as
opposed to arms-length market relations, provide firms with trust, fine grained
information and joint problem solving. Additional studies have found that embedded
relationships can also produce increased levels of tacit knowledge transfer and
knowledge spillovers (Reagans & McEvily, 2003; Uzzi & Gillespie, 2002; Uzzi &
Lancaster, 2003) as well as decreased cost of capital (Uzzi, 1999).
Decreased transaction costs, increased information and knowledge sharing and
problem solving all due to the trust inherent in embedded alliance networks provide
valuable benefits to firms who choose to transact within them (Gulati, 1998).
Although it is clear that these are benefits that reduce uncertainty in the context of
alliance relationships, it is less clear whether these experiences with past partners yield
increased success in new relationships with the same partner. However, since firms
29
who have worked together in the past have likely developed knowledge of each
firm’s capabilities and developed routines and trust that working with a past partner
will increase the success of the relationship, it is expected that:
H10: For firms entering into alliances with past partners, partner experience
will be positively related to alliance performance
Table 2.1 includes a summary of all hypothesized relationships articulated
throughout Chapter 2.
30

Table 2.1. Hypothesis summary

IV DV Prediction
H1a LNExpLic DealType Positive relation between
experience and likelihood of
entering same type again
H1b LNExpNRD DealType Positive relation between
experience and likelihood of
entering same type again
H1c LNExpRD DealType Positive relation between
experience and likelihood of
entering same type again
H2 Diversity DealType Positive relation between
diversity and DealType
H3 LNTotProd DealType Increased likelihood of
entering NRD and Lic over
R&D
H4a LNExpProd DealType Increased likelihood of
entering R&D over NRD and
Lic
H4b LNExpProd DealType Increased likelihood of
entering NRD and Lic over
R&D
H5 LNExpPart
(for ExpPart >0 only)
DealType Positive relation between
experience with a partner and
likelihood of entering again
with same partner
H6 ExpAlliance DealPerf Positve
H7a ExpLic*LicDeal DealPerf Positive
H7b ExpNRD*NRDDeal DealPerf Positive
H7c ExpRD*RDDeal DealPerf Positive
H7d Diversity DealPerf Positive
H8 ExpAlliance*TotProd DealPerf Positive
H9 ExpAlliance*ExpProd DealPerf Positive
H10 ExpLic
ExpNRD
(for ExpPart >0 only)
DealType Positive relation between
experience and likelihood of
entering same type again

31
Chapter 3: The Biotechnology Industry
The biotechnology industry will serve as the testing bed for the ideas presented
above. The biotechnology industry has seen many changes over the last 30 years, as it
was integral in fundamentally altering both the science and the business of the broader
life sciences industry. The following chapter will begin with an overview of the
history of the biotechnology industry, a review of academic research will follow,
concluding with justification for the appropriateness of this industry to test the ideas
outlined above.
History of the Biotechnology Industry
Many refer to the introduction of biotechnology into the life sciences industry
in the early 1970s as a revolution (i.e. Zucker & Darby, 1997). Others, however,
contest this label based on the broader history of life science, showing that the
biological heuristic is not new, but newly appreciated with recent discoveries
(Hopkins et al, 2007). Regardless of whether it is considered a revolution, the
scientific developments of the early 1970s fundamentally changed not just how
biological science is conducted, but how the industry is organized. In the postwar
1940s through 1960s, pharmaceutical science was characterized by synthetic
production of natural molecules and based on organic chemistry (Hopkins et al. 2007).
At this time little was known about the actual biological structures which the
chemicals were meant to target (referred to as “targets”; Gilsing & Nooteboom, 2006).
As a result drug discovery was based on a brute force method of screening the impact
of different chemicals on targets (Gilsing & Nooteboom, 2006). In the early 1970s
there was a shift toward a more intentional focus on the biological process, which
32
resulted in hypothesis-driven tests of compounds’ effects on targets, and also
allowed for intentional targeting of specific chronic diseases (Hopkins, et al, 2007).
This new focus resulted in many new blockbuster drugs for pharmaceutical companies.
Also in 1973, Stanley Cohen and Herbert Boyer discovered the basic process for
making recombinant DNA (r-DNA), which is a method of making proteins, such as
human insulin, in controlled manufacturing conditions (Pisano, 1991; BIO, 2008).
This discovery is often viewed as the advent of the biotechnology revolution (Zucker
& Darby, 1997).
In addition to the development of r-DNA, other improvements in process and
scientific knowledge contributed to this industry. High throughput screening (HTS)
was developed in the 1980s and continues to be refined. HTS is a process which
allows the combination of the random search and hypothesis driven search by using
robotics to test large quantities of compounds (Gilsing & Nooteboom, 2006;
McKelvey & Orsenigo, 2004). The robotics in this case can dispense very small
amounts of a liquid suspension with the same concentration into plates with one
thousand compartments. Each of these compartments can then be used to reproduce
the same experiment many times over in a short period of time. Nineteen-ninety also
marked the advent of the Human Genome Project (HGP), cosponsored by the National
Institutes of Health and the U.S. Department of Energy. The purpose of the HGP was
to identify all the genes in the human genome in an effort to make the sequences
available for further research. The $3 billion international project (Bylinsky, 1994)
actually took 13 years and was completed in 2003. Like a lot of basic scientific
research, the application of knowledge gained from the HGP is still highly ambiguous.
33
James Watson, the scientist who discovered the double-helix shape of DNA,
predicted that it will take geneticists 10,000 years to fully understand the functions of
the human genome (Bylinsky, 1994).
Promise of these scientific developments spawned the birth of hundreds of
dedicated biotechnology firms. In 1994 there were almost 250 publicly traded
biotechnology companies (Alster, 1994). By 2006 this number had grown to about
400 publicly traded companies with another 1,200 or so being privately held. This
explosion of new companies not only created a new sector of the life sciences industry
it fundamentally altered the way the broader life science industry was organized.
Pharmaceutical companies, pressured by shrinking pipelines due to patent expirations,
were looking for new ways to add to their product lines and supplement their pipelines
(Alster, 1994; Tully, 1994). The resulting partnerships, mergers and acquisitions
created an industry of densely connected organizations, including pharmaceutical
firms, biotech firms, universities and public research institutes (Powell, et al. 2005;
Roijakkers & Hagedoorn, 2006). In this process biotech companies have continued to
be increasingly successful. The first biotechnology drug was approved in 1982 and
the number of new approvals has continued to grow steadily. This is shown in Figure
3.1. Sales and revenues have increased in a similar manner, as demonstrated in Figure
3.2.

34
Figure 3.1: New Biotech Drug and Vaccine Approvals/ New Indication Approvals by Year

(Source: BIO: http://bio.org/speeches/pub/er/statistics)

35
Figure 3.2: Biotechnology Industry Sales and Revenue by Year
Biotechnology Sales and Revenue
0
10
20
30
40
50
60
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Year
USD$ (in billions)
Sales
Revenues

(Source of Data: BIO: http://bio.org/speeches/pub/er/statistics)

36
The scientific advancements in the biotechnology industry are credited with
driving the changes within the life sciences industry over the last several decades, but
additional regulatory factors have also contributed to the industry’s development. One
of the notable factors in the biotechnology industry is the involvement of universities
and their research scientists. Social networks between these researchers and biotech
companies have been found to increase innovation for biotech companies (Liebeskind
et al, 1996). The Bayh-Doyle Act of 1980 has been identified as a key instigator of
this collaboration in the United States (Gittelman, 2006). The Bayh-Doyle Act came
out of a desire to improve industrial competitiveness by allowing a way for companies
to obtain intellectual property rights (IPR) from universities for developments that
might otherwise have remained part of public domain (Gittelman, 2006). The result
was increased patenting of scientific discoveries by universities, making them an
excellent source of new knowledge and information (Owen-Smith & Powell, 2003).
Biotechnology companies were often funded by pharmaceutical companies in
early stages of the industry, acting as the external R&D divisions of these large
companies (Pisano, 1991). Also, due to the seeming promise of the new techniques
emerging from biotechnology companies venture capital and IPOs have been a huge
funding source, peaking in the year 2000 at $32.7 billion in funding raised, just prior
to the bursting of the dot-com bubble (Weintraub, 2003). This can be seen in the stock
index graph in Figure 3.3 below.

37
Figure 3.3: Biotechnology and Pharmaceutical Stock Indexes

(Source: Yahoo Finance http://quote.yahoo.com)
Life Sciences Indexes
0
200
400
600
800
1000
1200
1400
1600
Date
7/1/1994
1/3/1995
7/3/1995
1/2/1996
7/1/1996
1/2/1997
7/1/1997
1/2/1998
7/1/1998
1/4/1999
7/1/1999
1/3/2000
7/3/2000
1/2/2001
7/2/2001
1/2/2002
7/1/2002
1/2/2003
7/1/2003
1/2/2004
7/1/2004
1/3/2005
7/1/2005
1/3/2006
7/3/2006
1/3/2007
7/2/2007
1/2/2008
Date
Pharmaceutical Index (AMEX)
Biotechnology Index (NASDAQ)

38
In addition to funding variation, regulatory and legitimacy issues are also
factors that contribute to the dynamics of the biotechnology industry. Approval of all
drugs, devices and technologies that come out of biotech research laboratories is
required by the U.S. Food and Drug Administration (FDA). The average time it takes
for a drug to make it through the FDA approval process is approximately 8.5 years
(McNamara & Baden-Fuller, 2007). This does not include the time taken to develop
the drug prior to it entering into FDA trials. Genetech’s colon cancer drug, Avastin,
for instance, was in development and clinical trials for 14 years until any approved
uses were granted by the FDA (Weintraub, 2003). Navigating this bureaucratic
process is seen as a competitive advantage, and a resource that many biotechnology
companies sought out in alliances with pharmaceutical companies (Rothaermel, 2001).
Since the FDA acts as the gatekeeper between drugs and the market, it is subject to
regulatory pressure when potential problems arise. When problems with Merck’s
Vioxx, AstraZeneca’s Iressa and others become public knowledge, pressure is placed
on the FDA to impose greater restrictions on products in the clinical trial phase. This
can make success in an already risky industry increasingly difficult. Therefore, as it
has been shown, the life sciences industry has seen considerable change over the past
40 years, and as a result it has also been an object of interest and study in the academic
community.
Academic Research on the Biotechnology Industry
The biotechnology industry has been the subject of extensive academic
research in recent years. The themes pervading this research include issues of

39
governance, legitimacy and access to funding, innovation, and interorganizational
relationships. This field provides a valuable testing ground due to the proliferation of
alliance formation and interorganizational linkages that have taken place over the last
two to three decades (Powell et al. 2005). Robinson and Stuart (2001) looked at how
the network of past alliances serves as a governance mechanism within interfirm
transactions. They find that as a firm becomes more central in a network and more
proximate to other firms, access to funding increases and equity participation
decreases (Robinson & Stuart, 2001). Higgins and Gulati (2003) find that when
members of the upper echelons of a new firm have affiliations with powerful
downstream firms, such as large pharmaceutical organizations they are able to attract
more prestigious underwriters in initial public offerings (IPO). These upper echelon
connections also contribute to alliance formation between firms with similar
connections (Kim & Higgins, 2007). Stuart, Hoang and Hybels (1999) similarly show
that young companies that form alliances with more powerful established companies
are more able to attract venture capital and earn greater valuations in IPOs than firms
with no such affiliations. Finally, young firms are more able to attract revenue
generating alliances with multiple in-licensing agreements (Stuart, Ozdemir & Ding,
2007). Also, affiliations with scientists in universities increase the chances that firms
can acquire commercialization rights to discoveries from universities (Stuart et al,
2007). Firms with no affiliations also often enter into their first relationship at a
discount, justified by the alliance partner as payment for gathering information on the
viability of the firm (Nicholson, Danzon & McCullough, 2005). These studies show

40
how alliance relationships can be a valuable resource to firms, even beyond the
utility gained in the relationships themselves.
Alliance networks have also been explored in detail within the biotechnology
field. Roijakkers and Hagedoorn (2006) explore R&D partnering trends in the biotech
industry from 1975-1999, finding less than 50 relationships in 1975 with the trend
peaking at almost 700 in 1996, resulting in a highly interconnected network of firms.
Also, small biotechnology companies are involved in the majority of these
relationships in the 1980s with the pharmaceutical companies moving into more
central positions, forming many relationships in the 1990s (Roijakkers & Hagedoorn,
2006). Powell et al. (2005) explore the evolution of attachment in these networks,
showing that early attachment in the industry is of a commercial nature, later shifting
to more collaborative activities involving universities, venture capital and small firms.
In the long run, however, the preference for diverse linkages was the dominant
attachment logic (Powell, et al. 2005). Walker and colleagues find that social capital
and path dependence explain the development of interorganizational networks in this
industry over time (Walker, Kogut & Khan, 1997). Those firms that rely on social
capital, tending to share the same set of alliance partners, have enduring networks and
those that are not structurally equivalent tend to remain that way.
In addition to looking at the network structures, studies have looked at the role
of past interactions on performance and future interactions. Contrary to prior research
in other industries, R&D partnerships in the biotech industry seem to be negatively
related to the likelihood for future partnerships (Roijakkers et al. 2005). The authors

41
attribute this to the dual market structure of these relationships, containing a small
biotech firm and a large pharmaceutical firm, stating that they are not likely to be
characterized by equality and mutual dependence (Roijakkers et al. 2005). Work by
Hoang and Rothaermel (2005) offers a potentially different explanation for the
tendency for the negative relationship between past R&D partnering and future
partnerships. They find that partner specific experience negatively contributes to
R&D relationship performance (Hoang & Rothaermel, 2005). It is suggested that this
finding may be due to the fact that firms are inappropriately generalizing from their
prior experience with the same partners. This finding also speaks to the ambiguity in
collaborating on knowledge intensive efforts like research and development.
A common measure of performance within this industry is innovation (Joly &
de Looze, 1996; Trajtenberg, 2001). Sorensen and Stuart (2000) looked at the impact
of age on ability to innovate by measuring the number of patents each firm had over
time and looking at how often they were cited. They found that older organizations
were more effective at procuring patents for their innovations, but that they were less
connected to changing environmental demands, as demonstrated by fewer citations of
those patents than younger firms. Quintana-Garcia and Banavides-Velasco (2008)
find that technologically diverse firms have a higher innovative capacity, especially in
exploratory innovation. Exploratory innovation is classified as those patents that rely
on new technology, not those that are improvements on existing technology
(Quintana-Garcia and Banavides-Velasco, 2008). Close connection to scientists and

42
the university research system has also been shown to influence the ability of firms
to innovate (Gittelman, 2006).
As can be seen from the brief review of the history and research on the
biotechnology industry, much has been learned and observed in this sector, especially
surrounding interorganizational collaboration and alliance formation. The current
study will complement this existing literature by bringing additional understanding as
to how different types of alliance experience contribute to the type of alliance a firm
will enter, as well as which types of experience best predict successful alliance
outcomes.
This particular industry is well suited to this type of study due to the volume of
relationships that have taken place over the course of the last 20 years. In the drug
discovery process, often carried out collaboratively with biotech companies and
pharmaceutical companies, the chance that a newly discovered molecule will make it
to market is less than 1% (Rothaermel, 2001). Due to the patent system in the
pharmaceutical industry and the rapid speed at which generic drugs consume the
market on expiration (Oliver, 1995), new innovations are constantly being sought out
in the form of alliances with biotechnology companies. There is optimism within the
industry that a given deal may provide firms with the opportunity to create the next
“blockbuster” (Survey: Climbing the Helical Staircase, 2003), or develop the next
“home run” patent (Owen-Smith & Powell, 2003). In this way, partnerships are
commonly used to help firms access complementary assets and stay viable (Powell, et
al. 2005; Rothaermel, 2001; Greis, Dibner & Bean, 1995).

43
Chapter 4: Methods and Analysis
The ideas and hypotheses developed above are tested on a set of product and
technology alliance relationships within the biotechnology industry. The following
sections will outline the data set, measures, and the methods employed in hypothesis
testing.
Data
The dataset used in the analysis of this study was compiled primarily from the
MedTRACK database. MedTRACK is a database assembled by Life Science
Analytics, Inc. and contains information on public and private companies within the
life sciences industry, including information on alliances, mergers and acquisitions.
Information on these business deals is collected from publicly available information
such as press releases, press coverage, industry news and SEC filings.
The dataset developed for this study was drawn initially from the deals and
alliances database. Although MedTRACK includes both private and public companies
in the United States and from around the world, the sample drawn for this study
includes only those deals that have at least one US publicly traded biotechnology
company. Such a company may be headquartered anywhere in the world, assuming it
is traded on a US exchange. This not only limits the number of deals within the
dataset to a more manageable number, but it ensures access to more complete data due
to SEC filing requirements and also allows assumptions to be made about the
guidelines with which firms are required to comply under the U.S. Food and Drug
Administration. The deals and alliances were drawn from the database from 1990-

44
2007. The initial query of the MedTRACK deals and alliances database resulted in
2,042 relationships.
This list of relationships included information about the type of relationship
that was formed, the product area which was the object of the relationship, the FDA
stage of the product or technology at the signing, as well as a short description of the
deal as described in communication surrounding the deal. Table 4.1 includes a list of
the categories into which relationships were classified. Many relationships were
classified into more than one category so for the purposes of this study, only those that
fell into one of the 3 major categories outlined in table 4.1 were included, any
“hybrid” relationships were excluded from the analysis. A hybrid was defined as any
deal that included two or more broad categories of license, R&D and non-R&D deals.
If a deal included a marketing and distribution deal, it was classified as a non-R&D
and was included as such in the study. Additional data was gathered from
MedTRACK on the companies featured in the relationships. This included
information on the clinical trial stages of their products over time, the number of
products across different indication areas over time, as well as the year the firm was
founded. Appendix A includes a sample of the information pulled for each
relationship as well as the product pipeline data for each company by indication and
clinical trial phase data for each company by year.
Data on the daily stock market returns for each firm in the dataset for the event
study portion of this research were pulled from the Center for Research in Security
Prices (CRSP) database. These data were pulled based on the date of the alliance.

45
Since the event study method uses an event window to calculate the percentage of
abnormal returns, the accuracy of the dates for the deals used in this portion of the
study was extremely important. A review of the accuracy of the dates recorded by
MedTRACK was conducted and a search was conducted for those deals with incorrect
dates to determine the correct date. Where no solid information on the relationship
could be located, the deals were excluded. An additional cleaning of the deals was
conducted based on the dates of specific deals to ensure that the event window of two
deals did not overlap. If two deals or events were found to have been announced on
the same day, these deals were eliminated. Similarly, any deal that took place during
the event window of another event was also eliminated. The number of deals was
reduced further based on the availability of stock return data from the CRSP database
for the 180 days prior to the event period. This process resulted in approximately 525
deals which could be used in the event study portion of this research.
Measures
The initial data set described above was used to calculate a variety of
experience measures. The three types identified for this study include license
agreements, non-R&D agreements and R&D agreements. Ideally the dataset would
have also included analysis of joint ventures, however, the number of JV deals in the
data set was too small to produce significance, and those deals were excluded from
analysis. License experience, non-R&D experience and R&D experience (hypotheses
1a-c, 5, and 7a-c) measures were created by summing all previous relationships of a
given type in this data set. Table 4.1 below shows each of the types that were used to

46
classify relationships in MedTRACK and which type of relationship they were
classified into for construction of the deal type measure.
For each relationship of a certain type, a one was assigned to that type. For
relationships that were classified into two different types such as a manufacturing and
sublicense agreement, then the firm was assigned a one in a hybrid license/non-R&D
agreement category. The hybrid categories were only used in the Alliance Experience
(hypotheses 6, 8, 9) variable, as it was a measure of all types of alliance experience.
The experience measures were created by summing the type classifications for all
relationships that took place prior to the date of the current relationship in the data set.
Because the data are left censored, the first five years of relationships were used as
history of experience and not included in the alliance analysis.

47

Table 4.1: Deal Type Definitions and Categorizations
MedTRACK Deal
Classifications
Definition Deal Type Category
Co-Development Two companies enter into an
agreement to develop an existing
product for market.
R&D Agreement
Development The process of undertaking
commercialization of a product.
R&D Agreement
Research Systematic investigation to establish
facts.
R&D Agreement
Co-Promotion Two or more companies join hands to
promote a product developed by the
other.
Non-R&D Agreement
Manufacturing Two companies join hands to
manufacture a new product. It is an
organized action of making of goods
and services for sale.
Non-R&D Agreement
Cross-Distribution A mutual exchange of distribution
rights between two or more
companies related with their
technology/product/services.
Non-R&D Agreement
Co-Market Two companies agree on an
exclusive or non-exclusive basis to
bring a newly developed product into
the market.
Non-R&D Agreement
Commercialization The act of bringing a product to sales. Non-R&D Agreement
Distribution The right to promote or sell a product
with in a specific region.
Non-R&D Agreement
Marketing The exchange of goods for an agreed
sum of money.
Non-R&D Agreement
Sales Income received for goods and
services over some given period of
time.
Non-R&D Agreement
Supply An amount of something available for
use.
Non-R&D Agreement
Cross-license Two parties generally exchange their
licensing rights of their
patents/products/technology to
promote their business standards.
License Agreement
License A legal document giving official
permission to do something.
License Agreement
Sublicense A license giving rights of production
or marketing of a product or service
to a person or company that is not the
primary holder of such rights.
License Agreement

48

Portfolio diversity (hypotheses 2 and 7d) measures the diversity of a firm’s
portfolio when they entered into each specific alliance relationship. Diversity was
measured following Powell et al (2005) using Blau’s (1977) heterogeneity index
where H = 1 – Σp
i
2
, where p
i
is the fraction of the population in a given group.
Hetereogeneity was calculated for two points in time, one for a firm prior to the
current deal and one for a firm after entering the current deal. So for each deal, the
firm’s percentage of deals, excluding the current deal, falling into each of the three
experience categories were squared, summed and subtracted from one. Lower values
indicate greater homogeneity. Heterogeneity was calculated again for each firm,
including the current deal. The heterogeneity level excluding the current deal was
subtracted from the heterogeneity level including the current deal to get an overall
change in diversity of the portfolio upon entering the current deal. The change in
diversity was used in the multinomial regression analysis because prior to a firm’s
decision to enter into a specific type of deal, if diversity is a consideration, then
management will consider how the current deal may influence the overall diversity of
the portfolio. Since the multinomial regression analysis seeks to capture how specific
attributes may affect a firm’s decision to enter into a new relationship, then using the
change measure is most appropriate. The OLS analysis seeks to understand how
specific attributes of the deal that has been selected translate into deal performance
suggesting that the overall diversity of the portfolio, given the relationship just

49
announced is the relevant diversity measure. Therefore the heterogeneity measure
which includes the current deal was used in the OLS analysis.
Product area experience (hypotheses 4a, b, and 9) is a variable used to
measure the extent of experience a firm has working with products in a given
therapeutic area. MedTRACK classifies each relationship as relating to one of 17
therapeutic areas, including cancer, cardiovascular systems, genetic diseases and
disorders, etc. For those relationships where a therapeutic area was not reported, the
deal summary was used to identify a therapeutic area based on a dictionary of terms
developed by a medical doctor as being relevant in each area. The MedTRACK
database also tracks how many products a firm has in each of the therapeutic areas
each year. Due to the longevity of many of these products they have a cumulative
effect in the MedTRACK report, so a product released in 1995 in the cancer area may
continue to be reported in the cancer area for 10-15 years depending on how long the
product is either in the firm’s pipeline or on the market. This tracking provides a good
indication of how experienced a firm is in a given therapeutic area in any given year.
The therapeutic area of a given deal then is used to identify the product experience for
that firm in that area in the year the deal was executed.
Product development experience (hypotheses 3 and 8) is represented by a
simple count of the number of products a company has across all therapeutic areas
during the year the alliance was entered. This experience variable is representative of
the extent to which a firm has successfully developed unique products that are either
in FDA clinical trials or are on the market.

50
Partner experience (hypotheses 5 and 10) was a variable created to measure
the extent of experience a firm may have had with a specific partner. This measure
was also created as a count of the number of times a firm had worked with this
particular partner firm in the past. Again, this measure only captures those
relationships that were included in this data set from 1990-2007.
Alliance type
The alliance type variable was determined using the same classifications from
MedTRACK and the categorizations outlined in Table 4.1 for the focal alliance
relationship. This method of measuring the dependent variable results in a 3 group
categorical variable.
Alliance performance
Measuring the performance of alliance relationships is a difficult task due to
the varying timeline on which they are expected to produce results as well as the
somewhat subjective nature of distinguishing between results produced from the
alliance and those produced from extraneous factors (Kale, Dyer & Singh, 2002). Ray
et al (2004) also note that measurement of performance of firm capabilities or
resources should be done at the process level, as opposed to an aggregated firm level.
One proxy for measuring the success of an event, such as an alliance, is to measure the
abnormal stock returns during the announcement of the event, otherwise known as the
event study method (Anand & Khanna, 2000; Haleblian & Finkelstein, 1999).
Although stock returns are a firm level performance measure, limiting the
measurement to the window in which the event was announced provides a more

51
specific measure of the alliance than an overall measure such as return on assets.
This method measures the perceived expectation of success and is based on the
efficient market hypothesis. Many scholars have used this measure to evaluate the
effectiveness of significant events, such as alliances (Koh & Venkatraman, 1991;
Anand & Khanna, 2000) and acquisitions (Hayward, 2002). Although abnormal stock
returns are a proxy for success, they have been found to be highly correlated with firm
level outcome measures such as return on assets and innovative performance (Healy,
Palepu & Ruback, 1992; Higgins & Rodriguez, 2006) as well as with managerial
assessments of firm performance (Kale et al., 2002). Abnormal stock returns are
calculated using the event study method. In order to estimate the incremental value
created for a firm, the residuals are extracted from a standard asset-pricing model that
is used to predict firms’ returns over an event period. Daily stock market return data
for each firm were used to estimate the market model (Fama, 1976) over a 180 day
period prior to the event (Kale, et al. 2002).
r
it
= α
i
+ β
i
r
mt
+ ε
it
In this market model r
it
represents the daily returns to firm i on day t. The
corresponding daily returns on the value-weighted
1
S&P 500 are denoted by r
mt
, and α
i

and β
i
are firm-specific parameters. Finally, ε
it
is independent and identically
distributed (i.i.d.) normal error. The estimates obtained from this model are then used
to predict daily returns for a firm over the event period (i.e., one day before the event
and the event day or -1 to 0), as:

1
Value-weighted: Components, or the stocks that comprise the index, are weighted according to the
market value of the outstanding shares, where a stock with a higher market capitalization will have a
larger impact on the index than a stock with a lower market capitalization.

52
R
it
= α
i
+ β
i
r
mt
Here R
it
are the predicted daily returns, and α
i
and β
i
r
mt
are model estimates.
Therefore, the daily excess returns for each firm are calculated as:
ε
it
= r
it
- R
it

where ε
it
is the percentage of daily firm-specific excess returns.
The excess returns indicate the unanticipated movements in the stock price on
a daily basis for each firm over the event period. This variable is cumulated over the
event period to create a success measure indicated by the percentage change in stock
returns as a result of the event announcement. This percentage change is then
multiplied by the market capitalization of the firm immediately prior to the event
window to create an overall wealth effect for the event.
Control variables
In the multinomial regression analysis time and company age are included as
control variables. Since the number of each type of relationship, and the market trends
for alliances may vary over time dummy variables for the years 1996-2007 are
included in the analysis to control for any time-related variations. The age of the
company in years at the time of the alliance is also included to control for variations in
the age of the company. In addition, the OLS regression analysis controls for the size
of the firm using the market capitalization of the firm prior to the event period. Table
4.2 includes a summary of all variables.

53

Table 4.2 Variable Summary

Dependent Variable Variable
Name
Description of the Variables
Alliance type DealType
"-License agreement Identified as a license agreement or a
sales or supply agreement
"-Non-R&D agreement Identified as a commercialization,
manufacturing, marketing or co-
promotion agreement
"-R&D agreement Identified as a development or research
agreement
Alliance Performance DealPerf Market capitalization of a firm multiplied
by the abnormal stock return based on
event study of the alliance
announcement
Independent Variables
License experience ExpLic A sum of all previous deals of this type
since 1990 by this company
Non-R&D experience ExpNRD A sum of all previous deals of this type
since 1990 by this company
R&D experience ExpRD A sum of all previous deals of this type
since 1990 by this company
Product area experience ExpProd The number of products in the same area
as the current deal held by the company
Total alliance
experience
ExpAlliance A sum of the total number of deals since
1990 by this company
Partner experience ExpPart A sum of all previous deals signed with
the same partner as the focal relationship
since 1990
Product development
experience
TotProd The total number of products held by this
company
Portfolio Diversity Diversity Heterogeneity of the portfolio as
represented by one minus the squared
sum of the percentage of deals within
each category
Control Variables
Firm effects Dummy for the firm represented in a
given deal
Market capitalization MarkCap Price per share times the shares
outstanding at event date -10
Time Time Dummy for the year the deal was signed
Company Age CompAge The age of the company in years the
year the deal was signed

54
Analysis: Multinomial Logistic Regression
In order to test the hypotheses developed to answer the first research question
(hypotheses 1-5), which asks which types of experience and relationship
characteristics predict relationship type of future relationships, multinomial logistic
regression was used. Multinomial logistic regression is a method meant for the
analysis of data with categorical outcome variables of more than two qualitatively
different categories (Agresti, 1998). These categories cannot be ranked or ordered,
and thus must be treated as different but equally valid outcomes. In the case of
alliance relationships, choosing the type that will best meet the goals of the
organizations involved should be the objective, rather than one type being preferred
over another.
Multinomial logistic regression seeks to explain how explanatory variables
affect the probability that a certain response category will be chosen. The multinomial
logistic probabilities of observing a given outcome category are given by Aldrich &
Nelson, (1984, p. 73):
P
ji
≡ P(Y=j|X
i
) = exp[b’
j
X
i
]/D
i
Where D
i
= ] exp(bjXi) [
1
∑
=
J
j
and b’
j
X
i
represents Σb
kj
X
ik
. The unknown parameters to
estimate are the coefficients b
j
. There are also J sets of parameters in the model and
each set contains K coefficients. Compared to the dichotomous case where there are
only K parameters to estimate, in the polytomous case there are K(J-1) parameters to
estimate and therefore the sample size needed to achieve statistical significance is
considerably higher. The suggested a minimum of number of cases ranges from 10 to

55
30 cases for each predictor variable (Pedhazur, 1997). For this study there are
approximately 20 predictor variables suggesting the need for 200-600 cases in each of
the dependent variable categories. This data set includes approximately 200-450 cases
in each category. One set of coefficients is estimated for comparison between the
referent group, usually the last group, and another outcome group. Therefore for 3
outcome variables, two sets of coefficients will be estimated. In this way, the number
of cases falling into each cell of independent variable K with each outcome variable J
needs to have a sufficient number of cases. In this case the deal type has three
potential outcomes: license, non-R&D and R&D agreements.
Analysis: Ordinary Least Squares Regression
To address the second research question (hypotheses 6-10) which looks at the
relationship between firm alliance experience and alliance success, a regression
analysis is used with the results from the event study method, described above, as the
dependent variable. The measures of experience articulated earlier are used as
predictor variables to determine the extent to which experience predicts cumulative
excess returns. In addition to using the cumulative excess returns, the wealth effects
created were also used as a dependent variable. The wealth effects variable was
created by multiplying the cumulative excess returns by the market capitalization of
the firm on day -10 relative to each event (Anand & Khanna, 2000). The stock prices
and shares outstanding used to calculate the market capitalization were retrieved from
the CRSP database. Since the deals in this study cover a 10 year timespan, the wealth

56
effects variable was also calculated taking inflation into account. The consumer
price index as reported in the CRSP database was used to bring all cumulative wealth
effects rates up to December 2007, the most recent date of any alliance included in the
study.

57
Chapter 5: Results
Descriptive statistics for each variable are included in tables 5.1 through 5.3
below. Table 5.1 includes the distribution of deals by DealType each year. As can be
seen from the table, the number of alliances in this data set increases from 1996 and
then somewhat stabilizes around 2000. The R&D agreements represented the largest
proportion of the alliances with 46%, the non-R&D alliances the least with 23% and
the license agreements make up 31%. Although more even distribution of cases
across relationship types would be favorable as it would provide more power to the
analysis of the non-R&D agreements, there are enough cases in this group, given the
number of predictor variables to run the analysis.
Table 5.1 DealType*Year Crosstabulation [Count (% within year)]

Deal Type
Year 1 2 3
1996 3 (17%) 7 (39%) 8 (44%)
1997 5 (19%) 9 (35%) 12 (46%)
1998 23 (40%) 10 (18%) 24 (24%)
1999 15 (29%) 11(21%) 26 (50%)
2000 29 (28%) 24 (23%) 50 (49%)
2001 41 (35%) 23 (20%) 54 (46%)
2002 45 (34%) 28 (21%) 58 (44%)
2003 40 (35%) 28 (24%) 48 (41%)
2004 40 (38%) 21 (20%) 44 (42%)
2005 28 (28%) 26 (26%) 48 (47%)
2006 14 (17%) 22 (27%) 47 (57%)
2007 15 (25%) 18 (30%) 28 (46%)
Total 299 (31%) 227 (23%) 447 (46%)

DealType Legend: 1= License agreement, 2=Non-R&D agreement, 3=R&D agreement

Table 5.2 includes descriptive statistics for each of the variables included in the
multinomial logistic regression analysis. Due to the large number of firms that engage
in only a few relationships, but the small number of firms that engage in a very large

58
number of relationships, the variances for the initial count variables were quite large,
and often larger than the mean number of deals. This distribution is not uncommon in
practice, however, when using multinomial logistic regression, true overdispersion can
lead to over estimated significance in results as the model assumes that the mean and
variance are equal (Agrest, 2002). A transformation may be used to solve the problem
of apparent overdispersion. In this case, the natural logs of the experience variables
solved the overdispersion problem and thus were used in the analysis. Table 5.2
reflects the descriptive statistics of the natural logs of the count variables. Table 5.3
includes descriptive statistics for the subset of alliances where the focal firm had
worked with its partner at least one time before. Each of the experience variables in
table 5.3 except the number of products, which is not specific to the individual alliance,
counts only those deals where the two firms had worked together. As a result, the
experience numbers are considerably smaller. The natural logs were used again in this
model to avoid problems of overdispersion later in the analysis.

59

Table 5.2 Descriptive Statistics for Full Multinomial Logistic Regression Model
(N=973)

Variable Mean S.D. 1 2 3 4 5 6 7 8
1. CompAge
17.15 7.87

2. ExpAlliance
2.79 1.08
.261***
3. LNExpLic
1.46 1.03
.452*** .856***
4. LNExpNRD
1.08 1.00
.546** .719*** .532***
5. LNExpRD
1.71 1.02
.394*** .905*** .738*** .600***
6. LNExpPart
0.25 0.46
-.006 .132*** .098*** -.013 .146***
7. LNExpProd 1.08 1.07 .261*** .251*** .202*** .278*** .243*** .136***
8. LNTotProd 2.77 0.86 .520*** .725*** .621*** .655*** .675*** .112*** .435***
9. ChgDiversity 0.03 0.12 -.159*** -.345*** -.288*** -.224*** -.302*** -0.061* -.086*** -.183***
*p<.1 ** p < .05 *** p < .01 **** p <
.001

60

Table 5.3. Descriptive Statistics for Multinomial Logistic Regression for Alliance between Past Partners
(N=449)

Variable Mean S.D. 1 2 3 4 5 6 7 8
1. CompAge 16.37 7.31
2. ExpAlliance 2.82 1.11 .550***
3. LNExpLic 0.17 0.33 .017 .114*
4. LNExpNRD 0.13 0.29 .186*** .072 -.147**
5. LNExpRD 0.34 0.42 -.102* -.088 -.217*** -.229***
6. LNExpPart 0.93 0.37 .117* .211*** .170*** .180*** .231***
7. LNExpProd 0.39 0.46 .119* .083 .077 .200*** .237*** .392***
8. LNTotProd 2.79 0.86 .617*** .693*** .065 .146** -.046 .236*** .126**
9. ChgDiversity 0.02 0.09 -.109* -.265*** -.038 -.044 .015 .022 -.006 -.175***
*p<.1 ** p < .05 *** p < .01 **** p <
.001

61

Two models were run in the process of testing the hypotheses proposed for research
question one (H1-H5). The first model is a baseline model with only the control
variables included. The second model includes all the experience variables. In these
models the fit statistic is a log likelihood ratio, L
2
which measures the size and
significance of the gap between the cell counts predicted by the model and the actual
cell counts observed. The L
2
is distributed as chi-square. To test whether the full
model has a better fit than the baseline model, which includes only control variables,
the Baysian Inferential Criteria (BIC) can be used (Raferty, 1995). The BIC can be
expressed by the following formula:
BIC
k
= L
k
2
– df
k
log n
where L
k
2
= χ
2
Sk
is the deviance for model Mk and df
k
is the corresponding number of
degrees of freedom. Lower BIC values indicate a better model fit and therefore can be
used to compare fit between nested models. Using this formula on the nested models
in this research the BICs have been calculated and included in table 5.4. The BIC for
the full model (BIC = -72) is lower than the BIC for the baseline model (BIC = 49)
indicating the full model is a better fit to the data. Multinomial logistic regression
estimates parameters on the outcome variables in relationship to a referent, which is
typically the last outcome variable. The model was run using R&D agreements as the
referent group. Table 5.4 includes the parameter estimates for each model.

62

Table 5.4. Estimates (log-odds and odds) from selected multinomial logistic regression models
(N = 973)

Model 1: Baseline
Model 2:
Full
Ind. Variables License vs. R&D Non-R&D vs. R&D License vs. R&D Non-R&D vs. R&D
β exp( β) β exp( β) β exp( β) β exp( β)
CompAge 0.004 0.037 ** 1.038 -0.004 0.018
ExpAlliance 0.340 0.656 * 1.927
LNExpLic 0.507 ** 1.660 -0.321 * 0.725
LNExpNRD 0.060 0.770 *** 2.160
LNExpRD -0.779 *** 0.459 -1.340 *** 0.262
LNExpPart 0.143 0.456 ** 1.578
LNExpProd -0.004 0.315 *** 1.370
LNTotProd -0.007 -0.023
ChgDiversity 2.130 ** 8.415 1.400 * 4.055
Time Fixed effects by year

Intercept -148.970 -0.110 -145.430 -2.720

Pearson Chi-square/df 1.249 1.009
Chi-square 40.600 216.110
df 26 42
BIC 49 -72
*p<.1 ** p < .05 *** p < .01 **** p < .001

63

In the baseline model it can be seen that company age (CompAge) has a small
but significant impact on the tendency to enter non-R&D agreements (b = .037, p
< .05) but not license agreements (b = .004, p > .1) when compared with entering
R&D agreements. That is, older firms are more likely to enter into non-R&D
agreements as compared to younger firms but company age alone does not increase
the odds of a firm’s likelihood of entering a license agreement. When including the
additional predictor variables CompAge is no longer significant when comparing non-
R&D (b = .018, p > .1) or license deals (b = -.004, p > .1) to R&D deals.
Hypothesis 1 a, b and c state that firms will be more likely to enter into
relationship types with which they have past experience. In the full model which
compares license agreements to R&D agreements, the coefficient for license
experience (ExpLic) is positive and significant (b = .507, p < .05). This indicates that
a firm with license agreement experience is more likely to enter into new license
agreements than into R&D agreements. Taking the exponent of the beta coefficient
indicates the actual likelihood of the event taking place (Agresti, 1998). That is, a
firm with experience with one additional license agreement is 66% more likely to
enter into new license agreements, relative to entering R&D agreements. These
results support hypothesis 1a. Similarly, in the second equation in model 2 which
compares non-R&D agreements to R&D agreements, the coefficient for non-R&D
experience (ExpNRD) is positive and significant (b= .770, p < .01), indicating that for
each additional non-R&D agreement alliance a firm is 200% or two times more likely

64
to enter into a new non-R&D agreement than an R&D agreement. These results
provide support for hypothesis 1b. Finally, the coefficients for R&D experience in
each of these models provide support for H1c. When comparing likelihood of
entering licensing agreements to R&D agreements, firms with more R&D experience
will be 55% less likely to enter into license agreements than R&D agreements (b= -
.779, p < .01). As demonstrated in equation 2 of the full model, firms with more R&D
experience will be 75% less likely to enter into non-R&D relationships than R&D
relationships (b= -1.340, p < .01). In summary, these results demonstrate that firms
with increasing experience in a certain type of relationship have a higher likelihood of
entering into relationships in which they have more experience, as opposed to
relationships with which they have less experience.
Hypothesis 2 explores the role that overall portfolio diversity plays in choosing
new relationship types, stating that increased diversity will positively contribute to the
type of relationship a firm chooses to enter. Clearly, when using a comparative model
all three types can not result in increased diversity, but due to the lack of a theoretical
basis for assuming one type would be more likely to contribute to diversity than
another type, the importance of the variable was hypothesized without identifying
preference for one type over another. The coefficient for the change in diversity in the
model comparing license agreements to R&D agreements suggests that firms that are
entering into license agreements are much more likely to increase the diversity of their
portfolio (b= 2.130, p < .05) than when entering R&D agreements. Taking the
exponent of the beta we find that firm are 800% or 8 times more likely to increase

65
portfolio diversity when entering into a new license agreement than when entering
into a new R&D agreement. The results are less pronounced in the model comparing
non-R&D agreements to R&D agreements, as the coefficient is only marginally
significant (b = 1.400, p < .1). Therefore, hypothesis 2 is marginally supported in that
increased diversity does play a large and significant role in the selection of license
agreements and a marginally significant role in the selection of new non-R&D
agreements when selecting them over R&D agreements.
Hypothesis 3 suggests that additional product development experience will
make firms more likely to enter into license agreements and non-R&D agreements
than R&D agreements. Product development experience (TotProd) does not
significantly contribute to the likelihood of entering into any particular alliance type as
neither the license agreement model coefficient (b =-.007, p > .1) nor the non-R&D
agreement model coefficient (b = -.023, p > .1) are significant. Hypothesis 3,
therefore, is not supported.
Hypothesis 4 asserts two competing hypotheses, one that suggests an
exploration strategy and the other an exploitation strategy. Hypothesis 4a asserts that
firms with more experience in a product area will be more likely to enter into R&D
agreements than non-R&D agreements and license agreements in the same product
area. Hypothesis 4b suggests that firms with greater experience in a product area will
be more likely to enter into non-R&D and license agreements than R&D agreements.
Therefore, results in support of H4a would find negative and significant coefficients
for ExpProd in both models, suggesting a decreased likelihood of selecting license and

66
non-R&D agreements compared to R&D agreements as product area experience
increases. Results in support of H4b would find a positive and significant coefficient
for ExpProd in both models, suggesting an increased likelihood of entering into non-
R&D and license agreements with more product area experience. The coefficient for
product experience is positive and significant (b = .315, p < .01) when comparing the
likelihood of entering non-R&D agreements compared to R&D agreements, but the
coefficient for comparing license agreements to R&D agreements is not significant (b
= -.004, p > .1). This indicates that for those firms that have more product area
experience, there is a significant increased likelihood of those firms entering into non-
R&D agreements relative to R&D agreements, but no significant increased likelihood
of entering license agreements over R&D agreements. These results partially support
the exploitation hypothesis 4b.
In order to test hypothesis 5 which states that firms with past alliance
experience with a partner will be more likely to enter into alliance types with which
they have experience together, a similar multinomial logistic regression model was run
using the partner specific experience variables. The baseline and the full model can be
seen in Table 5.5. As can be seen, both models are significant, however, the BIC for
the full model is considerably lower than the BIC for the baseline model, indicating
that the full model more accurately describes the data. Like hypothesis one, the
experience variable for license agreements in the first equation and the experience
variable for non-R&D agreements in the second equation are of interest in hypothesis
5. Firms that have license experience with a particular partner are more likely to enter

67
into new license agreements with that same partner, as opposed to R&D agreements
(b = 1.130, p < .05). Similarly firms that have non-R&D agreement experience with a
particular partner are more likely to enter into new non-R&D agreements with that
partner (b = 2.775, p < .01). As a result, hypothesis 5 is confirmed.

68

Table 5.5. Estimates (log-odds and odds) from selected multinomial
logistic regression models for those alliances where partner experience is one or more
(N = 276)

Model 3:
Past
Partner
Experience
Baseline
Model 4:
Past
Parnter
Experience
Full
Ind. Variables License vs. R&D Non-R&D vs. R&D License vs. R&D Non-R&D vs. R&D
β exp( β) β exp( β) β exp( β) β exp( β)
CompAge 0.002 0.050 ** 1.051 -0.013 0.013
ExpAlliance 0.371 0.045
LNExpLic 1.130 ** 3.096 -0.216
LNExpNRD 0.358 2.775 *** 16.040
LNExpRD -1.514 *** 0.220 -1.86 *** 0.156
LNExpPart -0.149 -0.127
LNExpProd 0.664 0.775
LNTotProd -0.350 0.060

Time Fixed effects by year

Intercept -4.750 -12.690 -36.560

Pearson Chi-square/df 1.050 1.068
Chi-square 47.900 *** 122.710 ***
df 24 38
BIC 19 -14
*p<.1 ** p < .05
*** p < .01

69

Results of Event Study Multivariate Regression Analysis
The event study method results for the mean cumulative excess returns across
all alliances in the sample are summarized in table 5.6. The number in the event day
column represents the number of days prior to or after the event. The date of the
actual event announcement is indicated by a zero. An event day of -10 is indicative of
10 days prior to the alliance announcement. The mean daily excess return column
indicates the percentage of excess returns for the given event day. The Patell Z test
indicates whether the percentage is significantly different from zero. The final column
shows the percentage of alliances which reported a positive abnormal return for the
given event day. The actual announcement day has a positive excess return of 1.80%
(z-statistic = 9.796).
The cumulative returns over two different event windows are summarized in
table 5.7. These two windows are both commonly used event windows in alliance
studies (Anand & Khanna, 2000; Koh & Venkatraman, 1991; Kale et al. 2002). As
can be seen, both event window periods show statistically significant excess returns
over the whole set of firms in this study (2.54% excess returns with a z-statistic of
3.323 for the -10,+3 window and 2.13% excess returns with a z-statistic of 7.654 for
the -1, 0 window). Analysis was run using both event windows, however, there was
little difference in the results. Since the smaller window has been widely used in
event studies (Koh & Venkatraman, 1991; McConnell, et al, 1985; Merchant, et al.

70
2000) and it provides a more focused look at the performance of the announcement,
it was the window used in the analysis and reported in the following results.
Table 5.6: Event Study Results
N =553

Event Day Mean Daily
Excess
Return
Patell Z
Test
% of Events with
Positive Excess
Returns
-10 -0.10% -1.121 49%
-9 0.31% 1.378 48%
-8 0.45% ** 3.083 52%
-7 -0.02% -0.701 46%
-6 -0.03% -0.454 47%
-5 0.26% 0.817 46%
-4 0.19% 1.257 50%
-3 0.08% 0.804 47%
-2 -0.12% -0.872 44%
-1 0.33% 1.029 45%
0 1.80% *** 9.796 55%
1 0.04% 0.491 46%
2 -0.31% * -1.729 47%
3 -0.34% -1.353 45%
4 -0.08% -0.172 45%
5 0.19% 1.273 49%
6 0.04% 0.331 49%
7 -0.10% -0.735 44%
8 -0.21% -0.395 48%
9 0.07% 1.054 46%
10 -0.21% -1.124 46%
' *p<.1 ** p < .05 *** p < .01

Table 5.7: Summary of Excess Returns Across Event Windows

Event
Window
Mean Daily
Excess
Return
Patell Z
Test
% of Events with
Positive Excess
Returns

(-10,+3)
2.54%*** 3.323 54%
(-1,0)
2.13%*** 7.654 57%
' *p<.1 ** p < .05 *** p < .01

71
The wealth effects for each firm event were used as the dependent variables in the
regression analysis to test hypotheses 6-10. Fixed effects were included in both
models to control for time as well as unobserved firm level heterogeneity. Due to the
large number of firms, the firm coefficients were not enumerated for ease of
presentation of the variables of interest (Anand & Khanna, 2000). Table 5.8a includes
descriptive statistics and correlations for the key variables included in the OLS
analysis.
Model 1 of the ordinary least squares analysis includes each of the baseline
experience variables discussed in this study. The overall model is significant (F =
1.448, p < .01) with an R
2
of .25. Therefore, the baseline experience variables predict
25% of the variance in performance, as measured by wealth effects. To test
hypothesis 6 that overall alliance experience will lead to higher predicted performance,
or wealth effects, the overall count of alliances was used. The coefficient for
ExpAlliance in model 1 in table 5.8b is positive and highly significant (b = 67,528.801,
p < .001). Therefore, based on model 1 alliance experience in general does increase
the wealth effects of alliance relationships when controlling for specific types of
alliance experience. Hypothesis 6 is supported.

72
Table 5.8a Descriptive Statistics for Ordinary Least Squares Models
(N=512)

Variable Mean S.D. 1 2 3 4 5 6 7 8 9 10
1 ExpAlliance 22.81 30.17
2 ExpPart 0.43 1.11 .087**
3 ExpLic 5.25 6.81 .896*** .113**
4 ExpNRD 3.04 5.42 .740*** -.036 .620***
5 ExpRD 6.86 9.67 .957*** .084* .790*** .607***
6 TotProd 19.55 20.70 .834*** .118*** .791*** .664*** .765***
7 ProdExp 4.31 7.35 .261*** .211*** .242*** .215*** .227*** .409***
8 LicDeal 0.19 0.40 -.055 -.013 .010 -.031 -.084* -.042 -.029
9 NRDDeal 0.16 0.36 -.086** -.011 -.131*** .050 -.111** -.019 .107** -.211***
10 RDDeal 0.30 0.46 .013 -.001 -.002 -.072 .060 .007 -.046 -.320*** -.280***
11 ExpRD*RDDeal 2.31 6.54 .452*** .114*** .343*** .191*** .510*** .362*** .101** -.174*** -.152*** .542***
12 ExpNRD*NRDDeal 0.57 2.24 .112** -.035 .029 .297*** .070 .109** .126*** -.125*** .594*** -.166***
13 ExpLic*LicDeal 1.05 3.57 .255*** .065 .342*** .277*** .176*** .207** .130*** .598*** -.126*** -.191***
14 ExpAlliance*TotProd 965.94 2332.73 .914*** .101** .828*** .735*** .851*** .894*** .281*** -.041 -.066 .011
15 ExpAlliance*ExpProd 156.09 529.46 .514*** .310*** .479*** .409*** .460*** .542*** .742*** .012 -.002 .005
16 CompAge 16.39 7.96 .445*** -.030 .431*** .446*** .383*** .450*** .165*** -.029 .147*** -.071
17 MarkCap 6670900 18055400 .795*** 0.065 .609*** .395*** .837*** .642*** .156*** -.091** -.078* .054

73
Table 5.8a Descriptive Statistics for Ordinary Least Squares Models (Continued)
11 12 13 14 15 16
12 ExpNRD*NRDDeal -.090**
13 ExpLic*LicDeal -.104** -.075*
14 ExpAlliance*TotProd .411*** .102** .210***
15 ExpAlliance*ExpProd .253*** .138*** .238*** .563***
16 CompAge .181*** .245*** .097** .409*** .231***
17 MarkCap .470*** .032 .032 .753*** .388*** .363***

*p<.1 ** p < .05 *** p < .01

74

Table 5.8b: Cross-sectional analysis of alliance performance

Dependent Variable-Wealth Effects (in Dollars)

Independent variable 1 2 3
Constant -294302.627 -358651.062 -150158.245
ExpAlliance
67528.801 *** 62830.081 *** 62952.941 ***
ExpPart
-29865.830 -32782.495 2012.163
ExpLic
-85200.300 *** -74738.205 *** -81262.256 ***
ExpNRD
-81777.456 *** -74032.895 *** -84112.225 ***
ExpRD
-55159.174 ** -49536.368 ** -58453.916 ***
TotProd
-4200.814 -4787.959 -11876.063 **
ExpProd
-15722.318 *** -16246.129 *** 11766.334 *
ExpLic*LicDeal
-4182.197 1271.011
ExpRD*RDDeal
9483.535 ** 41431.537 ***
ExpNRD*NRDDeal
35457.385 *** 10979.292 ***
Diversity
110564.594 122601.733
ExpAlliance*TotProd 130.717 **
ExpAlliance*ExpProd -485.891 ***
MarkCap
-0.034 *** -0.035 *** -0.036 ***
CompAge
not included in model
Time
11 years fixed effects included
FirmFix
Fixed effects

N
489 489 489
R
2

0.25 0.27 0.32
F-Value
1.448 *** 1.574 *** 1.959 ***
*p<.1 ** p < .05 *** p < .01

Hypotheses 7a-c predicted that when a firm is entering into a specific type of
alliance relationship that experience with that alliance type will result in stronger
alliance performance. In order to test this set of hypotheses interaction terms were
used between the type of relationship being entered and the experience that a firm had
with that type of relationship. These interaction terms can be seen in model 2. Model
2 is also a significant model (F = 1.574, p < .01) which accounts for a total of 27% of
the variance in wealth effects (R
2
= .27). Hypothesis 7a predicted that experience in
license agreements will increase wealth effects when a firm is entering a new license

75
agreement. The interaction term ExpLic*LicDeal, however, was not a significant
predictor of performance (b = -4,182.197, p > .1), thus, H7a is not supported. H7b
predicted that non-R&D experience would bolster alliance performance when a firm
was entering into a new non-R&D deal. The results suggest that this hypothesis is
supported, as the coefficient was positive and significant (b = 35,457.385, p < .01).
This suggests that when a firm is entering a new non-R&D deal that an additional non-
R&D deal in that firm’s alliance portfolio will increase the performance of the new
deal by almost $35,500, as measured by an increase in stock valuation of the company.
Model 2 results also show support for H7c which states that experience with R&D
deals will increase the performance of new R&D deals. The coefficient is also
positive and significant (b = 9,483.535, p < .05) indicating an additional $9,483 of
wealth effects created for each additional R&D deal a firm has in its experience
portfolio.
Hypothesis 7d also makes a prediction based on a firm’s experience portfolio.
High portfolio diversity suggests that a firm has an even distribution of different types
of alliance experience, whereas low diversity suggests that experience is skewed
toward one type. Increased diversity, however, does not result in increased
performance (b = 110,564.594, p > .1) in this context. Therefore hypothesis 7d is not
supported.
Hypothesis 8 predicted that an interaction between product development
experience and overall alliance experience would result in stronger alliance
performance. This hypothesis, along with H9 was tested using model 3 from Table

76
5.8b. The overall model is significant (F = 1.574, p < .01) and has an R
2
of .32,
indicating that the entire model accounts for 32% of the variance in wealth effects.
The coefficient for the interaction between product development experience and
overall alliance experience (TotProd*ExpAlliance) is positive and significant (b =
130.717, p < .05). Therefore, support has been found for H8.
In looking at the interaction term between alliance experience and product area
experience in model 3, a significant but negative relationship is found (b = -485.891, p
< .01). As a result, hypothesis 9 is not supported.
Finally, to test hypothesis 10, which looks at the success of those alliances in
relation to partner specific experience that took place between firms that had worked
together in the past, the OLS model was run using the partner specific alliance type
experience variables and only those deals where the partner experience was one or
more. The descriptive statistics and correlations for this model are included in table
5.9a and results of the model are included in table 5.9b below. The overall model is
significant ( F = 1.835, p < .05) and the R
2
= .26, suggesting that this model accounts
for 26% of the variance in relationship performance. As can be seen, experience with
a partner (ExpPart) only marginally influenced the wealth effects, but in a negative
direction (b = -57,485.803, p > .1). As a result, hypothesis 10 is not supported.
Table 5.10 includes a summary of all the hypothesis findings.

77

Table 5.9a Descriptive Statistics for Ordinary Least Squares Models for Deals where Firms were Partners in the Past
(N=126)

Variable Mean S.D. 1 2 3 4 5 6 7 8
1
ExpAlliance
23.48 30.71

2
ExpPart
1.78 1.63
.221**
3
ExpLic 0.17 0.38 .040 -.104
4
ExpNRD 0.19 0.47 .053 0.139 -.053
5
ExpRD 0.61 0.96 .009 .336*** -.271*** -.136
6
TotProd 20.27 19.55 .783*** .320*** -.002 .139 -.007
7
ExpProd 0.68 0.87 -.075 .309*** -.048 .208** .423*** -.010
8
MarkCap 7503175.40 20733295.44 .901*** .118 -.073 .095 .028 .677*** -.067
9
CompAge
14.98 6.97
.484*** .126 -.004 .144 .095 .492*** .239*** .485***
*p<.1 ** p < .05 *** p < .01

78
Table 5.9b: Cross-sectional analysis of alliance performance for
alliances where firms were partners in the past

Independent variable
Constant 82307.154
ExpAlliance 3563.025
ExpPart -57485.803 *
ExpLic -151855.777
ExpNRD 176465.072 *
ExpRD 91921.808 *
TotProd -3983.872
ExpProd -52229.413
MarkCap -0.010 **
CompAge 1532.585
Time
11 years fixed effects
included

N
126
R
2

0.26
F-Value
1.835 **
*p<.1 ** p < .05 *** p < .01

79

Table 5.10. Hypothesis results summary

IV DV Prediction Result
H1a LNExpLic DealType Positive relation
between experience
and likelihood of
entering same type
again
Supported
H1b LNExpNRD DealType Positive relation
between experience
and likelihood of
entering same type
again
Supported
H1c LNExpRD DealType Positive relation
between experience
and likelihood of
entering same type
again
Supported
H2 Diversity DealType Positive relation
between diversity and
DealType
Supported
H3 LNTotProd DealType Increased likelihood of
entering NRD and Lic
over R&D
Not Supported
H4a LNExpProd DealType Increased likelihood of
entering R&D over
NRD and Lic
Not Supported
H4b LNExpProd DealType Increased likelihood of
entering NRD and Lic
over R&D
Partially
Supported
H5 LNExpPart
(for ExpPart >0 only)
DealType Positive relation
between experience
with a partner and
likelihood of entering
again with same
partner
Supported
H6 ExpAlliance DealPerf Positive Supported
H7a ExpLic*LicDeal DealPerf Positive Not Supported
H7b ExpNRD*NRDDeal DealPerf Positive Supported
H7c ExpRD*RDDeal DealPerf Positive Supported
H7d Diversity DealPerf Positive Not Supported
H8 ExpAlliance*TotProd DealPerf Positive Supported
H9 ExpAlliance*ExpProd DealPerf Positive Not Supported
H10 ExpPart
(for ExpPart >0 only)
DealPerf Positive Not Supported

80
Chapter 6: Discussion & Conclusion
The resource-based perspective asserts two things about a firm and its
resources. First, it suggests that firms will make strategic decisions which
capitalize on its capabilities, and second that utilization of capabilities or resources
will result in increased rents or abnormal returns for that firm. This research has
also emphasized the importance in recognizing that these capabilities are dynamic,
constantly changing, (Ray, et al. 2004) and often work in concert with one another
(Helfat, 2000), both to influence decisions and ultimately performance. This
research, therefore, sought to explore several different capabilities to better
understand how they might collectively contribute to strategic decision making
and performance. The capabilities explored included two aspects of the alliance
capability, the alliance type capability and the partner capability, as well as two
different types of technological capabilities, specifically overall product
development capability and product area capability. Each capability was tested for
its impact on a firm’s likelihood to enter into a new alliance of a certain type,
including license agreements, non-R&D agreements and R&D agreements. This
tested the extent to which firms appear to be utilizing capabilities in strategic
alliance decisions. In addition, the performance of each alliance was evaluated
based on the extent of experience in each capability area by using cumulative
percentage of excess stock returns to derive a measure, wealth creation, gained
through each transaction. This tested the extent to which capabilities contribute to
increased returns or competitive advantage for firms.

81
Past experience with a certain interaction type was found to significantly
influence a firm’s likelihood to enter into additional alliance relationships of that
specific type. Those firms with more license experience were more likely to enter
into new license agreements than into R&D agreements, and firms with more non-
R&D experience were more likely to enter into new non-R&D agreements than
into R&D agreements. Firms with more R&D experience were less likely to enter
into non-R&D agreements and license agreements than R&D agreements. These
findings support the idea that firm members see past experience with a given
alliance type as a significant part of the firm level alliance capability, as it is a
contributing factor to the likelihood of entering into each type again in the future.
In this way, the idea of an interaction capability is supported in that firms will
choose to enter into new relationships where it seems that resources gained from
similar past relationships will provide an advantage.
The idea that firms have a preference for diversity in the relationships that
they choose to enter is also confirmed in this research. Diversity is defined in
terms of the variety of experience with different alliance types. Specifically it was
found that as firms enter new license agreements and non-R&D agreements, it is
more likely that those types will increase their portfolio diversity than if they were
to enter into R&D relationships. This finding also suggests a potential trajectory
for how firms venture into the alliance space. Because of the way that diversity is
measured the likelihood that it will increase with new relationships is extremely
high in early decisions. For instance, if a firm has only one license agreement in

82
its experience portfolio, its diversity measure will be very low because 100% of
its deals fall into a single category. Selection of a second deal in either of the other
two categories will immediately double the diversity score, whereas selection of a
second deal in the same category as the first will result in an unchanged diversity
score. If a firm has a larger number of relationships, more evenly distributed
across the categories, and thus a higher overall diversity score, addition of a new
deal to any of the categories will result in a relatively smaller change in diversity
than for the firm with only one prior deal. This suggests that those deal types that
were found to increase diversity are likely those that cause the greatest jump in the
score, likely earlier in a firm’s alliance experience. In this case, the fact that
license and non-R&D deals are more likely to increase portfolio diversity suggests
that many companies begin their alliance experience with R&D deals, followed by
licensing agreements and finally by non-R&D deals. This finding makes sense if
it is put into the product development context, where biotechnology firms are
known for starting off as small research labs that often seek out alliance
relationships as a way to finance research activities. As products are developed,
the need for financing decreases, but the need for access to marketing expertise, or
distribution channels increases. In this way, the preference for diversity may have
as much to do with the product development cycle as it does with the desire to gain
a more well rounded alliance capability.
The specific product development capability tested in this research was not
found to contribute to the likelihood of a firm choosing a particular type of

83
relationship. When comparing license agreements and non-R&D agreements to
the likelihood of entering R&D agreements, product development capability had
no significant impact on type choice. It is possible that firms with high
technological capability of product development is less an indicator that they will
choose a specific type of relationship and more likely evidence that the proficiency
in developing products indicates the ability of a firm to correctly choose the most
appropriate type of alliance, as opposed to any one particular type, thus accounting
for the lack of significance in this analysis.
How a firm chooses to employ its product area capability in terms of
alliance type choice seems to suggest a decision to either exploit the capability, by
utilizing the gains from previous experience, or to explore within the capability by
seeking to expand knowledge in that particular area. Findings from this research
suggest the exploitation strategy, as experience in a product area contributes to the
likelihood of a firm entering into non-R&D agreements. This was found to be
significant when comparing non-R&D agreements to R&D agreements. Although
license agreements are another potential way in which the product area capability
might be exploited, this finding was not significant. Although it is not entirely
clear why there is a significant difference only in the likelihood to enter non-R&D
agreements this may have to do with the extent to which the product area
capability applies to functional areas. For instance, maybe a manufacturing firm
continues to enter into non-R&D manufacturing agreements with companies that
develop heart valve products because of its manufacturing expertise in the

84
materials necessary to make heart valves. In a similar way, firms that have
experience marketing diabetes products have an existing technological capability
in terms of the client base to which they market, as well as knowledge of the issues
relevant to the consumers in the diabetes market. In this way, the product area
capability may have more to do with having the necessary functional knowledge
than having the scientific knowledge to continue working in the same product area.
One final capability type explored in this study was the role of past partner
experience. The tendency for firms to enter into relationships types with which
they have past experience with partners with which they have worked before was
confirmed. In other words, when a firm has license experience with a particular
partner, it has a higher likelihood of entering into new license relationships with
that same partner. The same is also true of non-R&D agreements and R&D
agreements. It is also interesting to note that in comparing the magnitude of these
coefficients to that of the larger model, the effects of past type experience on the
likelihood of entering into the same type again in the future were considerably
higher among those firms that had worked together before. This suggests that in
addition to what is learned about interaction for certain alliance types, that
knowledge or capability is more highly valued when applying it again to the same
partner.
The analysis of the relationship between capabilities and alliance
performance reveals some interesting dynamics. At the baseline when only
considering pure experience counts, only general alliance experience has a positive

85
relationship with performance of the deal. All other types of experience,
including experience with a partner, experience with license agreements, R&D
agreements or non-R&D agreements, and experience in a specific product area,
were all negatively related with performance. These findings seem to support the
fact that simple measures of capabilities are insufficient (Priem & Butler, 2001a;
Newbert, 2007). Further analysis supports the idea that capabilities are complex,
dynamic work together in unique ways to create value for a firm (Eisenhardt &
Martin, 2000). The interaction between non-R&D deals and non-R&D experience
produced a significant and positive relationship to performance. R&D experience
was also a positive predictor of performance when only R&D deals were being
considered. In this way, the resources, or capabilities are only valuable to a firm
when the firm has the opportunity to utilize the capabilities (Barney, 1991). It also
suggests that there are distinct skills, routines and processes unique to each type of
alliance relationship. Experience with license agreements was not found to
contribute to the performance of new license agreements. This confirms Anand
and Khanna’s (2000) finding, and also expands it to suggest that other types of
relationship performance can be improved with experience with those types.
Portfolio diversity was not found to be a significant predictor of alliance
performance. This result seems to support the rationale provided above that the
tendency toward diversity is more a function of the product development life cycle
and less an intentional effort by managers to achieve a balanced or diverse
portfolio of relationship types.

86
The interaction of product development experience with overall alliance
experience was found to positively contribute to wealth creation of new alliance
relationships. Although product development experience alone does not contribute
to alliance wealth creation, it is understandable that there may be some synergy
between alliance experience and product development experience. Firms that
simply have a long history of successful product development may have been
successful largely due to a specific technological advantage. When alliance
experience is interacted with product development experience, however, an
indicator of how heavily a firm may have relied on alliance relationships to secure
necessary resources to successfully create those products. In this way, the findings
suggest that over time firms may develop the capability to successfully choose
which types of resources they should secure to develop products through specific
types of alliance relationships.
The same dynamic was not found to be true for a firm’s experience in a
specific product area. The interaction of product area with alliance experience
actually resulted in negative wealth creation for new alliance relationships.
Although the reasoning for this finding is somewhat unclear, perhaps it has to do
with the depth of knowledge that is required with additional work in a specific
product area. As a firm gains more experience in a specific product area, the firm
knowledge become more tacit, and more heavily ingrained in the specific routines
of that organization. Unless there is also a specific mechanism for learning or

87
codifying the tacit knowledge and routines within the organization, additional
alliance experience may not provide additional value to the firm.
The test of performance for firms with repeat partners also finds a
marginally significant negative impact. These two processes, increased partner
experience and increased product area experience may be relatively similar. The
knowledge that is gained within that particular area of expertise, be it a partner, or
a product area, may not be easily transferred to other areas or partners. Another
explanation may be the issue of legitimacy. Research in this field, as well as in
other fields has shown that initial partnerships act to validate a firm, to provide it
with legitimacy within the marketplace (Baum & Oliver, 1991; Kim & Higgins,
2007; Higgins & Gulati, 2003; Stuart, et al. 1999). For instance, Stuart, et al.
(1999) found that biotechnology companies that had prominent alliance partners
went to initial public offering more quickly and received higher valuations than
firms with no such alliance partners. Stuart et al, (2007) found that biotechnology
firms with in-licensing agreements are more likely to attract downstream alliances
that generate revenue than firms with fewer in-licensing agreements. Nicholson et
al. (2005) also found that firms are more likely to earn increased valuations
through IPO as well as to secure more funding from venture capital organizations
as a result of alliance partners. It may be that this dynamic is manifesting itself
through the partner and product area measures in that initial announcements of
collaboration for a specific product area, or with a specific partner provide the firm
its rewards, such that subsequent experience has relatively less value. In summary,

88
the ideas of the resource based view are supported by these findings,
demonstrating a tendency to make strategic decisions based on experiences or
capabilities and also that certain types of experience lead to increased performance,
even when considering a wide range of a firm’s experiences.
Implications for the Resource Based View
This study represents a modest attempt to build on our understanding of the
value of the resource based view by adding to the literature in several ways. First,
this study sought to use the research literature on alliances to independently
develop a set of capabilities that might contribute to strategic decision making as
well as alliance performance as called for by Dutta, Narasihman & Rajiv (2005).
Although many of the capability measurements used in this study have been used
in previous research, such as past partners (Gulati, 1995) and alliance type
experience (Anand & Khanna, 2000), to my knowledge no existing study has
attempted to simultaneously capture each of these capability types in the same
study. Secondly, RBV scholars have noted that capabilities are often highly
complex and the ways in which they interact, ambiguous. Ray et al (2004)
asserted that many resources or capabilities may indeed be bundled together in
order to handle one business process. Strategic alliances may be thought of as a
business process in that they are highly complex and ambiguous. Likely because
of this, current research has had difficulty identifying ways in which firm
resources can work together to produce successful relationships. In this way, this
study used several indicators to try to capture the ways in which firm resources can

89
work together to produce successful relationships by including all measures in
the same models, thus controlling for a wide variety of firm experience. The
interaction of these variables also demonstrated the ways in which controlling for
other types of experience can be insufficient, but that taking the specific strategic
decision into account, as in the case of the interactions between R&D experience
and R&D deals, or viewing product development experience through the lens of
alliance experience can produce fruitful insights into how capabilities can produce
synergies. Finally, by taking several years of data and calculating firm experience
levels for each relationship and then controlling for firm effects over the course of
time, this study captured, at least to some extent, the dynamic changing nature of a
firm’s capability over time. This study, therefore, has provided preliminary
validation for several of the ideas recently proposed in the research literature,
including the importance of viewing capabilities as dynamic (Eisenhardt & Martin,
2000), interactive (Helfat, 2000) and often requiring bundling to handle a single
business process (Ray et al, 2004).
Implications for Alliance Capability Research
One relatively unique aspect of this study in terms of the alliance capability
research is its attempt to understand both how prior experience contributes to the
decision to enter into new relationship as well as how that experience contributes
to the deal level outcomes. The level of specificity in terms of relationship types is
also unique to this study. Anand & Khanna (2000) examine license agreement
experience, but not in connection with other types of relationships, and only in

90
terms of performance. Wang & Zajac (2006) look at the decision to enter into
alliances or to make an acquisition, and they also consider the partner, but they do
not distinguish the type of alliance relationship. This study therefore expands
upon the current literature by providing some information about the role that R&D
agreements and non-R&D agreements play in capabilities development. Both
these types of relationships were found to positively contribute to performance
when entering into new relationships of that type, whereas the type experience in
general had a negative effect. Such a finding suggests that there are distinct skills,
routines and processes that are involved in each of the different types of alliance
relationships. The interaction of alliance experience with product development
experience also suggests that other firm level competencies and factors need to be
considered as part of the overall alliance capability, as it was found to positively
impact wealth creation for the specific relationship. In this way, this research
validates the need for the alliance capabilities research to cast a broad net when
working to understand more about the factors that work together to create
successful alliance relationships.
Limitations and Opportunities for Future Research
Several limitations in this study create opportunities for future research.
First, this study only measured capability in terms of a count variable. This
unilateral form of measurement presented difficulties with multicollinearity among
the predictors. Work by Kale et al. (2002) and Mayer and Salomon (2006) also
incorporated capabilities measures based on internal evaluation by firm members.

91
Resource-based view scholars have also advocated for the inclusion of mixed
methods to create measures of capabilities. Rouse and Daellenbach (1999) suggest
including things like organizational culture in consideration of firm level
capabilities and resources. Armstrong and Shimizu (2007) also recommend using
surveys and field research. Although this study did seek to overcome some of
these weaknesses by using a wide variety of capability measures, more qualitative
data would have provided a more rich understanding of the findings.
In addition, this study only looked at alliance relationships from the
perspective of one of the alliance partners. Although partner attributes were
somewhat captured in the experience variables as well as in the product area
variable, there was not full information for both partners. As suggested by Wang
and Zajac (2006) considering alliance decisions from a dyadic perspective may
have added a level of richness to the data which would have provided a more
holistic understanding of the results. Specifically in this study it may have shown
how two firms capitalize on both their experiences when choosing a new type of
relationship. This study was also limited to relationships from a single industry
making results less generalizable to other industries with different industry
dynamics. The lack of joint ventures in this research also made for a more limited
analysis of the different relationship types.
Finally, the performance measure in this study included only the event
study method over a short window. This method provides only a short term,
outsider’s perspective on performance, despite the fact that it has been found to

92
correlate with other performance indicators (Healy, Palepu & Ruback, 1992;
Higgins & Rodriguez, 2006; Kale, et al. 2002). This measure coupled with
managers and industry experts’ post hoc assessment of the success of the alliance
relationship would provide a better picture of whether the relationship was
successful and which aspects of the organizations’ capabilities may have
contributed to that success.
Future work in this area should work to combine the unique aspects of all
work done in the area of alliance capabilities. Incorporation of internal data as
used by Kale et al. (2002), data from both firms (Wang & Zajac, 2006) on a
variety of capabilities measures all tested simultaneously as was done in this study
will likely provide a more complete and thus productive picture of how firm
capabilities contribute to alliance decisions and outcomes. Additionally, in an
effort to further explore the relationship between internal and external capabilities,
future work should look at a wider range of variables that may explain the
differences between what predicts a decision and what predicts its performance in
the market.
Role of Communication in Alliance Capability Development
Inherent in the idea of an alliance capability and subsequent alliance
performance is the issue of communication. Many, if not all, of the processes,
routines and skills that are required to execute operational capabilities or dynamic
capabilities (Teece, et al, 1997) are created, enacted and disseminated through
communication. The work on the tacitness of knowledge alludes to the difficulty

93
that often accompanies communication of routines and competencies (Kogut &
Zander, 1992). Alliances in particular have added communication challenges in
that communication is supposed to happen across organizational levels as well as
across different organizational cultures. In addition, an extensive literature on
negotiation which deals with how two individuals can effectively communicate to
reach an agreement is somehow divorced from the literature on how a firm can
effectively execute an agreement. In this way, there are several areas of existing
research that might be brought together to shed light on the way in which
communication, both in the planning and executing stages may positively impact
the success of alliance relationships.
In addition, current work, including this research study, often fails to
examine the actual routines and processes that are involved in the creation and
execution of an alliance relationship. Some work has been done at the case study
level that exposes communication issues as sources of contention (Doz, 1996, for
example), yet to my knowledge, no systematic work has been done on this issue
within the context of alliance relationships. Kale et al (2002) conducted a study
using the dedicated alliance function as a variable in predicting the outcome of
alliance relationships and found it to have a positive effect. This suggests the
validity of examining the actual routines enacted by the individual within that
dedicated alliance function to understand the way in which experience and lessons
learned were applied to new relationships. Again, however, until such work is
actually done, it is likely that the alliance capability will remain a black box.

94
Future work in this area might borrow methods from some other
disciplines. Discourse analysis may be one valuable methodological approach that
would supplement the study of alliance capabilities. One might look at the extent
to which the discourse used to discuss alliance relationships recognizes the various
roles that different departments and individuals will play in the execution of an
alliance relationship. It might also be fruitful to see whether there is a discursive
process that captures learning from past relationships and incorporates them into
the discussions of new relationships. Capturing difference in discourse related to
an existing alliance relationship and comparing how those differences affect the
overall success of the relationship may shed light on areas of organizations that are
routinely left out of the communication process, or whose opinion is not sought,
but should be. Another methodological approach might be experimental methods,
as often used in small group research and in psychology research. For instance,
one might use an experiment to explore how well individuals with different areas
of expertise are able to share relevant aspects of that expertise in the context of
planning for new alliance relationships. This might be accomplished by assigning
subjects to a functional area, like finance, communication, operations,
manufacturing, and research. Subjects could then be supplied with a profile of
their expertise as someone who works in that role. Then armed with that
information they could be assigned a task to negotiate a new alliance relationship.
Analysis of how much of each professional’s profile information was entered into
the discussion could provide a gage as to how well individuals are able to apply

95
their expertise to the alliance situation and whether or not a thorough agreement
might be reached.
Conclusion
This research sought to understand the extent to which alliance and
technological capabilities contribute to a firm’s likelihood of entering into certain
alliance types, as well as whether those same capabilities contributed to the wealth
creation earned upon announcement of the alliance decision. Interaction, partner
and product area capabilities were found to contribute to the likelihood of entering
into specific types of new relationships, however, the product development
capability did not. Similar results were revealed in the test of capability influence
on wealth creation, such that general alliance experience, the specific alliance type
experiences of non-R&D and R&D agreements, and the interaction of general
alliance experience with total product development experience were all positively
related to increased alliance wealth creation.
This study has made a contribution to our understanding of how alliances
function in the organizational context. As can be seen from the findings presented
here, as well as the foregoing discussion of limitations and areas for future
research, the dynamics involved in capabilities research, and alliance research in
particular, are complex and changing. In this way, it is hoped that this study
represents a call for continued work in the area of alliance capability development
and has perhaps inspired new ways of thinking about and approaching capabilities
research.

96
References
Agresti, A. (1998). An introduction to categorical data analysis (2nd ed.).
Hoboken, NJ: John Wiley & Sons.

Agresti, A. (2002). Categorical data analysis (2
nd
ed.). Hoboken, NJ: John Wiley
& Sons.

Alster, N. (1994). Natural partners. Forbes, 153, 76-78.

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

Argote, L. (1999). Organizational Learning: Creating, Retaining and
Transferring Knowledge. New York: Springer.

Arino, A., & de la Torre, J. (1998). Learning from failure: Towards an
evolutionary model of collaborative ventures. Organization Science, 9(3),
306-325.

Armstrong, C.E. & Shimizu, K. (2007). A review of approaches to empirical
research on the resource-based view of the firm. Journal of Management,
33, 959-986.

Arya, B. & Lin, Z. (2007). Understanding collaboration outcomes from an
extended resource-based view perspective: The roles of organizational
characteristics, partner attributes and network structures. Journal of
Management, 33, 697-723.

Barney, J. (1991). Firm resources and sustained competitive advantage. Journal of
Management, 17, 99-120.

Barney, J. B. (2001). Is the resource-based “view” a useful perspective for
strategic management research? Yes. Academy of Management Review, 26,
41-56.

Barney, J. B., & Hansen, M. H. (1994). Trustworthiness as a source of competitive
advantage. Strategic Management Journal, 15, 175-190.

Barney, J.B., Wright, M, & Ketchen, D.J. (2001). The resource-based view of the
firm: Ten years after 1991. Journal of Management, 27, 625-641.

Baum, J. A. C. & Oliver, C. (1991). Institutional linkages and organizational
mortality. Administrative Science Quarterly, 36, 187-218.

97
B. I. O. (2008). Biotechnology industry facts. Retrieved Mar. 17, 2008, from
http://bio.org/speeches/pubs/er/statistics.asp?p=yes

Blau, P. (1977). Inequality and heterogeneity: A primitive theory of social
structure. New York Free Press.

Burt, R. S. (1995). Structural Holes: The social structure of competition (2
nd
ed.)
Cambridge, MA: Harvard University Press.

Bylinsky, G. (1994). Genetics: The money rush is on. Fortune, 129, 94-104.

Castells, M. (2000). The rise of the network society (2nd ed.). Oxford: Blackwell.

Child, J. (Ed.). (2001). Learning through strategic alliances. New York: Oxford
University Press.

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

Cohen, M. D. & Bacdayan, P. (1994). Organizational routines are stored as
procedural memory: Evidence from a laboratory study. Organization
Science, 5, 554-568.

Cohen, W. M., & Levinthal, D. A. (1990). Absorptive capacity: A new perspective
on learning and innovation. Administrative Science Quarterly, 35, 128-152.

Combs, J.G. & Ketchen, D.J. (1999). Explaining interfirm cooperation and
performance: Toward a reconciliation of predictions from the resource-
based view and organizational economics. Strategic Management Journal,
20, 867-888.

Contractor, F. J., & Lorange, P. (1988). Why should firms cooperate? The
strategy and economics basis for cooperative ventures. In F. J. Contractor
& P. Lorange (Eds.), Cooperative strategies in international business (pp.
3-30). Lexington, MA: Lexington Books.

Crocker, K., & Masten, S. E. (1988). Mitigating contractual hazards: Unilateral
options and contract length. The RAND Journal of Economics, 19, 327-343.

Daft, R. (1983). Organization theory and design. New York: West.

98
Deeds, D. L. & Hill, C. W. L. (1996). Strategic alliances and the rate of new
product development: An empirical study of entrepreneurial biotechnology
firms. Journal of Business Venturing, 11, 41-55.

Doz, Y. L. (1996). The evolution of cooperation in strategic alliances: Initial
conditions or learning processes? Strategic Management Journal, 17, 55-
83.

Dutta, S., Narasimhan, O. & Rajiv, S. (2005). Conceptualizing and measuring
capabilities: Methodology and empirical application. Strategic
Management Journal, 26, 277-285.

Duysters, G., Kok, G., & Vaandrager, M. (1999). Crafting successful strategic
technology partnerships. R&D Management, 29(4), 343-351.

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

Fama, E. F. (1976). Foundations of Finance. New York: Basic Books.

Gilsing, V., & Notteboom, B. (2006). Exploration and exploitation in innovation
systems: The case of pharmaceutical biotechnology. Research Policy, 35,
1-23.

Gittelman, M. (2006). National institutions, public-private knowledge flows, and
innovation performance: A comparative study of the biotechnology
industry in the US and France. Research Policy, 35, 1052-1068.

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

Grant, R. M. (1996). Toward a knowledge-base theory of the firm. Strategic
Management Journal, 17, 109-122.

Greis, N. P., Dibner, M. D., & Bean, A. S. (1995). External partnering as a
response to innovation barriers and global competition in biotechnology.
Research Policy, 24, 609-630.

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

99
Gulati, R. (1995b). Social structure and alliance formation patterns: A
longitudinal analysis., Administrative Science Quarterly, 40, 619-652..

Gulati, R. (1998). Alliances and networks. Strategic Management Journal, 19(4),
293-317.

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

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

Gulati, R., & Westphal, J. D. (1999). Cooperative or controlling? The effects of
ceo-board relations and the content of interlocks on the formation of joint
ventures. Administrative Science Quarterly, 44(3), 473-506.

Hagedoorn, J. (1993). Understanding the rationale of strategic technology
partnering: Interorganizational modes of cooperation and sectoral
differences. Strategic Management Journal, 14, 371-385.

Haleblian, J., & Finkelstein, S. (1999). The influence of organizational acquisition
experience on acquisition performance: A behavioral learning perspective.
Administrative Science Quarterly, 44(1), 29-56.

Hayward, M. L. A. (2002). When do firms learn from their acquisition experience?
Evidence form 1990-1995. Strategic Management Journal, 23(1), 21-39.

Healey, P. M., Palepu, K. G., & Ruback, R. (1992). Does corporate performance
improve after mergers? Journal of Finance Economics, 31, 135-175.

Helfat, C.E. (2000). Guest editor’s introduction to the special issue: The evolution
of firm capabilities. Strategic Management Journal, 21, 955-959.

Helfat, C.E. & Peteraf, M.A. (2003). The dynamic resource-based view:
Capability lifecycles. Strategic Management Journal, 24, 997-1010.

Higgins, M. C., & Gulati, R. (2003). Getting off to a good start: The effects of
upper echelon affiliations on underwriter prestige. Organization Science,
14(3), 244-263.

Higgins, M. J. & Rodriguez, D. (2006). The outsourcing of R&D through
acquisitions in the pharmaceutical industry. Journal of Financial
Economics, 80, 351-383.

100
Hoang, H., & Rothaermel, F. T. (2005). The effect of general and partner-
specific alliance experience on joint R&D project performance. Academy
of Management Journal, 48(2), 332-345.

Hopkins, M. M., Martin, P. A., Nightingale, P., Kraft, A., & Mahdi, S. (2007). The
myth of the biotech revolution: An assessment of technological, clinical
and organizational change. Research Policy, 36, 566-589.

Hoopes, D.G., Madsen, T.L, & Walker, G. (2003). Guest editors’ introduction to
the special issue: Why is there a resource-based view” Toward a theory of
competitive heterogeneity. Strategic Management Journal, 24, 879-902.

Inkpen, A. C., & Dinur, A. (1998). Knowledge management processes and
international joint ventures. Organization Science, 9(4), 454-468.

Joly, P. B., & de Looze, M. A. (1996). An analysis of innovation strategies and
industrial differentiation through patent applications: The case of plant
biotechnology. Research Policy, 25, 1027-1046.

Joskow, P. L. (1987). Contract duration and relationship specific investments:
Empirical evidence from coal markets. American Economic Review, 77,
168-185.

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

Kale, P., & Singh, H. (2007). Building firm capabilities through learning: The role
of the alliance learning process in alliance capability and firm-level
alliance success. Strategic Management Journal, 28, 981-1000.

Kalnins, A., & Mayer, K. J. (2004). Relationships and hybrid contracts: An
analysis of contract choice in information technology. Journal of Law,
Economics and Organization, 20, 207-229.

Khanna, T., Gulati, R., & Nohria, N. (1998). The dynamics of learning alliances:
Competition, cooperation and relative scope. Strategic Management
Journal, 19, 193-210.

Kim, J. W., & Higgins, M. C. (2007). Where do alliances come from? The effects
of upper echelons on alliance formation. Research Policy, 36, 499-514.

Kogut, B. (1988). Joint ventures: Theoretical and empirical perspectives. Strategic
Management Journal, 9, 319-332.

101

Kogut, B. (1989). The stability of joint ventures: Reciprocity and competitive
rivalry. Journal of Industrial Economics, 38, 183-198.

Kogut, B., & Zander, U. (1992). Knowledge of the firm, combinative capabilities,
and the replication of technology. Organization Science, 3(3), 383-397.

Kogut, B., & Zander, U. (1996). What firms do? Coordination, identity, and
learning. Organization Science, 7, 502-518.

Koh, J., & Venkatraman, N. (1991). Joint venture formations and stock market
reactions: An assessment in the information technology sector. Academy of
Management Journal, 34(4), 869-892.

Li, S., & Berta, W. B. (2002). The ties that bind: Strategic actions and status
structure in the us investment banking industry. Organization Science,
19(339-368).

Liebeskind, J. P., Oliver, A. L., Zucker, L., & Brewer, M. (1996). Social networks,
learning and flexibility: Sourcing scientific knowledge in new
biotechnology firms. Organization Science, 7(4), 428-443.

Lorenzoni, G., & Lipparini, A. (1999). The leveraging of interfirm relationships as
a distinctive organizational capability: A longitudinal study. Strategic
Management Journal, 20(4), 317-338.

Madhok, A. & Tallman, S.B. (1998). Resources, transactions and rents: Managing
value through interfirm collaborative relationships. Organization Science,
9, 326-339.

Mahoney, J. T., & Pandian, J. R. (1992). The resource-based view within the
conversation of strategic management. Strategic Management Journal, 13,
363-380.

Makadok, R. (2001). A pointed commentary on Priem and Butler. The Academy
of Management Review, 26, 498-505.

Mayer, K., & Argyres, N. (2004). Learning to contract: Evidence from the
personal computer industry. Organization Science, 15, 394-410.

Mayer, K. J., & Salomon, R. (2006). Capabilities, contractual hazard and
governance: Integrating resource-based and transaction cost perspectives.
Academy of Management Journal, 49, 942-959.

102
McNamara, P., & Baden-Fuller, C. (2007). Schareholder returns and the
exploration-exploitation dilemma: R&D announcements by biotechnology
firms. Research Policy, 36, 548-565.

McKelvey, M., & Orseningo, L. (2004). Pharmaceuticals as a sectoral innovation
system. In F. Malerba (Ed.), Sectoral systems of innovation. Cambridge,
MA: Cambridge University Press.

Mitchel, M. L. (1989). The impact of external parties on brand-name capital: The
1982 Tylenol poisonings and subsequent cases. Economic Inquiry, 27, 601-
619.

Mowery, D. C., Oxley, J., & Silverman, B. S. (1996). Strategic alliances and
interfirm knowledge transfer. Strategic Management Journal, 17, 77-91.

Nelson, R., & Winter, S. (1982). An evolutionary theory of economic change.
Cambridge, MA: Harvard University Press.

Newbert, S.L. (2007). Empirical research on the resource-based view of the firm:
An assessment and suggestions for future research. Strategic Management
Journal, 28, 121-146.

Nicholson, S., Danzon, P. M., & McCullough, J. (2005). Biotech-pharmaceutical
alliances as a signal of asset and firm quality. The Journal of Business,
78(4), 1433-1464.

Oliver, S. (1995). Make a good product. Forbes, 156, 90-91.

Owen-Smith, J., & Powell, W. W. (2003). The expanding role of university
patenting in the life sciences: Assessing the importance of experience and
connectivity. Research Policy, 32, 1695-1711.

Park, N. K., Mezias, J. M., & Song, J. (2004). A resource-based view of strategic
alliances and firm value in electronic marketplace. Journal of Management,
30(1), 7-27.

Pedhazur, E.J. (1997). Multiple regression in behavioral research (3
rd
ed.).
Orlando, FL: Harcourt Brace.

Penrose, E. (1959). The theory of the growth of the firm. New York: John Wiley.

Pisano, G. P. (1991). The governance of innovation: Vertical integration and
collaborative arrangements in the biotechnology industry. Research Policy,
20, 237-249.

103

Powell, W. W., White, D. R., Koput, K. W., & Owen-Smith, J. (2005). Network
dynamics and field evolution: The growth of interorganizational
collaboration in the life sciences. American Journal of Sociology, 110(4),
1132–1205.

Priem, R.L. & Butler, J.E. (2001a). Is the resource-based “view” a useful
perspective for strategic management research? Academy of Management
Review, 26, 22-40.

Priem, R.L. & Butler, J.E. (2001b). Tautology in the resource-based view and the
implications of externally determined resource value: Further comments.
Academy of Management Review, 26, 57-66.

Provan, K. G. (1982). Interorganizational linkages and influence over decision
making. The Academy of Management Journal, 29(2), 443-451.

Quintana-Garcia, C., & Benavides-Velasco, C. A. (2008). Innovative competence,
exploration and exploitation: The influence of technology diversification.
Research Policy, In press.

Raferty, A. E. (1995). Bayesian model selection in social research. Sociological
Methodology, 25, 111-163.

Ray, G., Barney, J. B. & Muhanna, W.A. (2004). Capabilities, business processes,
and competitive advantage: Choosing the dependent variable in empirical
tests of the resource-based view. Strategic Management Review, 25, 23-37.

Reagans, R., & McEvily, B. (2003). Network structure and knowledge transfer:
The effects of cohesion and range., Administrative Science Quarterly, 48,
240-267.

Robinson, D. T., & Stuart, T. E. (2001). Just how incomplete are incomplete
contracts? Evidence from biotech strategic alliances, Columbia Law
Seminar: NYU/Columbia Law Seminar.

Roijakkers, N., & Hagedoorn, J. (2006). Inter-firm R&D partnering in
pharmaceutical biotechnology since 1975: Trends, patterns and networks.
Research Policy, 35, 431-446.

Roijakkers, N., Hagedoorn, J., & van Kraneburg, H. (2005). Dual market
structures and the likelihood of repeated ties-evidence from pharmaceutical
biotechnology. Research Policy, 34, 235-245.

104
Rothaermel, F. T. (2001). Complementary assets, strategic alliances, and the
incumbent's advantage: An empirical study of industry and firm effects in
the biopharmaceutical industry. Research Policy, 30, 1235-1251.

Rothaermel, F. T., & Deeds, D. L. (2006). Alliance type, alliance experience and
alliance management capability in high-technology ventures. Journal of
Business Venturing, 21, 429-460.

Rousse, M.J. & Daellenbach, U.S. (1999). Rethinking research methods for the
resource-based perspective: Isolating sources of sustainable competitive
advantage. Strategic Management Journal, 20, 487-494.

Simonin, B. L. (1999). Ambiguity and the process of knowledge transfer in
strategic alliances. Strategic Management Journal, 28, 595-623.

Sorensen, J. B., & Stuart, T. E. (2000). Aging, obsolescence, and organizational
innovation. Administrative Science Quarterly, 45(1), 81-112.

Spender, J. C. (1996). Making knowledge the basis of a dynamic theory of the firm.
Strategic Management Journal, 17, 45-62.

Stuart, T. E., Hoang, H., & Hybels, R. C. (1999). Interorganizational endorsements
and the performance of entrepreneurial ventures. Administrative Science
Quarterly, 44(2), 315-349.

Stuart, T. E., Ozdemir, S. Z., & Ding, W. W. (2007). Vertical alliance networks:
The case of university-biotechnology-pharmaceutical alliance chains.
Research Policy, 36, 477-498.

Survey: Climbing the helical staircase. (2003). The Economist, 366, 3.

Sydow, J., & Windeler, A. (1998). Organizing and evaluating interfirm networks:
A structurationist perspective on network processes and effectiveness.
Organization Science, 9, 265-284.

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

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

Trajtenberg, M. (2001). Innovation in israel 1968-1997: A comparative analysis
using patent data. Research Policy, 30, 363-389.

105

Tsoukas, H. (1996). The firm as a distributed knowledge system: A constructionist
approach. Strategic Management Journal, 17, 11-25.

Tully, S. (1994). Pill pushers get merger fever. Fortune, 129, 129.

Uzzi, B. (1996). The sources and consequences of embeddedness for the economic
performance of organizations: The network effect. American Sociological
Review, 61(4), 674-698.

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

Uzzi, B. (1999). Embeddedness in the making of financial capital: How social
relations and networks benefit firms seeking financing. American
Sociological Review, 64(4), 481-505.

Uzzi, B., & Gillespie, J. J. (2002). Knowledge spillover in corporate firms
financing networks: Embeddedness and the firm's debt performance.
Strategic Management Journal, 23, 595-618.

Uzzi, B., & Lancaster, R. (2003). Relational embeddedness and learning: The case
of bank loan managers and their clients. Management Science, 49(4), 383-
399.

Van De Ven, A. H. (1976). On the nature, formation and maintenance of relations
among organizations. Academy of Management Journal, 1, 24-36.

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

Wang, L., & Zajac, E. J. (2007). Alliance or acquisition? A dyadic perspective on
interfirm resource combinations. Strategic Management Journal, 28, 1291-
1317.

Weintraub, A. (2003). Five hurdles for biotech: The science is hot and the stocks
are up. Here's what's needed to deliver on the promise. In Business week
(Vol. 3835, pp. 60).

Wernerfelt, B. (1984). A resource-based view of the firm. Strategic Management
Journal, 5, 171-180.

Williamson, O. (1975). Markets and hierarchies. New York: Free Press.

106
Zaheer, A., McEvily, B., & Perrone, V. (1998). Does trust matter? Exploring
the effects of interorganizational and interpersonal trust on performance.,
Organization Science: A Journal of the Institute of Management Sciences,
9, 141-159.

Zaheer, A., & Venkatraman, N. (1995). Relational governance as an
interorganizational strategy: An empirical test of the role in economic
exchange. Strategic Management Journal, 16, 373-392.

Zajac, E. J., & Olsen, C. P. (1993). From transaction cost to transactional value
analysis: Implications for the study of interorganizational strategies.
Journal of Management Studies, 30, 131-145.

Zollo, M., Reuer, J., & Singh, H. (2002). Interorganizational routines and
performance in strategic alliances. Organization Science, 13, 701-713.

Zollo, M. & Winter, S. G. (2002). Deliberate learning and the evolution of
dynamic capabilities. Organization Science, 13, 339-351.

Zucker, L., & Darby, M. R. (1997). Present at the biotechnology revolution:
Transformation of technological identity for a large incumbent
pharmaceutical firm. Research Policy, 26, 429-446.

107
Appendix A
Appendix A: Sample Data from Medtrack

Data on Specific Deals
Partner(s) Date
Deal Size
(in $ MM) Deal Type(s)
Stage at
Signing
Current
Stage
Merck & Co Inc (US Public)
Ariad Pharmaceuticals Inc (US Public)
7/12/2007 0 Co-development,
Collaboration,
Commercialization
PhaseII PhaseII
GlaxoSmithKline plc (US Public)
Synta Pharmaceuticals Corp. (US
Public)
10/10/2007 0 Commercialization,
Development
PhaseIII PhaseIII

Indication
Area of
Interest
Milestone
(in $ mm)
Upfront (in $ mm)
Deal Title
Hematological Malignancies Cancer 652 75 Ariad Pharmaceuticals
and Merck Announce
Global Collaboration to
Jointly Develop and
Commercialize AP23573
for Cancer
Melanoma Cancer 835 80 GlaxoSmithKline and
Synta Pharmaceuticals
Announce Development
and Commercialisation
Collaboration for STA-
4783

108
Appendix A: Sample Data from Medtrack (Continued)
Summary
ARIAD Pharmaceuticals and Merck announced that they have entered into a global collaboration to jointly develop and
commercialize AP23573, ARIAD's novel mTOR inhibitor, for use in cancer. It is expected that AP23573 will enter into Phase III
clinical development for the treatment of metastatic sarcomas beginning this quarter. The companies anticipate conducting a
broad-based global development program in which clinical trials and biomarker studies will be conducted concurrently in
multiple cancer indications. Both companies will share overall responsibility for global commercialization and development of
AP23573. In the U.S., ARIAD will distribute and sell AP23573 for all cancer indications and book all sales, and ARIAD and
Merck will co-promote and will each receive 50 percent of the income from such sales. Outside the U.S., Merck will distribute,
sell and promote AP23573 and book all sales. On a global basis, ARIAD will be responsible for manufacturing the active
pharmaceutical ingredient used in the product, and Merck will be responsible for the formulation of the finished product
(tablets). In the U.S., ARIAD will have primary responsibility for development of AP23573 in the metastatic sarcoma indication.
Merck and ARIAD will have joint responsibility in the U.S. for development of all other cancer indications being pursued.
Outside the U.S., Merck will have primary responsibility for development in all cancer indications being pursued.
GlaxoSmithKline and Synta Pharmaceuticals announced the execution of a global collaboration agreement for the joint
development and commercialisation of STA-4783, a first-in-class, small-molecule, oxidative stress inducer that is entering
Phase III clinical development for the treatment of metastatic melanoma. Under the terms of the agreement, the companies
will share responsibility for development and commercialisation of STA-4783 in the US, and GSK will have exclusive
responsibility for development and commercialisation of STA-4783 outside the US.

109
Appendix A: Sample Data from Medtrack (Continued)
Clinical Trial Pipeline Data by Year

Array BioPharma Inc (US
Public)

Trial
Phase PC
Phase
I
Phase
II
Phase
III PA A M PM D F NA
2008 1 8 5 0 0 0 0 0 0 0 0 0
2007 2 7 5 0 0 0 0 0 0 0 0 0
2006 4 2 2 0 0 0 0 0 0 0 0 0
2005 5 1 0 0 0 0 0 0 0 1 0 0
2004 3 1 1 0 0 0 0 0 0 0 0 0
2003 1 1 0 0 0 0 0 0 0 0 1 1
2002 0 1 0 0 0 0 0 0 0 0 1 1
2001 0 1 0 0 0 0 0 0 1 0 0 0
2000 0 0 0 0 0 0 0 0 0 0 1 1

Product Pipeline by Indication Area

Array BioPharma Inc (US
Public)
Auto BLS Cancer Cardio CNS Dental Derm DS GDDS INF Kidney Met MSD
2008 3 0 10 0 0 0 0 0 0 1 0 0 0
2007 3 0 10 0 0 0 0 0 0 1 0 0 0
2006 2 0 5 0 0 0 0 0 0 1 0 0 0
2005 2 0 3 0 0 0 0 0 0 1 0 0 0
2004 0 0 3 0 0 0 0 0 0 1 0 0 0
2003 0 0 1 0 0 0 0 0 0 1 0 0 0
2002 0 0 0 0 0 0 0 0 0 1 0 0 0
2001 0 0 0 0 0 0 0 0 0 0 0 1 0
2000 0 0 0 0 0 0 0 0 0 0 0 1 0

110
Appendix A: Sample Data from Medtrack (Continued)

Product Pipeline by Indication Area (continued)
Misc Opth Resp SubAbs WH
2008 0 0 0 0 0
2007 0 0 0 0 0
2006 0 0 0 0 0
2005 0 0 1 0 0
2004 0 0 1 0 0
2003 0 0 1 0 0
2002 0 0 1 0 0
2001 0 0 1 0 0
2000 0 0 0 0 0

Abstract (if available)

Abstract Strategic alliances, or interorganizational relationships, are a prevalent strategy used to achieve organizational goals. Along with the increasing prevalence of alliances has come a heightened need to understand the dynamics of these relationships, as approximately half of the alliances entered fail (Kale, Dyer & Singh, 2002). This research, therefore, seeks to expand on our current understanding of alliance capability development by exploring the effects of specific types of alliance and technological experience as well as their interactions on new alliance formation and performance. Using the resource based view of the firm as a theoretical framework, this research finds that firms are significantly more likely to enter into alliance types, and with specific alliance partners, with which they have past experience. The RBV notion that firm resources will be utilized in strategic decisions is supported by this finding. Experience in a particular product area was also found to increase the likelihood of firms entering into non-R&D type relationships suggesting the use of alliance relationships to capitalize on or seek to exploit product specific resources in the alliance context. Alliance experience in general was found to positively influence the performance of alliance relationships, while specific types of alliance experience were only found to influence performance when entering into that type of relationship. In this way, firms seem to develop distinct capabilities in specific alliance types that are not necessarily easily transferred from one relationship type to another.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Empirical essays on relationships between alliance experience and firm capability development

PDF

Empirical essays on alliances and innovation in the biopharmaceutical industry

PDF

Understanding alliance building and communication across multiple technology platforms among California's immigrant-serving NGOs

PDF

Three essays on young entrepreneurial firms

PDF

Effects of economic incentives on creative project-based networks: communication, collaboration and change in the American film industry, 1998-2010

PDF

CEO selection performance: does board experience matter?

PDF

The evolution of knowledge creation online: what is driving the dynamics of Wikipedia networks

PDF

The power of visual imagery and its role in public relations

PDF

The influence of new media marketing public relations on the South Korean film industry -- in relation to the U.S. film industry

PDF

The evolution of multidimensional and multilevel networks in online crowdsourcing

PDF

The pop show: racial performance and transformation in global arts industries

PDF

Why we follow influencers: the role of Asian female fashionistas in Los Angeles

PDF

The rise and fall of Howard Dean: the media's role in the Iowa caucus "scream" debacle

PDF

Role enactment in interactive media: a role-play perspective

PDF

Bounded technological rationality: the intersection between artificial intelligence, cognition, and environment and its effects on decision-making

PDF

The Digital Revolution and its impact on the beauty industry

PDF

Introverts in public relations: how the industry can nurture them into the practice

PDF

The civic responsibility of professional sports teams: Exploring the role of a team in its community

PDF

Staying ahead of the digital tsunami: strategy, innovation and change in public media organizations

PDF

The elemental rhetoric of performance

Asset Metadata

Creator Stephens, Kimberlie J. (author)

Core Title The role of capabilities in new alliance creation and performance: a study of the biotechnology industry

School Annenberg School for Communication

Degree Doctor of Philosophy

Degree Program Communication

Publication Date 08/04/2009

Defense Date 04/21/2009

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag OAI-PMH Harvest,resource based view,strategic alliances

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Fulk, Janet (committee chair), Mayer, Kyle (committee member), Monge, Peter (committee member)

Creator Email kimberliestephens@yahoo.com,kimpothoven@yahoo.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m2464

Unique identifier UC1492021

Identifier etd-Stephens-3095 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-173205 (legacy record id),usctheses-m2464 (legacy record id)

Legacy Identifier etd-Stephens-3095.pdf

Dmrecord 173205

Document Type Dissertation

Rights Stephens, Kimberlie J.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu