Content uploaded by Kulwant Singh
Author content
All content in this area was uploaded by Kulwant Singh on Feb 25, 2019
Content may be subject to copyright.
SPECIAL ISSUE ARTICLE
Creating and taming discord: How firms manage
embedded competition in alliance portfolios to limit
alliance termination
Navid Asgari
1
| Vivek Tandon
2
| Kulwant Singh
3
| Will Mitchell
4
1
Strategy and Statistics Area, Fordham University,
Gabelli School of Business, New York, New York
2
Department of Strategic Management, Temple
University, Fox School of Business, Philadelphia,
Pennsylvania
3
Department of Strategy and Policy, National
University of Singapore Business School,
Singapore
4
Strategic Management Area, University of
Toronto, Rotman School of Management,
Toronto, Ontario, Canada
Correspondence
Navid Asgari, Strategy and Statistics Area,
Fordham University, Gabelli School of Business,
45 Columbus Avenue, New York, NY 10023.
Email: nasgari@fordham.edu
Funding information
NUS Research Grant, Grant/Award Number: R-
313-000-113-112; Social Sciences and Humanities
Research Council of Canada
Research Summary:Firms with resources that make
them attractive allies are also desirable partners for com-
petitors so that competition among partners is embedded
in alliance portfolios. We develop a framework in which
competition within a portfolio creates benefits for a focal
firm but threatens partners, increasing the hazard of alli-
ance termination. We then propose four mechanisms for
managing the threat of competition to partners reflecting
aspects of portfolio configuration: alliance governance,
social cohesion, social structure of competition, and part-
ner similarity. We test our framework using a sample of
204 biopharmaceutical firms with alliance portfolios com-
prising 1,621 alliances between 1990 and 2000. The
study addresses the interplay of competition and coopera-
tion in alliance portfolios, and more generally, key
aspects of value chain integration strategy.
Managerial Summary:Alliance portfolios comprise a
focal firm’s set of direct partners, some of which compete
with each other because of overlapping resources, capa-
bilities, and strategies. The threat of actual or perceived
competition from other partners may cause some firms to
terminate their alliance with the focal firm. We develop a
framework comprising four mechanisms related to alli-
ance portfolios—alliance governance, social cohesion,
social structure of competition, and partner similarity—
that allows focal firms to attenuate the hazard of termina-
tion of their alliances. We find support for our framework
in a study of 204 biopharmaceutical firms with alliance
portfolios comprising 1,621 alliances between 1990 and
2000. We improve understanding of how firms can man-
age competition and cooperation within their alliance
portfolios.
Received: 2 November 2015 Revised: 21 August 2017 Accepted: 26 August 2017
DOI: 10.1002/smj.2784
Strat Mgmt J. 2018;1–27. wileyonlinelibrary.com/journal/smj Copyright © 2018 John Wiley & Sons, Ltd. 1
KEYWORDS
alliance portfolios, biopharmaceutical industry,
competition and cooperation, termination, value chain
integration
1|INTRODUCTION
Creating and managing alliance portfolios (Hoffmann, 2007; Lavie, 2007; Wassmer, 2010) is a cen-
tral part of a value chain integration strategy (Kapoor & Adner, 2012; Mitchell, 2014). Yet, the pro-
cess of creating alliance portfolios commonly involves competition among focal firms’allies
(Gomes-Casseres, 1996; Vassolo, Anand, & Folta, 2004) because desirable partners often possess
relevant resources (Dussauge, Garrette, & Mitchell, 2000; Garrette, Castañer, & Dussauge, 2009;
Hennart, 1988; Stuart, 1998). Research is only beginning to explore the tensions that may arise if
firms in a portfolio fear that competing partners may access resources they do not want to disclose.
We argue that competition among a focal firm’s portfolio partners increases the hazard of termina-
tion of the focal firm’s alliance with a partner facing such competition, but that firms may attenuate
these risks via aspects of portfolio configuration involving alliance governance, social cohesion,
social structure of competition, and partner similarity.
We build arguments using the literature on alliance portfolios (Lavie, 2007), value chain integra-
tion (Mitchell, 2014), and studies of competition among allies (Bresser, 1988; Cui, Calantone, &
Griffith, 2011; Gimeno, 2004; Gomes-Casseres, 1996; Silverman & Baum, 2002). Several studies
examine the effects of dyadic competition on alliances (e.g., Blodgett, 1992; Singh & Mitchell,
1996), but provide little understanding of how competition among a focal firm’s partners influences
the stability of alliance portfolios (Greve, Baum, Mitsuhashi, & Rowley, 2010; Greve, Mitsuhashi, &
Baum, 2013), and particularly, of what factors moderate the challenges posed by competition
between partners. Network studies, commonly based on board interlocks, examine how competition
between a firm and its indirect partner impacts termination of the relationship with its direct partner
(e.g., Hernandez, Sanders, & Tuschke, 2015), but do not examine how this risk is affected by char-
acteristics of the relationship between the firm and its partners. Studies of alliance formation, mean-
while, examine how relationship-specific characteristics influence alliance formation (e.g., Gimeno,
2004), but do not assess alliance termination; the impact on formation and termination is likely to
differ because ex-ante arm’s-length negotiations at formation qualitatively differ from ex-post deeper
ties at termination. The underlying issue we address is how a focal firm can manage discord and
portfolio instability arising from competition between its partners.
We differentiate between alliance portfolios and dyads. The fundamental distinction is that the
focal firm in a dyad has only one partner, while a portfolio involves multiple partners (Hoffmann,
2007; Lavie, 2007; Shipilov & Li, 2011; Shipilov, Li, & Greve, 2011; Wassmer, 2010; Wassmer,
Li, & Madhok, 2017). With the presence of multiple allies, a portfolio creates the potential for dis-
cord among partners (Simmel, 1950) so that the termination of a given alliance in a portfolio may
be contingent on the nature of the focal firm’s other ties within the portfolio. The potential for dis-
cord and termination requires a focal firm to manage its portfolio as a social group so that it can
maintain its collective properties and usefulness. In the absence of common membership in a focal
2ASGARI ET AL.
firm’s portfolio, competition between two firms in dyadic alliances has no bearing on the termina-
tion of a focal firm’s alliances with either of them.
We propose a two-part framework that first details the focal firm’s benefits and the threat to a
partner from competition within the portfolio, and then identifies approaches that can moderate the
impact of this competition, along with the mechanisms for implementing these approaches. The
argument begins with the idea that competition among partners can create benefits for a focal firm
arising from opportunities to combine resources (Vasudeva & Anand, 2011), bargaining power
(Lavie, 2007), and hedging against risk and dependence (Singh & Mitchell, 1996). These benefits
come at a cost, however, raising concerns among partners of resource leakage to competitors,
reduced rents, and the loss of the focal firm’s attention to its alliances (Hernandez et al., 2015;
Kogut, 1989; Singh & Mitchell, 1996). We argue that the benefits to focal firms from greater com-
petition are capped by growth in the size and complexity of a portfolio, but that the threat to partners
increases with competition in the portfolio. Therefore, we predict that competition among partners
increases the hazard of termination of a focal firm’s alliance.
We then evaluate how termination hazards can be contained. We identify four mechanisms that
may help firms allay partners’concerns about competition within the portfolio: investing in equity,
increasing social cohesion of the portfolio, increasing the number of direct paths between partners,
and seeking strategically similar partners. These mechanisms potentially reduce the likelihood of
concerns about competition, limit the salience of the focal firm as a source of concern about the loss
of resources to competitors, and help partners recognize and manage the negative effects of competi-
tion they face. Hence, we propose that focal firms can moderate the negative effects of competition
on termination. We test the hypotheses with a longitudinal sample of 204 firms with portfolios com-
prising 1,621 alliances in the biopharmaceutical industry between 1990 and 2000; this sector is
marked by widespread use of alliance portfolios, often including competing partners.
We contribute to the literature on alliance portfolios, and more generally, value chain integration.
We extend studies of competitive influences beyond dyadic alliances to alliance portfolios, and
show that the value of alliance portfolios is both enhanced and undermined by incorporating com-
peting partners. We identify competition between partners as a significant potential influence on the
termination of the focal firm’s alliances and highlight mechanisms that moderate the impact of this
competition, thereby offering insights on the interplay of competition and cooperation within alli-
ance portfolios. This approach expands explanations of alliance termination, which currently focus
on failed cooperation, successful completion, exogenous change, or the emergence of alternate
opportunities (Greve et al., 2013; Singh & Mitchell, 1996). In doing so, we respond to calls from
alliance researchers to study “how and why firms change the configuration of their alliance portfo-
lios over time”(Wassmer, 2010, p. 162) using “multi-firm panel data on the evolution of alliance
portfolios in particular industries”(Lavie & Singh, 2012, p. 803). The focus on how the tension
between cooperation and competition affects alliance termination provides a complementary per-
spective to the role of alliance formation in the evolution of alliance portfolios.
2|COMPETITION AND ALLIANCE TERMINATION
An alliance portfolio is a set of direct partnerships by which firms obtain resources that meet exist-
ing and future needs at lower cost than through market exchange or internal development
(Hoffmann, 2007; Lavie, 2006). The central notion of alliance portfolios is that the focal firm man-
ages the collective properties of its portfolio by balancing the benefits of resource exchange against
the costs of maintaining the portfolio and of sharing resources with partners that may be current or
ASGARI ET AL.3
potential competitors (Lavie & Singh, 2012). Through their collective properties, alliance portfolios
differ from networks or collections of independent ties; activities and outcomes of individual alli-
ances in a portfolio are not independent of each other (Vassolo et al., 2004; Wassmer, 2010).
A key means to create value through alliance portfolios is to orchestrate the resources within
portfolio partners to identify and capitalize on resource combinations (Parkhe, 1993). Firms with
similar resources are likely to be actual or potential competitors, particularly if they operate in the
same industry (Wernerfelt, 1984). Therefore, competition among partners is an unavoidable and
embedded aspect of alliance portfolios (Bresser, 1988; Gomes-Casseres, 1996). Theory offers inde-
terminate predictions of the impact of such competition on alliance termination (Burt, 1992; Gnya-
wali & Park, 2011; Kogut, 1989; Shipilov, 2009). With limited exceptions (e.g., Lavie, 2007), few
studies have examined the interplay of competition and cooperation within alliance portfolios. Yet,
this interplay influences the benefits that partners separately and collectively gain from their alliance
(Khanna, Gulati, & Nohria, 1998), and how focal firms form, operate, and terminate alliances within
their portfolios.
Figure 1 summarizes the key actors and relationships within an alliance portfolio. Fis the focal
firm that establishes the alliance portfolio, comprising at least two partners. We evaluate the impact
of competition among F’s partners on the termination of its alliance with partner P. We distinguish
between F’s partners that compete with P, which we label PCs, and those that do not compete with
P, which we label PNCs. Our baseline hypothesis evaluates the hazard of termination of the alliance
between Fand P, owing to P’s concerns about F’s alliances with PCs. In subsequent hypotheses,
we evaluate moderators of this hazard.
1
We focus on how competition between partners Pand PCs in a portfolio influences termination
of F’s alliances with P, departing from studies that evaluate alliance termination between direct
competitors (Blodgett, 1992; Singh & Mitchell, 1996) or termination of board memberships due to
competition from indirect partners in board-interlock networks (Hernandez et al., 2015). Competi-
tion between Pand PCs is likely to increase the risk that the F–Palliance terminates. This focus
aligns with the alliance portfolio level of analysis, while offering insights into the complexity of
managing portfolios containing cooperative and competitive relationships. Despite investigation of
the causes of alliance instability (e.g., Borys & Jemison, 1989), explanations for termination of
inter-firm cooperation are incomplete (Greve et al., 2010, 2013).
FIGURE 1 Focal firms and partners in an alliance
portfolio
F: Focal firm; P: Partner firm; PC: Other firms in F’s
portfolio that compete with Pin the market for technology;
PNC: Other firms in F’s portfolio that do not compete with
P; Solid lines: Cooperative relationship between F,PCs,
and PNCs.
Example portfolio statistics
Size of F’s portfolio: Six alliances.
Competition for F–Palliance: Pcompetes with three of F’s
other partners
1
Alliances between Fand PNC may indirectly affect the hazard of termination of F’s alliance with Pif the time spent with PNC limits
F’s capacity to maintain its alliance portfolio (Singh & Mitchell, 1996); we address this issue empirically by considering portfolio
characteristics.
4ASGARI ET AL.
In the study that most closely matches ours, Hernandez et al. (2015) evaluated how Pmay termi-
nate its board-interlock tie with Fto prevent the loss of strategic knowledge to competitors through
common board memberships. The study assessed P’s characteristics, network ties formed by board
overlaps, and resolution of the knowledge loss problem by closing a network and pruning partners
connected to competitors. In contrast, we examine formal alliances within portfolios and study how
characteristics of F,P, and the F–Prelationship, and indirectly, of the P–PCs relationships, affect
termination of the F–Palliance. Because the basic threat that PCs pose for Pis the loss of resources
through F’s opportunistic behavior or its failure to protect P’s property rights, we examine mecha-
nisms that can moderate the effects of P–PCs competition on the termination of F–Palliances. The
nature of board-interlock networks limits examination of such alliance-specific characteristics. Her-
nandez et al. (2015, p. 1254) observed that board interlocks and alliances differ significantly, and
call for studies to “assess the different risks of direct and indirect links to rivals as well as the
defenses that are effective for each kind of risk.”
2.1 |Hypothesis 1: Competitors and termination of alliances
Competition between Pand F’s portfolio partners, PCs, is likely to affect the hazard of termination
of the F–Palliance, but will have varying impact on Fand on P. In Panel A of Table 1, we summa-
rize the benefits to Fof competition between Pand PCs, and how each benefit is mirrored in con-
cerns for P. Although the benefits and concerns are most significant when Pand PCs are close
competitors, our arguments do not require actual competition, but only that Pperceives PCstobe
competitors. Our subsequent references to competition between Pand PCs include both actual and
perceived competition.
Competition among partners may be beneficial to Fin three ways. First, competition among
partners provides Fwith combinatorial opportunities through potential access to a more diverse, yet
overlapping, range of resources (Gnyawali & Park, 2011; Lavie, 2007; Mahmood et al., 2013;
Rothaermel & Boeker, 2008; Vasudeva & Anand, 2011). Second, competition among partners will
increase F’s bargaining power and its ability to extract better terms from these partners (Lavie,
2007; Shipilov, 2009). Third, competition in the portfolio allows Fto hedge risks and reduce its
TABLE 1 Impact of competition among partners in alliance portfolios
Panel A. Competition between partners: Benefits to focal firms and partners’concerns
Benefits to focal firm (F) from competition between partners
(Pand PCs)
Partner’s(P) concerns arising from competitors (PCs)
in portfolio
•Combinatorial opportunities from sharing among partners with
overlapping resources (Gnyawali & Park, 2011; Mahmood,
Chung, & Mitchell, 2013; Vasudeva & Anand, 2011)
1. Leakage of resources to other partners (Katila,
Rosenberger, & Eisenhardt, 2008; Oxley &
Wada, 2009)
•Bargaining opportunities (Lavie, 2007; Shipilov, 2009) Loss of rents (Lavie, 2006)
•Hedging of risks and dependence, when partners have overlapping
resources (Jiang, Tao, & Santoro, 2010; Singh & Mitchell, 1996)
Other alliances distracting F from the F—P alliance
(Singh & Mitchell, 1996)
Panel B. Potential mechanisms to reduce effects of competition between partners.
How to address partner’s(P) concerns Mechanisms
1. Reduce the likelihood of concerns about competition 1a. Governance arrangements of F–Palliance: F’s equity
investment in F–Palliance
1b. Portfolio social cohesion: Density of F’s portfolio
2. Reduce the salience of the focal firm as a conduit for
resource transfers
2. Social structure of competition of P’s alliances: Direct paths
between Pand PCs
3. Help the partner recognize and address challenges 3. Similarity between Fand P: Strategic similarity between Fand P
ASGARI ET AL.5
dependence on Pby accessing substitute resources from alliances with PCs (Jiang et al., 2010;
Singh & Mitchell, 1996). These arguments suggest that Fwill gain from and seek to retain alliances
with partners that are competitors. However, the benefits of competition within the portfolio will be
capped as the number of competitors increases, owing to the costs and complexity of managing a
larger and more complex portfolio as well as discord within the portfolio.
The benefits of competition for Fraise three related concerns for P.
2
First, F’s gains from acces-
sing resources from competing partners will raise P’s concerns about the possible loss of its
resources to competitors (Borys & Jemison, 1989; Bresser, 1988; Hernandez et al., 2015; Katila
et al., 2008; Khanna et al., 1998; Mansfield, 1985; Oxley & Wada, 2009). The core benefit of the
portfolio for F—the relatively efficient combination and integration of resources from multiple
partners—may allow PCs to access resources that Pintended to share only with F. PCs are likely to
be able to absorb resources from Pthrough F; being competitors with Pwill provide PCs with the
capacity to understand and absorb transferred resources (Cui et al., 2011; Dussauge et al., 2000;
Hennart, 1988; Lane & Lubatkin, 1998). Therefore, Fmay serve as a conduit for the transfer of
resources among competitors in its portfolio. Fneed not actively transfer resources to PCs to raise
the hazard of termination of the F–Palliance. Instead, it is only necessary for Pto be concerned
about this possibility. Similarly, Pdoes not have to observe F’s active transfer of its resources to
monitor PC. Instead, the fact that PC is a competitor will increase the likelihood that Pwill monitor
PC to detect such transfers.
Second, F’s bargaining power arising from its alliances with PCs will increase the threat to P’s
performance (Lavie, 2007). Greater competition for Pin F’s portfolio lowers F’s dependence on
Pand increases P’s concerns about its relational rents being bargained away. Therefore, Pwill be
concerned about cooperation between Fand PCs.
Third, Pwill be concerned that alliances with PCs will reduce F’s dependence on and attention
to the F–Palliance. Firms have limited alliance carrying capacities, so that attention to one alliance
may come at the cost of reduced focus on another alliance (Singh & Mitchell, 1996). Concern about
the diversion of F’s attention (Ocasio, 1997) to other alliances will grow with competition between
Pand PCs.
These arguments suggest that competition from PCs will threaten Pand increase the likelihood
that Pwill terminate its alliance with F, to safeguard its resources, to signal its dissatisfaction with
F, or in the event of actual resource transfers, to commence the process of obtaining remedies from
For PCs. All three of P’s concerns increase with the competition it faces in the portfolio.
In sum, Fbenefits from competition between Pand PCs, increasing the likelihood that Fwill
seek to maintain its alliance with P. In contrast, competition between Pand PCs threatens P,
increasing the likelihood that Pwill respond (Katila et al., 2008), including by terminating its alli-
ance with F.
We expect P’s propensity to terminate the F–Palliance to dominate F’s propensity to retain the
alliance, for two reasons. First, the continuation of an alliance requires acquiescence of both partners
so that no one partner is able to enforce continued cooperation (Ahuja, 2000). Second, the benefits
of increased competition for Fwill be offset by the associated costs and complexity of managing
2
Several explanations arise for why Pwould ally with Fdespite the risk of loss of resources to PCs: (a) Pmight initially be unaware
of F’s alliances with PCs; (b) PCs might have emerged as competitors after Pestablished its alliance with F; (c) Pmay initially have
mistakenly viewed PCs as lacking the capacity to absorb its resources; (d) assurances from Fmay have eased P’s concerns about mis-
appropriation and unintended resource transfers; (e) Pmay have formed the alliance as part of a strategy to match PCs’access to For
to monitor PCs (Cui et al., 2011; Reuer & Ragozzino, 2006; Singh & Mitchell, 1996; Wassmer, 2010). To the extent that P accounts
for competitive concerns in its decision to ally with F, the impact of competition on the termination of alliances is likely to be attenu-
ated; this makes our tests conservative.
6ASGARI ET AL.
greater competition among partners within its portfolio; in contrast, P’s concerns will continue to
increase with competition within the portfolio. Hence, we propose that greater competition for Pin
the portfolio increases the hazard of termination of its alliance with F.
Hypothesis 1 (H1) The greater the competition between a partner and other firms in
the focal firm’s alliance portfolio, the greater the hazard of termination of the alliance
between the focal firm and the partner.
Two issues arise. First, it is conceivable that the threats to Pwill be limited at low levels of com-
petition from PCs, and will only affect the hazard of alliance termination when competition
increases beyond a threshold level. This suggests a potential nonlinear relationship between compe-
tition and alliance termination. We expect this possibility to depend on the context-specific nature of
benefits, threats, and competition; consequently, we do not hypothesize this relationship, but investi-
gate it empirically.
Second, we primarily evaluate alliance termination from P’s perspective because it faces the
negative consequences of competition. Factors that allay P’s concerns are most relevant in miti-
gating the challenges posed to alliance stability by competition; identifying and examining such
factors is the primary focus of our study. From F’s perspective, competition between Pand PCs
is beneficial, making it unlikely that it will terminate the F–Palliance due to competition. How-
ever, F’s benefits from and its dependence on its alliance with P, and the availability of substitute
alliances (Greve et al., 2013; Singh & Mitchell, 1996), will influence the extent to which Fwill
risk unsettling Pthrough its alliances with PCs and will influence how it resists P’s efforts to ter-
minate the F–Palliance. We deal with this issue empirically by including controls for F’sdepen-
dence on the F–Palliance, the nature of F’s relationships with P,andF’sandP’s alliances and
characteristics.
2.2 |Moderating the impact of competition on alliance termination
Focal firms will seek to limit the hazards of alliance termination arising from competition among
partners to benefit from such competition while limiting portfolio instability. Termination is costly
because of the difficulty of unravelling relationships, dismantling governance and coordinating
structures, dealing with disruption, establishing replacement alliances, and mitigating negative repu-
tation effects (Gulati & Singh, 1998; Singh & Mitchell, 1996). Theoretically, the hazard of termina-
tion due to competition may be mitigated by reducing the incidence of the problem caused by
competition, whether by making the source of the problem less salient, or by helping the partner rec-
ognize and address the problem.
We identify four mechanisms that moderate the effects of competition on the hazard of termina-
tion of the F–Palliance. These mechanisms reduce the likelihood that P’s concerns about competi-
tion will arise, reduce the salience of Fas a conduit for the competitive threat when it occurs, and
help Precognize and address challenges. Panel B of Table 1 summarizes the mechanisms. As we
elaborate in the following sections, governance arrangements for the F–Palliance and portfolio
social cohesion can reduce the incidence of the problems caused by competition between Pand
PCs; the social structure of competition will reduce the salience of Fin transferring resources to
PCs; and similarity between Pand Fwill help Precognize and address the challenges. We focus on
these moderating mechanisms as options for F,P, or the combination of Fand Pto reduce the nega-
tive effects of competition on P(Panel A, Table 1). The four mechanisms may not align
ASGARI ET AL.7
diametrically with P’s concerns, but can allay these concerns, reducing the effect of competition
from PCs on the hazard of termination of the F–Palliance.
The mechanisms that moderate alliance termination may overlap with those that influence alli-
ance formation (Gimeno, 2004). However, we expect these influences to have different impact
because the ex-ante arm’s-length negotiations between potential allies at the time of formation trans-
form at termination into ex-post interdependent relationships based on alliance-specific investments
and resource exchange (Greve et al., 2010; Gulati & Singh, 1998; Williamson, 1991). For example,
though appropriation concerns may influence governance arrangements for potential alliances, after
the alliances are formed, the governance arrangements will have substantially different impact on
preventing or failing to prevent appropriation (Gulati & Singh, 1998) and on facilitating or hindering
the termination of formal alliances. Greve et al. (2010) concluded that explanations for alliance for-
mation and termination are distinct. Evaluating influences on alliance termination therefore provides
a complementary perspective that adds to the holistic understanding of alliance formation and termi-
nation, and of portfolio evolution.
2.3 |Hypothesis 2: Alliance governance
Governance arrangements specify mechanisms for managing the alliance and the relationship
between partners. These arrangements facilitate the establishment and operation of alliances by spec-
ifying the means, focus, and incentives for cooperation and exchange (Sampson, 2007; Williamson,
1991). The governance agreement typically establishes procedures for information exchange, part-
ners’monitoring, and penalties for breaches, serving as a hostage-taking device and reducing part-
ners’opportunism (Williamson, 1983, 1991).
We consider governance in terms of equity investments in each alliance, and in the next hypoth-
esis, social cohesion in the portfolio. Stronger governance can be achieved through the exchange or
purchase of equity. Relative to less hierarchical governance modes, equity investment is associated
with greater formalization through contractual agreement as well as the establishment of charters,
routines, and mechanisms for coordinating exchange between partners (Gulati & Singh, 1998). In
turn, this formalization increases the effectiveness of information reporting and monitoring
(Sampson, 2007), mutual hostage-taking (Williamson, 1983, 1991), and alignment of interests
(Oxley, 1997; Pisano, 1989).
Greater equity exchange in an alliance will reduce P’s concerns about competition from PCsin
three ways. First, an equity-based governance arrangement between Fand Pwill facilitate social
and structural interaction, providing Pwith more information to monitor and control the resources it
shares with F(Gulati, 1995; Oxley & Wada, 2009; Sampson, 2007), potentially limiting resource
transfers through Fto PCs. Second, greater equity investment works as a mutual hostage-taking
device (Williamson, 1983), which may reduce F’s ability to bargain away P’s rent. Third, greater
equity investment is more likely to align F’s and P’s interests, ensuring that Fpays sufficient atten-
tion to its alliance with Pand to protecting P’s interests, even in the presence of its ties with PCs
and PNCs. Therefore, greater equity investments can reduce P’s concerns about competition from
PCs, moderating the impact of such competition on the hazard of termination of the F–Palliance.
Hypothesis 2 (H2) The greater the equity investment in a focal firm-partner alliance,
the weaker the impact of competition between that partner and other partners in the
alliance portfolio on the hazard of termination of the focal firm-partner alliance.
8ASGARI ET AL.
2.4 |Hypothesis 3: Portfolio social cohesion
The broader structure of relationships between the partners in a focal firm’s portfolio, particularly
the cohesiveness of the portfolio’s social structure, can mitigate P’s concerns of competition in the
portfolio. Social cohesion in a focal firm’s portfolio can serve as a governance mechanism by creat-
ing and enforcing a set of norms that encourage cooperation and discourage opportunism between
partners. We focus on portfolio density as a social structural mechanism to allay P’s concerns. Port-
folio density measures the intensity of ties between partners (Burt, 1992; Haunschild & Beckman,
1998; Krackhardt, 1999), and therefore, indicates social cohesion in the portfolio. A high-density
portfolio has more ties among firms than a low-density portfolio.
The larger number of relationships between partners in a dense portfolio increases partners’
embeddedness and social cohesiveness, facilitating the emergence and enforcement of common
norms supporting cooperation (Coleman, 1988; Krackhardt, 1999) and generalized trust between
partners in the portfolio (Gomes-Casseres, 1996; Krackhardt, 1999). These norms of cooperation
align interests and reduce concerns about opportunistic behavior by partners in the portfolio. Higher
portfolio density also provides Pwith more paths and information sources for monitoring
PCs. Thus, a dense portfolio of partners can allay P’s concerns about inappropriate transfers
between partners. Collectively, these arguments support the view that greater portfolio density
increases social cohesion, reduces P’s concerns about PCs, and moderates the effects of competition
on the hazard of termination of the F–Palliance.
Hypothesis 3 (H3) The greater the portfolio density, the weaker the impact of compe-
tition between partners on the hazard of termination of the focal firm-partner alliance.
2.5 |Hypothesis 4: Social structure of competition
The third mechanism moderating the impact of competition on the termination of the F–Palliance
is the social structure of competition that Pfaces, which can reduce the salience of F in facilitating
resource transfers between Pand PCs. In Hypothesis 1, we argued that Pis more likely to terminate
its alliance with Fwhen Fmight be a conduit through which its resources are lost to
PCs. Conversely, the existence of alternate conduits or pipes (Podolny, 2001) through which P’s
resources may be transferred to PCs will reduce the salience of Fas the conduit through which
resources may be lost. Paradoxically, concerns about any particular alliance or partner facilitating
the loss of resources will be limited by the existence of other paths through which resources may be
transferred.
We focus on how direct paths linking Pand PCs reduce the salience of Fas a conduit between
Pand PCs. Figure 1 illustrates how direct ties allow P’s resources to be transferred to PCs without
Fserving as a conduit. To preview our empirical analysis, we adopt a measure of direct paths based
on direct ties between Pand PC, while controlling for the overall structure of the links between
these firms.
Direct paths linking Pand PCs reduce the salience of Fas a conduit through which P’s
resources may be lost to PCs in two ways. First, direct paths offer PCs a direct route to P’s
resources, reducing the threat of Fas a conduit, and P’s concerns that Fis the primary source of
resource leakage to PCs (Haunschild & Beckman, 1998). Second, direct paths will close structural
holes between Pand PCs (Verspagen & Duysters, 2004), reducing F’s bargaining power over
Parising from its structural position (Burt, 1992; Coleman, 1988). Additionally, direct paths reflect
P’s and PCs’embeddedness within F’s portfolio, which may discourage PCs from seeking to access
ASGARI ET AL.9
P’s resources through F; the closer social structure may punish PCs if they access such resources
(Coleman, 1988; Krackhardt, 1999).
These arguments suggest that direct paths between Pand PCs reduce the salience of Fas a
source of concern to Pover the transfer of resources to PCs, moderating the effects of competition
on the hazard of termination of the F–Palliance.
Hypothesis 4 (H4) The greater the number of direct paths between a focal firm’s part-
ner and the partner’s competitors in the portfolio, the weaker the impact of competi-
tion between these partners on the hazard of termination of the focal firm-partner
alliance.
2.6 |Hypothesis 5: Similarity between Fand P
The fourth mechanism is strategic similarity between Fand P, which can reduce the negative impact
of competition between Pand PCs. Greater strategic similarity between these firms will help
Punderstand F’s intent and actions, and allow Pto detect whether its resources are transferred to
PCs. This will also allow Pto deal with Fmore effectively, reducing the hazard of termination of
the F–Palliance.
Similarity between firms can be judged on multiple dimensions such as culture, size, age, com-
petencies, product-market activities, and marketing, technology, or cost orientation. We focus on the
notion of strategic similarity (Ramaswamy, 1997; Spender, 1989), which refers to commonality in
strategy, operational focus, technology, and challenges. Strategic similarity usually results from
firms operating in the same industry or market segment, which creates overlapping leadership expe-
riences, strategies, and resources (Porac & Thomas, 1990). Firms in the same industry or sector are
more likely to agree and comply with norms and practices; this will facilitate understanding, interac-
tion, and the development of trust as well as the formation and stability of alliances (Borys & Jemi-
son, 1989; Coleman, 1988).
We propose that strategic similarity between Pand Fwill moderate the effects of competition
between Pand PCs on the hazard of termination of the F–Palliance in three ways, all of which arise
from greater shared understanding between Fand P. First, greater strategic similarity will facilitate
P’s understanding of F’s incentives and strategy, helping Panticipate F’s ability to transfer P’s
resources to PCs (Cohen & Levinthal, 1990; Katila et al., 2008; Lane & Lubatkin, 1998). This under-
standing and anticipation will help Pcraft more complete contracts ex ante and devise noncontractual
mechanisms (Reuer & Ariño, 2007; Reuer & Devarakonda, 2016) to limit resource leakage to PCs
through F. Second, strategic similarity will help Pto anticipate F’s bargaining strategies and how
Fmay opportunistically encourage competition between Pand PCs (Rowley, Greve, Rao, Baum, &
Shipilov, 2005); Pmay be more effective at resisting and countering F’s bargaining because greater
strategic similarity will provide Pwith greater understanding of F’s pay-off structure.
Third, strategic similarity will make it easier for Fto pay sufficient attention to its alliance with
Peven when Fis involved in multiple alliances with PCs because greater mutual understanding
eases communication and fosters common understanding, making it less costly for Fto interact with
P(Lane & Lubatkin, 1998; Ocasio, 1997; Singh & Mitchell, 1996). As a result, Pwill be less con-
cerned about how competition will affect F’s capacity to work with multiple partners. In turn, strate-
gic similarity will allow Pto monitor Fmore effectively, to prevent or detect resource transfers, or
in the event that its resources are transferred to PCs, to retaliate or seek compensation (Katila
et al., 2008).
10 ASGARI ET AL.
Therefore, greater strategic similarity between Fand Pwill help allay P’s concerns and moder-
ate the effect of competition between Pand PCs on termination of the F–Palliance:
Hypothesis 5 (H5) The greater the strategic similarity between a focal firm and its
partner, the weaker the impact of competition between that partner and the focal
firm’s other partners on the hazard of termination of the focal firm-partner alliance.
In sum, we predict that competition between a partner Pand its competitors PCs within the focal
firm’s portfolio increases the hazard of termination of the F–Palliance (H1). We expect the hazard
of termination of the F–Palliance to be moderated by greater equity investment in the alliance (H2),
greater portfolio density (H3), more direct paths between Pand PCs (H4), and strategic similarity
between Fand P(H5).
3|DATA AND METHODS
We locate our study in the global biopharmaceutical industry between 1990 and 2000. The industry
experiences high rates of change, uncertainty, and technological challenges, which encourage exten-
sive collaboration and formation of alliance portfolios (Rothaermel & Boeker, 2008; Vassolo et al.,
2004). Pharmaceutical firms compete widely across product markets and their related technological
areas, providing variation to study competition and alliances. Extensive reporting on the industry
ensures the availability of detailed information on firms and their alliances (Lerner, Shane, & Tsai,
2003). Our focus on a single industry controls for industry-level heterogeneity in technological
change or product market influences and opportunities, which can confound alliance termination.
We started our data collection with 519 firms listed in the Osiris database as formed before 2000
with the primary North American Industrial Classification System codes for biopharmaceuticals
firms. We excluded firms without financial or other data in Recap (www.recap.com), a reliable
source of information on biopharma alliances (Lerner et al., 2003; Schilling, 2009). We used Recap
to identify all alliances formed between 1990 and 1999, which we tracked until 2000 to identify ter-
minations. For alliances that existed in 1990, we collected data back to 1986 or to their year of for-
mation. We excluded ties that were exchanges of letters of intent, warrants, or other forms of
cooperation that did not amount to alliances; that had no data in Recap; or for which there was no
supporting information in Factiva (www.dowjones.com/factiva). This resulted in our initial data set
of 296 focal firms with 3,432 alliances. We then excluded alliances formed with nonprofit organiza-
tions (such as universities and medical centers), nonlisted firms, and those with missing data. We
excluded multiparty alliances, which differ from alliance portfolios (there were 106 such alliances;
we expect competition among partners in multiparty alliances and among partners in alliance portfo-
lios to have broadly similar effects, with the focal firm seeking to manage the hazard of alliance ter-
mination). We also dropped alliances where the reported year of formation was the same as the
reported termination year; these alliances were likely not formally established.
It was important to remove terminations arising from successful completion (Greve et al., 2013).
We excluded 18 alliances that were terminated because they achieved their objectives. We excluded
29 alliances that ended as their original contracts scheduled, which we interpreted as unrelated to
competition. Finally, we estimated models only for firms with alliance portfolios (at least two alli-
ances); we evaluated potential bias of this specification in our empirical analysis. Our final sample
comprised 204 firms with 1,621 alliances.
ASGARI ET AL.11
3.1 |Dependent variable
The dependent variable for all hypotheses is the hazard of Termination of an alliance. A dummy
variable (0–1) noted if an alliance was terminated during the study period. Though Recap reports
the formation of most alliances, it reported termination for only 10.5% of the alliances during our
study period. This low proportion reflects, in part, the reluctance of firms to publicize the termina-
tion of alliances (Schilling, 2009).
To address limited reporting of termination, we followed Asgari, Singh, and Mitchell (2017) in
using an intensive search process to establish whether alliances were terminated by year 2000. We
searched Factiva using partners’names and alliance characteristics to locate press releases relating
to each alliance. These press releases identified termination for an additional 12.5% of alliances. For
alliances without termination dates, we followed the stream of press releases referring to each alli-
ance until the final press release. We adopted the year of the final press release as its termination
year. The process estimated termination dates for another 34% of alliances. This is a reasonable
approximation method because capital markets, business media, and industry analysts are sensitive
to alliance announcements and track them closely in this industry (Kale, Dyer, & Singh, 2002). This
procedure is more reliable than the alternative of assuming that alliances terminate after fixed
periods of between three and five years. In total, we established that 926 (57.1%) of the 1,621 alli-
ances in our sample were terminated during the study period. We treated the remaining 695 (42.9%)
alliances as right-censored at the end of the study period.
3.2 |Explanatory variables
The explanatory variable for Hypothesis 1 is Competition between Pand PCs, which we computed
as the number of F’s partners that had at least one therapeutic area in common with Pin a year. As
Pmay have operated in areas not reflected in its alliances with F, we collected information on P’s
alliances with firms not in F’s portfolio to obtain fuller information about P’s therapeutic areas. We
constructed Competition in two steps.
First, we identified the therapeutic areas that each firm operated in through the technologies and
resources it sold or licensed to other firms. Recap identifies 22 therapeutic areas. We screened all
alliances and listed all the therapeutic areas associated with them in which a firm served as the out-
licensor (i.e., seller) of a technology or service.
Second, we counted the number of F’s portfolio partners that had at least one therapeutic area in
common with Pin the previous year. To determine the therapeutic areas in which a firm operated in
any given year, we assume that the firm was active in an area from two years prior to its out-
licensing year in that area until the termination of the licensing agreement or censoring at the end of
our study period. The count served as the measure of competition. This measure of competition
assumes Pis a source of concern even before operating in a therapeutic area, as entry would have
been signaled through clinical trials and the regulatory approval processes. Our measure evaluates
competition among partners in markets for technology (Arora, Fosfuri, & Gambardella, 2001),
where concerns about resource transfers to competitors are likely to be strong (Mansfield, 1985).
3
The explanatory variable for Hypothesis 2 is the dollar value of Equity investment in each alli-
ance, as reported by Recap. Equity investments were made in 19.9% of all alliances. We treated all
alliances for which no equity was reported as non-equity alliances.
3
The Competition measure does not include competition between F and Pbecause we are investigating competition between Pand
other partners of F; sensitivity analysis included the degree of competition between Fand Pas a control variable, finding equivalent
results.
12 ASGARI ET AL.
For Hypothesis 3, the explanatory variable is Portfolio density, which we measured as the pro-
portion of the number of ties between all partners in the focal firm’s portfolio to the total number of
possible ties in that portfolio. We computed this measure for each year of the study and used a one-
year lag. Our analysis of portfolio density reinforces our evaluation of portfolio characteristics and
complements the use of within-portfolio measures.
The explanatory variable for Hypothesis 4 is Direct paths, which is the number of ties through
which P’s resources could be transferred to PCs, other than through F, measured as the number of
PCs that have a direct path to Pthrough an alliance. We tracked these ties from formation to termi-
nation or censoring; for alliances whose termination we could not establish, we used the average
duration of ties in our sample, five years. We lagged this variable by a year. Consideration of direct
ties between Pand PCs evaluates another dimension of the interplay of competition and coopera-
tion, complementing our primary focus on competition and cooperation between Fand P.
For Hypothesis 5, the explanatory variable is Strategic similarity, an indicator variable that took
the value of 1 if the partners operated in the same four-digit SIC. The strategies, operational focus,
resources, and challenges (which define strategic similarity) for firms competing in the same indus-
try will vary, but nonetheless, tend to be more similar than those competing in different industries.
Operating within the same SIC, particularly in an industry as subject to regulatory oversight as bio-
pharma, will tend to result in similarity of strategic characteristics.
3.3 |Control variables
We controlled for three sets of factors that may influence alliance termination, as Table 2 outlines:
(a) characteristics of the alliance portfolio (Portfolio size, Portfolio closeness centrality);
(b) characteristics of the F–Palliance (F–P alliance experience, Exclusive alliance, R&D alliance,
Marketing alliance, Reciprocal interdependence, Shock-affected alliance); and (c) Fand
Pcharacteristics (F’s pre-portfolio alliance experience, F’s R&D, P’s R&D, F–P technological
relatedness, F–P technological relatedness
2
). We lagged time-varying variables by a year. We
obtained financial data from Osiris and Compustat.
Table 3 provides descriptive statistics. No major correlations arise among the predictor variables;
the largest is between direct paths and competition, at r = 0.36. Several moderate correlations are
present among controls. F’s pre-portfolio alliance experience correlates with competition (r = 0.74),
portfolio size (r = 0.76), and F’s R&D (r = 0.53). Portfolio size correlates with competition (r
= 0.77) and F’s R&D (r = 0.63). The results are consistent if we omit correlated control variables.
3.4 |Methods
To test the hypotheses, we conduct survival analyses by creating annual spells for alliances and esti-
mating the hazard of termination for each F–Palliance. We adopt frailty models, which control for
observation-specific heterogeneity by incorporating latent observation-specific effects in the hazard
function, thus accounting for time-dependent omitted variable bias (Gutierrez, 2002; Hougaard,
1986; Lancaster, 1979; Vaupel, Manton, & Stallard, 1979).
4
In a frailty model, the hazard is mod-
elled as:
4
Some focal firms may be able to identify partners resilient to the threat of competition and amenable to being managed by their part-
ners, increasing the possibility that the hazard of terminating the F–Palliance is influenced by F’s capabilities rather than by competi-
tion within the portfolio. These capabilities would lower termination hazards without affecting portfolio competition, reducing
support for the hypotheses; hence, the tests are conservative.
ASGARI ET AL.13
TABLE 2 Control variables
Variable Measure Purpose
2a. Characteristics of alliance portfolio
Portfolio
characteristics:
Portfolio size
Total number (incremented by 1 and then logged)
of F’s partners.
Addresses F’s experience with alliances, and
breadth of resources and partners that Fhas
access to (Hoffmann, 2007; Wassmer et al.,
2017); may affect F’s propensity to terminate
individual alliances.
Portfolio structure:
Portfolio closeness
centrality
Mean of closeness centrality of all partners in F’s
portfolio (based on Stanford Network Analysis
Project library; http://snap.stanford.edu/)
Controls for connectedness of the partners in F’s
portfolio in their own networks; partner
connectivity can impact the bargaining power
of partners, their ability to find substitute
partners, and thus, the stability of portfolio
(Lavie, 2007).
2b. Characteristics of focal firm-partner alliance
Alliance experience:
F–Palliance
experience
Total number of alliances between Fand Puntil
prior year, apart from the current focal alliance.
Controls for repeated ties, which build trust,
mitigate opportunism, and reduce the hazard of
termination (Gulati, 1995; Gulati & Singh,
1998).
Exclusivity of focal
firm partner
alliance:
Exclusive alliance
Dummy, taking the value 1 if alliance contract
contained exclusivity clauses. Based on
Recap’s reporting of the existence of this
clause in an alliance agreement.
Exclusivity may moderate impact of competition
among partners by preventing replication in
alliances with other firms. Also signals
commitment, aligned interests, and greater
property rights protection (Gimeno, 2004),
affecting termination.
Alliance focus: R&D
alliance
Dummy, taking value 1 if alliance primarily
focused on R&D and technology exchange, or
if a Marketing alliance also had substantial
R&D or technology.
Controls for focus and type of alliance, which
may affect termination (Kogut, 1989;
Mitchell & Singh, 1996). Also addresses
noncompetition drivers of alliance termination.
Alliance focus:
Marketing alliance
Dummy, taking value 1 if an alliance primarily
focused on marketing, or if an R&D alliance
also had a substantial marketing component.
Controls for focus and type of alliance, which
may affect termination (Kogut, 1989;
Mitchell & Singh, 1996). Also addresses
noncompetition drivers of alliance termination.
Nature of alliance
relationship:
Reciprocal
interdependence
Dummy, taking the value of 1 if interdependence
between partners was reciprocal or 0 if
interdependence was sequential (or in two
cases, pooled). Based on modification of Gulati
and Singh’s (1998) approach, relying on value
creation logic of each alliance. Based on Recap
reports.
Greater reciprocal interdependence indicates
partners’reliance on the alliance, the
complexity of coordination, and difficulty of
terminating the alliance (Gulati & Singh, 1998;
Hoehn-Weiss, Karim, & Lee, 2017; Reuer &
Devarakonda, 2016). Reciprocal dependence
generally creates greater reliance than
sequential or pooled interdependence.
Stability of alliance:
Shock-affected
alliance
Dummy, taking value 1 if focus of an alliance
was affected by the combinatorial chemistry
technological discontinuity of 1995. Computed
from Recap information.
Alliances disrupted by discontinuities are more
likely to terminate (Asgari et al., 2017;
Mitchell & Singh, 1996).
2c. Focal firm and partner characteristics
Alliance experience:
F’s pre-portfolio
alliance experience
Total number of alliances formed by Fapart from
the alliances in F’s portfolio, until prior year.
Controls for the firm’s experience and learning on
maintaining alliances (Gulati, Lavie, & Singh,
2009).
Focal firm
characteristics:
F’s R&D
Value of F’s total investment in R&D, measured
at end of each year.
Controls for firm’s technological competencies,
which indicates resources, and its attractiveness
to partners.
Partner
characteristics:
P’s R&D
Value of P’s total investment in R&D, measured
at end of each year.
Controls for partner’s technological competencies,
which indicates the potential attractiveness of
its resources to competitors and to the focal
firm.
14 ASGARI ET AL.
ht
jjXj,αj
=αjtjjXj
,ð1Þ
where α
j
, known as a frailty, is an unobserved, observation-specific effect on the hazard rate. “The
frailties are positive quantities not estimated from the data but instead assumed to have mean 1 (for
purpose of identifiability) and variance θ[which] is estimated from the data. If α
j
< 1, then the
effect is to decrease the hazard, and thus such subjects are known to be less frail that their counter-
parts. If α
j
> 1, then these frailer subjects face an increased risk”(Cleves, Gutierrez, Gould, &
Marchenko, 2010, p. 311). The goal is to estimate coefficients after removing α
j
from the model. To
do so, Equation (1) is rewritten as a survival function:
St
jjXj,αj
=St
jjXj
αj
:ð2Þ
To convert this equation into an unconditional survivor function, the frailty α
j
is removed by tak-
ing the integral of the function, which requires α
j
to have a functional form. We followed common
practice in assuming the inverse-Gaussian distribution, leading to the final hazard function:
hθtjjXj
=−
d
dt Sθtj
XjÞ
Sθtj
Xj
g−1
:ð3Þ
Due to identifiability issues, the unshared-frailty models we employed are incompatible with
Cox regression. Hence, we employed parametric methods with αj based on an inverse-Gaussian dis-
tribution with mean 1 and variance θ, and assumed a Weibull distribution for the time-to-response
link function. As the error terms for observations pertaining to a firm may be correlated, we cluster
the errors on focal firms. Thus, we estimated proportional hazard frailty models with errors clustered
for focal firms to test our hypotheses.
4|RESULTS
Table 4 presents the results. Model 1 includes controls. Models 2–6 test Hypotheses 1–5 individu-
ally. Model 7 includes all variables, while Model 8 presents a parsimonious model that incorporates
only the significant predictor interactions. All models have greater explanatory power than the
controls-only Model 1, based on log-likelihood ratio Chi-squared tests; the models with significant
interactions improve on the competition-only Model 2 (Model 5: p< .05; Model 6: p< .10; Models
7 and 8: p< .05).
The results support Hypothesis 1. The significant positive coefficient of Competition in Models
2–7 (Model 2: b = 0.034, p= .008; Model 7: b = 0.051, p= .001) shows that greater competition
TABLE 2 (Continued)
Variable Measure Purpose
Partner relatedness:
F–Ptechnological
relatedness
Relatedness (cosine similarity) between P’s
technological resources and that of all other
members of portfolio including F, based on P’s
and other portfolio members’vectors of
technological subclasses.
Controlled for overlapping technological
resources, potential value of P’s resources and
for P’s importance to Frelative to F’s other
partners (Katila et al., 2008).
Partner relatedness:
F–Ptechnological
relatedness
2
Square of technological relatedness. Controls for possibility that extreme levels of
technological relatedness may reduce
attractiveness of P’s resources because of the
great distance or similarity of resources.
(Cohen & Levinthal, 1990; Katila et al., 2008).
ASGARI ET AL.15
TABLE 3 Descriptive statistics and correlations
Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17)
(1) Competition 1
(2) Equity investment 0.04 1
(3) Portfolio density −0.04 −0.02 1
(4) Direct paths 0.36 −0.03 0.15 1
(5) Strategic similarity 0.10 0.03 0.02 0.04 1
(6) Portfolio size (logged) 0.77 0.01 −0.11 0.23 0.05 1
(7) F–Palliance experience 0.08 0.13 −0.02 0.04 −0.01 0.06 1
(8) F’s pre-portfolio alliance
experience
0.74 0.00 −0.06 0.23 0.04 0.76 0.08 1
(9) Exclusive alliance −0.01 0.25 −0.01 −0.09 −0.05 −0.05 0.00 −0.02 1
(10) R&D alliance 0.05 0.1 0.02 0.06 −0.06 −0.01 0.06 0.00 0.07 1
(11) Marketing alliance −0.04 −0.03 0.03 −0.03 0.06 −0.01 −0.05 −0.03 −0.09 −0.40 1
(12) Reciprocal interdependence 0.13 0.17 0.01 0.03 0.00 0.09 0.06 0.07 0.21 0.10 −0.10 1
(13) Shock-affected alliance 0.03 −0.05 0.00 −0.08 0.25 0.01 −0.10 0.03 −0.09 −0.09 0.06 0.03 1
(14) F’s R&D 0.56 −0.01 −0.06 0.10 0.08 0.63 0.04 0.53 −0.05 −0.02 0.03 0.03 0.09 1
(15) P’s R&D −0.13 0.03 0.09 0.07 0.12 −0.19 0.10 −0.16 0.00 −0.02 0.01 0.06 0.02 −0.27 1
(16) Portfolio closeness centrality 0.21 0.02 0.14 0.12 0.04 0.18 0.01 0.20 −0.01 0.01 −0.01 0.07 0.04 0.09 0.1 1
(17) F–Ptechnological
relatedness
0.18 0.04 0.05 0.23 0.02 0.11 0.15 0.09 −0.07 0.04 0.05 0.12 −0.08 −0.08 0.36 0.23 1
Mean 5.51 2.08 0.02 0.15 0.48 2.31 0.21 6.53 0.39 0.89 0.02 0.35 0.38 5.30 4.14 0.23 0.24
S.D. 6.86 6.01 0.08 0.45 0.5 0.77 0.54 8.18 0.49 0.32 0.14 0.48 0.49 7.87 7.12 0.04 0.22
Minimum 000001.10000000000.11 0
Maximum 41 125.8 1 5 1 3.89 4 39 1111133.1 37.3 1 0.87
16 ASGARI ET AL.
TABLE 4 Inverse-Gaussian frailty models with Weibull link-functions: Impact of competition and moderators on termination of the F–Palliance (positive hazard rate coefficient indicates
higher likelihood of termination)
1
Controls
2
(H1)
3
(H2)
4
(H3)
5
(H4)
6
(H5)
7
(H1-H5) 8 Parsimonious
H1 (+): Competition 0.034 0.033 0.026 0.042 0.050 0.051 0.057
(0.013) (0.013) (0.015) (0.013) (0.014) (0.016) (0.014)
0.008 0.011 0.075 0.002 0.000 0.001 0.000
H2 (−) Competition × Equity investment 0.000 0.000
(0.002) (0.002)
0.898 0.953
H3 (−) Competition × Portfolio density 0.730 0.530
(0.731) (0.743)
0.318 0.475
H4 (−) Competition × Direct paths −0.035 −0.033 −0.034
(0.012) (0.012) (0.012)
0.002 0.005 0.004
H5 (−): Competition × Strategic similarity −0.029 −0.027 −0.027
(0.012) (0.012) (0.012)
0.018 0.028 0.022
Equity investment −0.058 −0.059 −0.061 −0.059 −0.059 −0.059 −0.060 −0.059
(0.015) (0.015) (0.018) (0.015) (0.015) (0.015) (0.018) (0.015)
0.000 0.000 0.001 0.000 0.000 0.000 0.001 0.000
Portfolio density −0.706 −0.667 −0.670 −1.290 −1.064 −0.675 −1.496 −1.054
(0.649) (0.635) (0.634) (1.016) (0.711) (0.650) (1.092) (0.722)
0.277 0.294 0.291 0.204 0.135 0.299 0.171 0.144
Direct paths 0.206 0.119 0.120 0.060 0.605 0.124 0.530 0.593
(0.116) (0.116) (0.115) (0.118) (0.200) (0.116) (0.199) (0.201)
0.076 0.303 0.296 0.610 0.002 0.285 0.008 0.003
Strategic similarity −0.029 −0.052 −0.052 −0.053 −0.056 0.109 0.089 0.093
(0.135) (0.134) (0.134) (0.134) (0.133) (0.166) (0.167) (0.166)
0.828 0.699 0.697 0.693 0.673 0.514 0.592 0.574
ASGARI ET AL.17
TABLE 4 (Continued)
1
Controls
2
(H1)
3
(H2)
4
(H3)
5
(H4)
6
(H5)
7
(H1-H5) 8 Parsimonious
Portfolio Size (logged) 0.241 0.129 0.129 0.114 0.077 0.128 0.071 0.079
(0.138) (0.141) (0.141) (0.141) (0.145) (0.141) (0.142) (0.143)
0.080 0.361 0.361 0.418 0.592 0.362 0.620 0.580
F–Palliance experience 0.035 0.032 0.032 0.033 0.032 0.037 0.038 0.037
(0.094) (0.094) (0.094) (0.094) (0.096) (0.091) (0.094) (0.094)
0.710 0.736 0.734 0.727 0.739 0.689 0.687 0.693
F’s pre-portfolio alliance experience −0.013 −0.022 −0.022 −0.021 −0.019 −0.022 −0.019 −0.020
(0.010) (0.012) (0.012) (0.012) (0.012) (0.012) (0.012) (0.012)
0.221 0.070 0.070 0.078 0.115 0.065 0.114 0.107
Exclusive alliance −0.376 −0.387 −0.387 −0.378 −0.391 −0.380 −0.378 −0.385
(0.117) (0.116) (0.116) (0.116) (0.116) (0.118) (0.117) (0.117)
0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001
R&D alliance −0.190 −0.213 −0.213 −0.208 −0.213 −0.212 −0.209 −0.212
(0.163) (0.164) (0.164) (0.164) (0.164) (0.164) (0.164) (0.164)
0.241 0.194 0.195 0.204 0.193 0.195 0.202 0.194
Marketing alliance −0.291 −0.273 −0.273 −0.262 −0.260 −0.292 −0.270 −0.278
(0.345) (0.341) (0.340) (0.338) (0.337) (0.342) (0.335) (0.338)
0.399 0.423 0.422 0.438 0.441 0.393 0.420 0.412
Reciprocal interdependence (vs. sequential) −0.123 −0.146 −0.145 −0.148 −0.146 −0.148 −0.150 −0.149
(0.132) (0.133) (0.133) (0.132) (0.132) (0.133) (0.132) (0.132)
0.352 0.270 0.274 0.263 0.267 0.265 0.257 0.261
Shock affected alliance 0.426 0.414 0.414 0.413 0.418 0.425 0.426 0.427
(0.121) (0.122) (0.122) (0.122) (0.123) (0.120) (0.121) (0.121)
0.000 0.001 0.001 0.001 0.001 0.000 0.000 0.000
F’s R&D
(×100,000 scaling)
−0.031 −0.035 −0.035 −0.034 −0.033 −0.035 −0.032 −0.033
(0.009) (0.010) (0.010) (0.009) (0.010) (0.010) (0.009) (0.010)
0.001 0.000 0.000 0.000 0.001 0.000 0.001 0.001
−0.012 −0.012 −0.011 −0.011 −0.011 −0.012 −0.011 −0.011
18 ASGARI ET AL.
TABLE 4 (Continued)
1
Controls
2
(H1)
3
(H2)
4
(H3)
5
(H4)
6
(H5)
7
(H1-H5) 8 Parsimonious
P’s R&D
( × 100,000 scaling)
(0.009) (0.009) (0.009) (0.009) (0.009) (0.009) (0.009) (0.009)
0.201 0.218 0.220 0.222 0.230 0.219 0.233 0.229
Portfolio closeness centrality 1.668 1.436 1.437 1.378 1.309 1.412 1.260 1.293
(1.159) (1.164) (1.164) (1.184) (1.184) (1.168) (1.193) (1.184)
0.150 0.217 0.217 0.245 0.269 0.227 0.291 0.275
F–PTechnological relatedness −1.648 −1.841 −1.843 −1.867 −2.027 −1.892 −2.074 −2.064
(0.867) (0.837) (0.840) (0.828) (0.828) (0.833) (0.824) (0.825)
0.057 0.028 0.028 0.024 0.014 0.023 0.012 0.012
F–PTechnological relatedness squared 1.725 1.907 1.909 1.906 2.171 1.942 2.177 2.190
(1.181) (1.150) (1.152) (1.144) (1.131) (1.151) (1.133) (1.131)
0.144 0.097 0.097 0.096 0.055 0.091 0.055 0.053
Constant −2.754 −2.472 −2.470 −2.440 −2.388 −2.553 −2.446 −2.469
(0.416) (0.425) (0.426) (0.432) (0.426) (0.423) (0.426) (0.423)
0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Log pseudo likelihood −1311.4 −1308.6 −1308.6 −1307.8 −1305.6 −1306.9 −1303.7 −1304.1
LL ratio vs. M1 (df ) 5.6(1)** 5.6(2)* 7.2(2)** 11.7(2)*** 9.1(2)** 15.5(5)*** 14.7(3)***
LL ratio vs. M2 (df ) 0.02 (1) 1.6 (1) 6.1 (1) ** 3.5 (1) * 9.9 (4) ** 9.1 (2) **
Robust s.e. in parentheses; numbers below are p-values. Cases: 4,261 observations (1,621 alliances, 204 firms). ***p< .01, **p< .05, *p< .10 (LL ratio Chi-squared test).
ASGARI ET AL.19
between the focal firm’s partners increases the hazard of terminating the F–Palliance. The result
holds both alone (Model 2) and in combination with the moderators (Model 3–7).
The results do not support Hypothesis 2, which predicted that greater equity investment will
reduce the impact of competition on alliance termination. The interaction term Equity investment x
Competition is insignificant in Models 3 and 7 (Model 7: b = 0.000; p= .953). Although the inter-
action is insignificant, the main effect of Equity investment (Model 7; b = −0.060, p= .001) associ-
ates with reduced termination hazards. These results suggest that equity investment in an F–P
alliance does not moderate the increased hazard posed by competition between P–PCs, but instead
directly reduces the hazard of termination.
The results do not support Hypothesis 3, which predicts that greater portfolio density moderates
the negative impact of competition on alliance termination. The interaction of Competition xPortfo-
lio density has no significant impact in Models 4 or 7 (Model 7: b = 0.530; p= .475). Portfolio
density also has no direct impact on termination (Model 7: b = −1.496; p= .171).
The results support Hypothesis 4, which predicts that direct paths moderate the negative impact
of competition on alliance termination. The interaction Competition xDirect paths is negative in
Models 5 and 7 (Model 7: b = −0.033; p= .005), indicating that direct paths between Pand PCs
moderate the impact of competition on the termination of the F–Palliance. The main effect of
Direct paths is positively associated with termination hazards (Model 7; b = 0.530, p= .008), pos-
sibly because controlling for competition, P–PC alliances may substitute for F–Palliances and also
close structural holes, weakening F’s structural position. The marginal effects of this moderation are
material: a 1% increase in competition increases the hazard ratio by 0.27% for firms with no direct
ties, while a firm with one direct path experiences only a 0.09% increase (Appendix C in File S1;
Figures in Appendix D in File S1). Thus, a firm that adds a first direct path will reduce the respon-
siveness of the hazard ratio to competition from 0.27 to 0.09, that is, a two-thirds reduction.
Finally, the results support Hypothesis 5, which predicts that strategic similarity reduces the haz-
ard of alliance termination from greater competition. Models 6 and 7 (Model 7: b = −0.027; p
= .028) show that the interaction term Competition ×Strategic similarity reduces the hazard of ter-
mination of alliances. This result indicates that with greater competition among partners, alliances
between focal firms and partners that are strategically similar have lower hazards of termination.
The main effect of Strategic similarity does not influence termination hazards, indicating effective-
ness as a moderator only in the presence of competition. In the absence of Strategic similarity, the
marginal effects show that a 1% increase in competition increased the hazard ratio by 0.31%, while
a similar increase for strategically similar firms increased the hazard ratio by only 0.16%, that is, a
50% reduction (Appendix C in File S1; Figures in Appendix D in File S1).
Several control variables in Table 4 are notable. Termination hazards were higher for Shock-
affected alliances. Termination hazards were lower for Exclusive alliances and for cases with greater
F’s R&D. F–P technological relatedness had a nonmonotonic relationship with termination, first
declining and then increasing; the U-shape may indicate that greater relatedness initially facilitates
exchange, but ultimately, creates redundancy.
Two insignificant controls concerning alliance experience are also notable. F–P alliance experi-
ence was insignificant, suggesting that repeated ties with specific partners do not reduce termination
hazards. F’s pre-portfolio alliance experience moderately reduced the hazard of termination in most
models, although not in the full model. Investigations revealed that when entered alone both F–P
alliance experience and F’s pre-portfolio alliance experience were significant in some models that
also included quadratic effects, but were not significant in the model including the full set of con-
trols. These results suggest that repeated interactions with specific partners have limited impact in
20 ASGARI ET AL.
the context of the multiple and competitive contexts of alliance portfolios, where the stresses of
competition are strong; it is also possible that benefits of partner-specific experience and costs of
resource satiation and redundancy from repeated ties may offset each other. The limited results for
F’s pre-portfolio alliance experience may indicate that this experience creates an ability to replace
alliances when partners end collaboration as a result of competition.
5
In sum, the results show that competition between Pand PCs increased the hazard of termina-
tion of F–Palliances. In turn, the direct paths and strategic similarity mechanisms reduced the
effects of competition. By contrast, equity investments and portfolio density did not moderate the
effect of competition, although the focal firm’s equity investments had a direct impact on lower
termination.
4.1 |Robustness tests
We conducted four sets of robustness tests. First, Appendices A1 (Models 1–7) and A2 (Models
8–16) in File S1 report results of alternate measures, additional controls, and additional tests. The
results were robust to splitting equity investment between Fand P, assuming that each firm invests
relative to its size (Models 1–3); adding competition between Fand P(Model 5); using an ordinal
measure of interdependence that included the two cases of pooled interdependence (Gulati & Singh,
1998) (Model 6); adding a dummy for early-stage alliances (Rothaermel & Deeds, 2004) (Model 7);
adding a dummy to distinguish the pre- and post-1995 combinatorial chemistry shock periods
(Model 8); adding a count of the number of F’s partners that were also P’s partners (Model 9);
repeating the analysis with two alliances previously excluded as outliers (Model 10); replacing Port-
folio size with Structural holes, to account for F’s positional advantages (Burt, 1992) (Model 11);
replacing Portfolio size (logged) with the unlogged Portfolio size (Model 12), replacing F’s pre-
portfolio alliance experience with the logged measure (Model 13); and removing Portfolio size, F–
P’s alliance experience, and F’s pre-portfolio alliance experience because of high correlations
(Models 14–16). We also assessed possible nonlinearity in competition, finding no quadratic effect
(Model 4).
Second, we evaluated possible selection bias in our sample. As our study focuses on alliance
portfolios, we excluded firms with only a single alliance, raising the possibility that our sample only
included firms with strong cooperation and portfolio management capabilities, which consequently
had fewer alliances terminated. We therefore employed a two-stage residual technique, which treated
Portfolio size as endogenous (we did not use the more common Heckman-type models to deal with
possible selection bias because of nonlinearity in our second stage models). The first stage used all
independent and control variables of the main model, and introduced F’s knowledge diversity as the
exclusion restriction. Knowledge diversity (measured by 1 minus the Herfindahl index of the distri-
bution of the firm’s patents across different technological classes) allows a firm to form multiple
alliances and to collaborate across sectors, thus increasing its portfolio size. First-stage residuals
were added to the second-stage frailty models. These models yielded similar results to our main
5
Prior research concerning alliance experience and termination commonly finds null results. Cui et al. (2011) found that the effect of
partner firm’s partnering experience (akin to F’s pre-portfolio alliance experience) was insignificant. Park and Russo (1996) did not
find any impact of the partners’general alliance experience. Relatedly, Reuer and Ariño (2007)) found that firms that have collabo-
rated with each other in the past do not become less likely to negotiate enforcement provisions, while Reuer and Devarakonda
(2016)) found that prior ties between two partners have no effect on the absence of steering committees in alliances between them;
these results suggest that partner-specific alliance experience may fail to breed trust. Finally, Gulati et al. (2009) showed that general
alliance experience may not be a good predictor of value. Creation in alliances, while partner-specific experience, may affect value
creation only to the extent that the assets of a new partner differ from those of the firm’s prior partners. Thus, the limited effects of
experience in our study are consistent with prior work.
ASGARI ET AL.21
models. Third, we evaluated possible bias from the endogeneity of equity investments in an alliance.
Firms with greater relationship specific assets (Hennart, 1988) and dependence on their partners
(Parkhe, 1993) may invest greater equity in the alliance to safeguard against partner opportunism.
However, greater equity may also reduce alliance termination, as predicted in Hypothesis 2. We
repeated the two-stage residual inclusion models previously described, treating equity investments
as endogenous and partners’average knowledge diversity (measured as the average knowledge
diversity of the focal firm and partner weighted by size) as the exclusion restriction; average knowl-
edge diversity indicates that partners collaborate across different resource bases and are more likely
to require complex coordination as facilitated by greater equity. The results were unchanged. Fourth,
we tested a Cox proportional hazards model, which offers the advantage of not specifying a distribu-
tion for the hazard rate (Appendix B in File S1); the Cox model yielded similar results.
5|DISCUSSION
Alliance portfolio research emphasizes the positional benefits of information and power arbitrage
when firms have multiple alliances but has under-emphasized associated costs. We argue that com-
petition between partners within an alliance portfolio can be beneficial for focal firms while simulta-
neously threatening the partners. We find that competition between the focal firm’s partners
increases the risks of termination of its alliances. We also find that focal firms can moderate termi-
nation risks by allying with firms that have multiple direct links with competitors in the current port-
folio and by selecting strategically similar partners. The work contributes to alliance portfolio
research and to studies of value chain integration.
5.1 |Contributions
Our article contributes to alliance portfolio research in two ways. First, we propose mechanisms for
attenuating the impact of competition among partners within alliance portfolios. In doing so, we
extend understanding of the options available for managing the risks of resource loss between firms
within alliance portfolios. There are substantial challenges to managing the collective properties of
alliance portfolios while maintaining cooperative relationships with partners linked through competi-
tive ties. Research on alliance portfolios often underplays these complexities, treating alliance port-
folios as collections of dyads or subsets of broader stable networks. Our results suggest that it is
essential to distinguish alliance portfolios from these other structures in order to theorize effectively
and to evaluate how focal firms can establish, operate, and evolve their portfolios.
Second, we demonstrate that moderating mechanisms have differing impact on termination risks
within portfolios. We find that direct paths in the form of alliances between competitors as well as
strategic similarity between partners moderate the threat of competition. Both of these mechanisms
rely on partner selection, which suggests that they are viable choices for focal firms. By contrast,
portfolio density did not moderate the risks of competition for alliance termination, possibly because
the density construct incorporates both competitive and noncompetitive ties within the portfolio.
While portfolio density might build social cohesion along with norms of cooperation and trust, these
may be too diffuse to influence individual alliance termination decisions. Furthermore, strong social
cohesion may exacerbate P’s concerns regarding competitors in the portfolio; a dense portfolio with
associated norms of cooperation and information exchange may increase information exchange with
competitors in the portfolio, the very outcome of concerns to P.
22 ASGARI ET AL.
These implications recommend closer attention to embedded competition and to how focal firms
can manage this competition. Although studies evaluate the effects of a focal firm partnering with
competitors, the impact of competition among partners on the configuration and reconfiguration of
alliance portfolios has attracted less attention. Competition in board-interlock networks may be inci-
dental, arising from indirect ties established by independent organizations that are pursuing specific
relationships. In contrast, competition arising from direct ties established by a focal firm within its
portfolio will tend to have direct impact and requires specific management. Firms can potentially
ameliorate these effects through the design of alliances and alliance portfolios as well as by imple-
menting moderating mechanisms. In reinforcing the importance of competition among partners as a
driver of alliance termination, and consequently, of portfolio reconfiguration, we extend explana-
tions of portfolio evolution beyond their current focus on path dependence, trust, and technological
change (Asgari et al., 2017; Lavie & Singh, 2012). Alliance portfolios are ideal structures for evalu-
ating the interplay of competition and cooperation.
In parallel, the contributions reinforce the view that competition is a key influence on coopera-
tion. Though firms can choose who to compete against and cooperate with, focal firms have rela-
tively little influence on competition among their partners. Firms that establish direct ties with
competing partners may serve as conduits for unintended resources transfers, even if they formally
establish safeguards between alliances. We align with studies that highlight the risks of indirect ties
in addition to their strengths, and emphasize how embeddedness within networks can foster discord
and instability in addition to facilitating trust and sharing (Gomes-Casseres, 1996; Krackhardt,
1999). In demonstrating that actual and potential competition from partners of a current ally influ-
ence cooperation and portfolio change, we highlight the social structure of competition within port-
folios as providing a complementary perspective to portfolios as stable coalitions that evolve
through trust-based, path-dependent patterns (Gulati, 1995; Uzzi, 1996).
The study also contributes to the emerging body of work on industry ecosystems and value
chain integration (Adner & Kapoor, 2016). There is emerging recognition that firms can gain com-
petitive advantage by actively managing relationships among firms that operate within their value
chain (Mitchell, 2014). Creating and managing such eco-systems, though, typically also generates
competitive relationships among value chain partners. This article demonstrates potential limits in
value chain integration that arise from such competition and highlights mechanisms for attenuating
the limits.
5.2 |Limitations and future research
The study’s limitations suggest at least four avenues for future research. First, future studies should
explore how competition occurs within portfolios in industries with different dynamics, such as dif-
ferent product life cycles, technological foundations, intellectual property rights protection, and
bases of competition. In industries where product development times and life cycles are shorter than
in biopharmaceuticals, and where intellectual property regimes are weaker, the impact of competi-
tive rivalry and resource loss may be stronger on alliance termination.
Second, it would be valuable to investigate follow-on strategies after firms terminate alliances. It
is unlikely that a firm that terminates an alliance would eschew cooperation or alternate strategies
for accessing resources it previously obtained from the alliance. Following firms to establish their
subsequent actions will provide insights into portfolio reconfiguration.
Third, our focus on competitive ties explores only a part of the interdependencies that can affect
the evolution and performance of an alliance portfolio. For instance, task interdependencies between
partners can shape the ability of the focal firm to terminate alliances. When there is greater
ASGARI ET AL.23
interdependence, alliance value derives from a tightly knit chain of activities (Hoehn-Weiss et al.,
2017), where the sum of independent relationships is less than the total value of the alliance portfo-
lio (Vassolo et al., 2004). In such portfolios, terminating a particular tie can have cascading effects
that hinder and reduce the value of the interactions for the rest of the portfolio; this, in turn, reduces
the flexibility of the focal firm to terminate the alliance, even if the dyadic tie has little value.
Extending this reasoning to examine second-order effects of alliance formation and termination on
other alliances will be a fruitful line of research.
Fourth, it would be useful to investigate implications for value creation and capture. We find that
managing connections between competing partners can mitigate the strains imposed on the focal
firm-partner relationship. Whether such stabilization contributes to the performance of the focal firm
warrants study. Indeed, connecting portfolio interdependencies, portfolio structures, and alliance ter-
mination to value creation and value capture promises to be a productive line for research. Variance
in partners’capabilities, as well as in competitive and cooperative relationships, suggests that value
creation and capture will diverge between focal firms and partners as well as across partners. The
interplay of value creation, which in alliances takes place across firm boundaries, and value capture,
which arises within firm boundaries, offers a potentially rich context for investigating the interplay
of cooperative and competitive relationships within alliance portfolios. A related issue is how safe-
guards to prevent resource transfers between partners on firms’internal organization affect perfor-
mance, potentially hindering internal learning and sharing, and so affecting the firm’s ability to
create and capture value.
Alliance portfolios offer opportunities to examine inter-firm cooperation and the cessation of
such cooperation in contexts with more complex interactions than dyadic alliances. Our study shows
that the interplay of competition and cooperation in alliance portfolios opens opportunities for
research on how focal firms and their partners balance the benefits and challenges of cooperation
and competition. The results provide a platform for studying how alliance portfolios form, operate,
perform, evolve, and terminate.
ACKNOWLEDGEMENTS
We would like to thank Phanish Puranam, Samina Karim, the editors of the Special Issue—Andrew
Shipilov, Jeffrey Reuer, Dovev Lavie, and Werner Hoffmann—and two anonymous reviewers for
their valuable inputs. Further, Yong Xue’s excellent research assistance was instrumental for the
development of the article.
Kulwant Singh acknowledges support from National University of Singapore Research Grant R-
313-000-113-112. Will Mitchell acknowledges support from the Social Sciences and Humanities
Research Council of Canada.
REFERENCES
Adner, R., & Kapoor, R. (2016). Innovation ecosystems and the pace of substitution: Re-examining technology S-curves. Strategic
Management Journal,37(4), 625–648.
Ahuja, G. (2000). The duality of collaboration: Inducements and opportunities in the formation of interfirm linkages. Strategic Man-
agement Journal,21(3), 317–343.
Arora, A., Fosfuri, A., & Gambardella, A. (2001). Markets for technology: The economics of innovative and corporate strategy. Cam-
bridge, MA: MIT Press.
Asgari, N., Singh, K., & Mitchell, W. (2017). Alliance portfolio reconfiguration rollowing a technological discontinuity. Strategic
Management Journal,38(5), 1062–1081.
Blodgett, L. L. (1992). Factors in the instability of international joint ventures: An event history analysis. Strategic Management Jour-
nal,13(6), 475–481.
24 ASGARI ET AL.
Borys, B., & Jemison, D. B. (1989). Hybrid arrangements as atrategic alliances: Theoretical issues in organizational combinations.
Academy of Management Review,14(2), 234–249.
Bresser, R. K. F. (1988). Matching collective and competitive strategies. Strategic Management Journal,9(4), 375–385.
Burt, R. S. (1992). Structural holes: The social structure of competition. Cambridge, MA: Harvard University Press.
Cleves, M., Gutierrez, R. G., Gould, W., & Marchenko, Y. V. (2010). An introduction to survival analysis using Stata. College Sta-
tion, TX: Stata Press.
Cohen, W. M., & Levinthal, D. A. (1990). Absorptive capacity: A new perspective on learning and innovation. Administrative Sci-
ence Quarterly,35(1), 128–152.
Coleman, J. S. (1988). Social capital in the creation of human capital. American Journal of Sociology,94,95–120.
Cui, A. S., Calantone, R. J., & Griffith, D. A. (2011). Strategic change and termination of interfirm partnerships. Strategic Manage-
ment Journal,32(4), 402–423.
Dussauge, P., Garrette, B., & Mitchell, W. (2000). Learning from competing partners: Outcomes and durations of scale and link alli-
ances in Europe, North America and Asia. Strategic Management Journal,21(2), 99–126.
Garrette, B., Castañer, X., & Dussauge, P. (2009). Horizontal alliances as an alternative to autonomous production: Product expansion
mode choice in the worldwide aircraft industry 1945–2000. Strategic Management Journal,30(8), 885–894.
Gimeno, J. (2004). Competition within and between networks: The contingent effect of competitive embeddedness on alliance forma-
tion. Academy of Management Journal,47(6), 820–842.
Gnyawali, D. R., & Park, B. J. R. (2011). Co-opetition between giants: Collaboration with competitors for technological innovation.
Research Policy,40(5), 650–663.
Gomes-Casseres, B. (1996). The alliance revolution: The new shape of business rivalry. Cambridge, MA: Harvard University Press.
Greve, H. R., Baum, J. A. C., Mitsuhashi, H., & Rowley, T. J. (2010). Built to last but falling apart: Cohesion, friction, and with-
drawal from interfirm alliances. Academy of Management Journal,53(2), 302–322.
Greve, H. R., Mitsuhashi, H., & Baum, J. A. (2013). Greener pastures: Outside options and strategic alliance withdrawal. Organiza-
tion Science,24(1), 79–98.
Gulati, R. (1995). Social structure and alliance formation patterns: A longitudinal analysis. Administrative Science Quarterly,40(4),
619–652.
Gulati, R., Lavie, D., & Singh, H. (2009). The nature of partnering experience and the gains from alliances. Strategic Management
Journal,30(11), 1213–1233.
Gulati, R., & Singh, H. (1998). The architecture of cooperation: Managing coordination costs and appropriation concerns in strategic
alliances. Administrative Science Quarterly,43(4), 781–814.
Gutierrez, R. G. (2002). Parametric frailty and shared frailty survival models. Stata Journal,2(1), 22–44.
Haunschild, P. R., & Beckman, C. M. (1998). When do interlocks matter? Alternate sources of information and interlock influence.
Administrative Science Quarterly,43(4), 815–844.
Hennart, J. F. (1988). A transaction costs theory of equity joint ventures. Strategic Management Journal,9(4), 361–374.
Hernandez, E., Sanders, W. G., & Tuschke, A. (2015). Network defense: Pruning, grafting, and closing to prevent leakage of strategic
knowledge to rivals. Academy of Management Journal,58(4), 1233–1260.
Hoehn-Weiss, M. N., Karim, S., & Lee, C. H. (2017). Examining alliance portfolios beyond the dyads: The relevance of redundancy
and nonuniformity across and between partners. Organization Science,28(1), 56–73.
Hoffmann, W. H. (2007). Strategies for managing a portfolio of alliances. Strategic Management Journal,28(8), 827–856.
Hougaard, P. (1986). Survival models for heterogeneous populations derived from stable distributions. Biometrika,73(2), 387–396.
Jiang, R. J., Tao, Q. T., & Santoro, M. D. (2010). Alliance portfolio diversity and firm performance. Strategic Management Journal,
31(10), 1136–1144.
Kale, P., Dyer, J. H., & Singh, H. (2002). Alliance capability, stock market response, and long-term alliance success: The role of the
alliance function. Strategic Management Journal,23(8), 747–767.
Kapoor, R., & Adner, R. (2012). What firms make vs. what they know: How firms’production and knowledge boundaries affect
competitive advantage in the face of technological change. Organization Science,23(5), 1227–1248.
Katila, R., Rosenberger, J. D., & Eisenhardt, K. M. (2008). Swimming with sharks: Technology ventures, defense mechanisms and
corporate relationships. Administrative Science Quarterly,53(2), 295–332.
Khanna, T., Gulati, R., & Nohria, N. (1998). The dynamics of learning alliances: Competition, cooperation, and relative scope. Strate-
gic Management Journal,19(3), 193–210.
Kogut, B. (1989). The stability of joint ventures: Reciprocity and competitive rivalry. Journal of Industrial Economics,38(2),
183–198.
Krackhardt, D. (1999). The ties that torture: Simmelian tie analysis in organizations. In S. B. Andrews & D. Knoke (Eds.), Research
in the sociology of organizations (Vol. 16, pp. 183–210). Greenwhich, CT: JAI Press.
Lancaster, T. (1979). Econometric methods for the duration of unemployment. Econometrica,47(4), 939–956.
Lane, P. J., & Lubatkin, M. (1998). Relative absorptive capacity and interorganizational learning. Strategic Management Journal,
19(5), 461–477.
Lavie, D. (2006). The competitive advantage of interconnected firms: An extension of the resource-based view. Academy of Manage-
ment Review,31(3), 638–658.
ASGARI ET AL.25
Lavie, D. (2007). Alliance portfolios and firm performance: A study of value creation and appropriation in the U.S. software industry.
Strategic Management Journal,28(12), 1187–1212.
Lavie, D., & Singh, H. (2012). The evolution of alliance portfolios: The case of Unisys. Industrial and Corporate Change,21(3),
763–809.
Lerner, J., Shane, H., & Tsai, A. (2003). Do equity financing cycles matter? Evidence from biotechnology alliances. Journal of
Financial Economics,67(3), 411–446.
Mahmood, I. P., Chung, C. N., & Mitchell, W. (2013). The evolving impact of combinatorial opportunities and exhaustion on innova-
tion by business groups as market development increases: The case of Taiwan. Management Science,59(5), 1142–1161.
Mansfield, E. (1985). How rapidly does new industrial technology leak out? Journal of Industrial Economics,1(12), 217–223.
Mitchell, W. (2014). Why Apple’s product magic continues to amaze–skills of the world’s # 1 value chain integrator. Strategy and
Leadership,42(6), 17–28.
Mitchell, W., & Singh, K. (1996). Survival of businesses using collaborative relationships to commercialize complex goods. Strategic
Management Journal,17(3), 169–195.
Ocasio, W. (1997). Towards an attention-based view of the firm. Strategic Management Journal,18(S1), 187–206.
Oxley, J. E. (1997). Appropriability hazards and government in strategic alliances: A transaction cost approach. Journal of Law, Eco-
nomics, and Organization,13(2), 387–409.
Oxley, J. E., & Wada, T. (2009). Alliance structure and the scope of knowledge transfer: Evidence from U.S.-Japan agreements. Man-
agement Science,55(4), 635–649.
Park, S. H., & Russo, M. V. (1996). When competition eclipses cooperation: An event history analysis of joint venture failure. Man-
agement Science,42(6), 875–890.
Parkhe, A. (1993). Strategic alliance structuring: A game theoretic and transaction cost examination of interfirm cooperation. Academy
of Management Journal,36(4), 794–829.
Pisano, G. P. (1989). Using equity participation to support exchange: Evidence from the biotechnology industry. Journal of Law, Eco-
nomics, and Organization,5(1), 109–126.
Podolny, J. M. (2001). Networks as the pipes and prisms of the market. American Journal of Sociology,107(1), 33–60.
Porac, J. F., & Thomas, H. (1990). Taxonomic mental models in competitor definition. Academy of Management Review,15(2),
224–240.
Ramaswamy, K. (1997). The performance impact of strategic similarity in sorizontal sergers: Evidence from the U.S. banking indus-
try. Academy of Management Journal,40(3), 697–715.
Reuer, J. J., & Ariño, A. (2007). Strategic alliance contracts: Dimensions and determinants of contractual complexity. Strategic Man-
agement Journal,28(3), 313–330.
Reuer, J. J., & Devarakonda, S. V. (2016). Mechanisms of hybrid governance: Administrative committees in non-equity alliances.
Academy of Management Journal,59(2), 510–533.
Reuer, J. J., & Ragozzino, R. (2006). Agency hazards and alliance portfolios. Strategic Management Journal,27(1), 27–43.
Rothaermel, F. T., & Boeker, W. (2008). Old technology meets new technology: Complementarities, similarities, and alliance forma-
tion. Strategic Management Journal,29(1), 47–77.
Rothaermel, F. T., & Deeds, D. L. (2004). Exploration and exploitation alliances in biotechnology: A system of new product develop-
ment. Strategic Management Journal,25(3), 201–221.
Rowley, T. J., Greve, H. R., Rao, H., Baum, J. A. C., & Shipilov, A. V. (2005). Time to break up: Social and instrumental anteced-
ents of firm exits from exchange cliques. Academy of Management Journal,48(3), 499–520.
Sampson, R. (2007). R&D alliances and firm performance: The impact of technological diversity and alliance organization on innova-
tion. Academy of Management Journal,50(2), 364–386.
Schilling, M. A. (2009). Understanding the alliance data. Strategic Management Journal,30(3), 233–260.
Shipilov, A. V. (2009). Firm scope experience, historic multimarket contact with partners, centrality, and the relationship between
structural holes and performance. Organization Science,20(1), 85–106.
Shipilov, A. V., & Li, S. X. (2011). The missing link: The effect of customers on the formation of relationships among producers in
the multiplex triads. Organization Science,23(2), 472–491.
Shipilov, A. V., Li, S. X., & Greve, H. R. (2011). The prince and the pauper: Search and brokerage in the initiation of
status-heterophilous ties. Organization Science,22(6), 1418–1434.
Silverman, B. S., & Baum, J. A. C. (2002). Alliance-based competitive dynamics. Academy of Management Journal,45(4), 791–806.
Simmel, G. (1950). The sociology of Georg Simmel. Glencoe, IL: The Free Press.
Singh, K., & Mitchell, W. (1996). Precarious collaboration: Business survival after partners shut down or form new partnerships.
Strategic Management Journal,17(S1), 99–115.
Spender, J. C. (1989). Industry recipes: An enquiry into the nature and sources of managerial judgement. Oxford, England:
Blackwell.
Stuart, T. E. (1998). Network positions and propensities to collaborate: An investigation of strategic alliance formation in a
high-technology industry. Administrative Science Quarterly,43(3), 668–698.
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.
26 ASGARI ET AL.
Vassolo, R. S., Anand, J., & Folta, T. B. (2004). Non-additivity in portfolios of exploration activities: A real options-based analysis
of equity alliances in biotechnology. Strategic Management Journal,25(11), 1045–1061.
Vasudeva, G., & Anand, J. (2011). Unpacking absorptive capacity: A study of knowledge utilization from alliance portfolios. Acad-
emy of Management Journal,54(3), 611–623.
Vaupel, J. W., Manton, K. G., & Stallard, E. (1979). The impact of heterogeneity in individual frailty on the dynamics of mortality.
Demography,16(3), 439–454.
Verspagen, B., & Duysters, G. (2004). The small worlds of strategic technology alliances. Technovation,24(7), 563–571.
Wassmer, U. (2010). Alliance portfolios: A review and research agenda. Journal of Management,36(1), 141–171.
Wassmer, U., Li, S., & Madhok, A. (2017). Resource ambidexterity through alliance portfolios and firm performance. Strategic Man-
agement Journal,38(2), 384–394.
Wernerfelt, B. (1984). A resource-based view of the firm. Strategic Management Journal,5(2), 171–180.
Williamson, O. E. (1983). Credible commitments: Using hostages to support exchange. American Economic Review,73(4), 519–540.
Williamson, O. E. (1991). Comparative economic organization: The analysis of discrete structural alternatives. Administrative Science
Quarterly,36(2), 269–296.
SUPPORTING INFORMATION
Additional Supporting Information may be found online in the supporting information tab for this
article.
How to cite this article: Asgari N, Tandon V, Singh K, Mitchell W. Creating and taming
discord: How firms manage embedded competition in alliance portfolios to limit alliance ter-
mination. Strat Mgmt J. 2018;1–27. https://doi.org/10.1002/smj.2784
ASGARI ET AL.27
