Alliance Formation in the Midst of Market and Network: Insights From Resource Dependence and Network Perspectives

Abstract

Alliances are often formed as a response to challenges from both market and social forces. Although the resource dependence logic posits that firms enter into alliances to stabilize resource flows between different markets and also to increase market power in their primary industry, it remains unclear whether the social power of firms, generated from alliance networks, may motivate firms to respond differently to the dependence logic of alliance formation. By incorporating social network theory, we argue that a firm’s social network advantages in the primary industry may serve as critical contingency conditions of the dependence logic. Analyses of firms in the U.S. computer industry from 1994 to 2007 suggest that a firm’s centrality advantage marginally reduces the positive effects of market dependencies on alliance formation, whereas a firm’s brokerage advantage enhances the market dependence effect.

Keywords

alliance formation market dependence alliance network centrality brokerage

Introduction

Firms are embedded in both market and social influences (Podolny, 2001; Powell, 1990), and strategic alliances are often used by firms to address challenges from these different forces. One dominant market explanation for alliance formation, as informed by resource dependence theory (RDT; Pfeffer & Nowak, 1976), is that alliances are formed to stabilize resource flows and to enhance market power (Casciaro & Piskorski, 2005; G. F. Davis & Cobb, 2010). As such, market entry via alliances has been viewed as a form of dependence-reducing strategy (Hillman, Withers, & Collins, 2009). In contrast, alliances are also formed to cultivate interorganizational relationships, build up social networks, and gain social power (Gulati, 1998; Rosenkopf & Padula, 2008; Stuart, 1998). Although each stream of research has generated valuable insights into the determinants of alliance formation, relatively little systematic investigation has been done to understand whether the social forces complement or substitute the market forces in alliance activities. In this study, we aim to develop a theoretical framework that combines insights from both RDT and social network theory (SNT) and explore the following research question: Will firms with varying network advantages respond differently to the market demands of resource dependence in their decisions for alliance formation?

Prior research has suggested two major functions of strategic alliances. One is to enhance a firm’s market power by stabilizing the resource flow, and the other is to enhance its social power by affiliating with other important players. The market logic of RDT is that if a firm is heavily dependent on a given market, it will be constrained by actors in that market. Forming alliances with these actors can be a response to mitigate the constraints and accordingly stabilize the revenue flow and enhance a firm’s market power relative to its competitors (Barringer & Harrison, 2000; Hillman et al., 2009; Pfeffer & Nowak, 1976). Meanwhile, the social logic of SNT posits that power advantages may emerge from embedded social networks (Brass & Burkhardt, 1993; Powell, Koput, & Smith-Doerr, 1996; Shipilov, 2009). By forming alliances, a firm is likely to be situated in a unique network position, such as centrality or brokerage (Burt, 1992), which reflects a firm’s power advantage in its social hierarchy (Borgatti & Foster, 2003; Gulati, 1999; Kogut, Shan, & Walker, 1992).

This dual implication of alliances captures the two sides of the same coin. It raises an interesting question as to whether the social logic of SNT complements or substitutes the market logic of RDT in driving alliance formation. The linkage of these two types of logics relies on the well-established notion that a firm’s economic activities are embedded in its social relations (Granovetter, 1985). Both logics share many common assumptions on dependence as firms are constrained by interdependences in both market and network relationships (Hillman et al., 2009; Pfeffer, 1987). Given that a firm can also use network relationships to gain power and access resources (Bae & Gargiulo, 2004), the social logic may constrain or promote the market logic. However, our understanding of how socially embedded networks affect the dependence logic is very limited (Pfeffer, 2003). In particular, it remains unclear how a firm’s network position (e.g., centrality or structural hole) affects its reaction to the market dependence logic. In an effort to fill this gap, we argue that these two logics may not always be compatible with each other, and the compatibility (or incompatibility) relies on how a firm reproduces its patterns of interconnections in the alliance network so as to maintain its network advantage.

We theorize that the reproducing pattern of interconnection serves as a mechanism to explain why firms with different network advantages may respond to the market dependence logic of alliance formation in distinctive ways. Specifically, we contend that central firms, compared to peripheral firms, tend to reproduce the pattern of interconnections by forming alliances within the primary industry to retain their power position, reducing their response to the market dependence logic of alliance formation in a different industry. Conversely, to retain their power position of structural holes (Burt, 1992), firms tend to reproduce the pattern of interconnections by forming alliances with disconnected firms, such as firms in different industries, thereby increasing their response to the market dependence logic of alliance formation.

This study makes two major contributions. First, building on and extending the growing body of research that has made an effort in integrating RDT and SNT (for reviews, see Borgatti & Foster, 2003; Hillman et al., 2009), our study attempts to understand the compatibility of market and social logics on alliance formation with the purpose to bridge the divided research in this area. We reveal that the social logic of SNT on brokerage advantage complements the market logic of RDT in alliance formation, while the social logic of SNT on centrality advantage substitutes the market logic of RDT. This integrative approach has the potential to clarify the network boundary of RDT, an important research topic that has been emphasized by scholars in recent RDT research (e.g., Hillman et al.; Pfeffer, 2003).

Second, our integrative approach also represents an important advancement in alliance research, which has often been divided by either market or social perspectives. For example, prior research has approached the question of alliance formation from different theories, such as transaction costs economics (Hennart, 1988), resource-based theory (Eisenhardt & Schoonhoven, 1996), organizational learning (Lavie & Rosenkopf, 2006), and institutional theory (Oliver, 1997), among others. From these perspectives, alliances have often been regarded as an effective means to realize economic rationality in the market (e.g., to reduce transaction costs or to access external resources) or to pursue social influences (e.g., to strive for normative rationality in an institutional context or to gain social influences in interfirm relationships). In our view, there is a need to integrate these two different logics of alliance formation, which may provide insights into the interesting debate on the relationship between environmental determinism from an RDT perspective and strategic choice from an SNT perspective. Our study represents a worthwhile effort towards this goal.

Theoretical Background and Hypotheses

Although RDT and SNT provide distinct approaches to explain firm decisions, both share some basic assumptions on dependence and power (Hillman et al., 2009; Pfeffer, 2003), which are derived from economic and social exchanges, respectively (Burt, 1980; Emerson, 1962). Pfeffer has long recognized that organizations in RDT are “not autonomous, but rather are constrained by a network of interdependencies with other organizations” (1987: 26). In other words, firms are constrained not only by market forces in resource flows but also by social forces in interorganizational relations. Firms are thus motivated to reduce others’ market power by forming close collaborations (e.g., joint ventures) with them (Pfeffer, 2003) or seek network advantages by strategically positioning in interorganizational networks (Burt, 2002; Coleman, 1988).

The shared assumptions in SNT and RDT offer a rich avenue for theoretical integration that will advance our understanding of interorganizational relationships (Bae & Gargiulo, 2004; Burt, 1992; Elg, 2000; Gulati, 1995; Lomi & Pattison, 2006). Scholars have advocated the “blending” of different constructs, thoughts, and logics in two theories (Mayer, 2013; Okhuysen & Bonardi, 2011; Suddaby, Hardy, & Huy, 2011) that can help clarify the mechanisms of social networks in response to the dependence logic of a firm’s alliance activities.

We argue that the shared assumptions of these two theories, however, do not necessarily lead to compatible logics in predicting a firm’s strategic decisions because they emphasize different sources of power and different mechanisms to explain firm behavior. While SNT emphasizes network position as a source of social power, RDT regards alternative resources as a source of market power. Given that alliances can be used to reduce exchange uncertainties shaped by the market dependence mechanism as well as change the network patterns shaped by the social power mechanism, it is meaningful to examine the compatibility (or incompatibility) between these two theoretical perspectives (RDT and SNT) in an effort to provide a more holistic view of alliance formation. We start with the distinctive power mechanisms underlying the market and social logics and then explain why firms with centrality or brokerage advantage may have different responses to the market dependence logic.

Market Dependence Logic of Alliance Formation

In line with the market logic, RDT emphasizes that a firm’s strategic behavior is decided by its resource dependence, which is defined as “the product of importance of a given input or output to the organization” (Pfeffer & Salancik, 1978: 51). Market dependence captures the situation in which the output of one firm (e.g., a supplier) is the input for another firm (e.g., a buyer; Pfeffer, 1972b), gauging the extent to which a firm vertically relies on the end market to stabilize the economic exchange of products or services (i.e., revenue flows; Pfeffer, 1972a). When a firm sells a large proportion of its outputs to a particular market or industry, then the target market becomes a critical source of resource to the firm’s survival as its dependence on other markets is proportionally reduced (Finkelstein, 1997; Pfeffer, 1972a). A firm’s dependence concentration provides opportunities for actors in the target market to gain control as the firm has to comply with external demands from these actors. These actors in turn may create uncertainties and constraints (e.g., in terms of price, quality, and delivery time) for the focal firm.

The dependence logic predicts that firms will form alliances in the target industry to mitigate uncertainties and constraints (Pfeffer, 1976; Pfeffer & Nowak, 1976; Thompson, 1967). Essentially, the vertical integration strategy allows the firm to control critical resources (e.g., end markets; Galbraith & Stiles, 1984), exploit the partners’ complementary resources (e.g., distribution channels; Chung, Singh, & Lee, 2000; Eisenhardt & Schoonhoven, 1996; Hagedoorn, 1993; Park, Chen, & Gallagher, 2002), coordinate activities between exchange partners (Pfeffer & Salancik, 1978), and/or consolidate market power (Gimeno, 2004; Kim & Singal, 1993). Among various alliance forms, joint ventures reflect firms’ strong commitments in the target industry and are also more effective in reducing resource constraints from partners. We thus begin with the following baseline hypothesis:

Hypothesis 1: A firm’s market dependence on a given industry increases the likelihood of the firm’s joint venture formation in that industry.

An Integrative Approach of Alliance Formation

As noted, alliance activities not only have an impact on a firm’s market position but also influence the reproducing pattern of the firm’s alliance network. The social logic of SNT thus comes into play with the market logic of RDT. Researchers have repeatedly called for more research to consider the boundary conditions under which RDT is more or less predictive (Hillman et al., 2009; Pfeffer, 2003). We argue that the intersection of RDT and SNT, especially their focus on power, may provide a possible area of integration for examining the compatibility or incompatibility of market and social logics (Okhuysen & Bonardi, 2011; Suddaby et al., 2011; Zahra & Newey, 2009).

From an RDT perspective, the source of market power can be derived from a firm’s effort to achieve a coalition effect against competitors through alliances within its primary industry as well as alliances across industries to stabilize various value-chain activities (Holmqvist, 2004; Lavie, 2007; Lavie & Rosenkopf, 2006; Rothaermel & Deeds, 2004). In particular, establishing alliances across industries not only reduces a firm’s dependence on the target industry but also improves the firm’s bargaining position relative to other firms in the primary industry (Pfeffer, 1976; Pfeffer & Salancik, 1978).

Different from RDT’s focus on market power in economic exchanges, SNT emphasizes firms’ social power in social exchanges. Strategic alliances create networks in which most firms are embedded (Gulati, 1995). The social networks shape economic actions of firms by creating unique opportunities and constraints (Dacin, Ventresca, & Beal, 1999; Uzzi, 1996). Studies have shown that social power originated from alliance networks within a firm’s primary industry constitutes a valid basis to identify a firm’s network advantages, such as centrality and structural holes (Yang, Lin, & Peng, 2011). According to SNT, we argue that firms may reproduce certain alliances to maintain their network advantages.

The focus of power in both RDT and SNT carries two important implications: First, the concept of power (Emerson, 1962; Pfeffer & Salancik, 1978) is useful for guiding network research by understanding the power-maintaining mechanism via network reproduction. Second, SNT (Bonacich, 1987; Burt, 1992; Coleman, 1990) may also enrich resource dependence research by further clarifying the boundary conditions of the market dependence prediction of alliance formation.

Social Power Resulting From Alliance Networks

By developing interorganizational relationships, such as alliances, firms occupy certain network positions, which result in different types and levels of power (Bae & Gargiulo, 2004; Gulati, 1995; Guler & Guillén, 2010; Shipilov, 2009). Centrality and brokerage represent two distinct types of network advantages that a firm can leverage (Gnyawali & Madhavan, 2001). The centrality advantage refers to the extent to which a firm occupies a central position in ties to other network members, denoting a firm’s ability to access resources (Freeman, 1979). The brokerage advantage refers to the extent to which a firm connects otherwise disconnected firms in the network. As such, the firm is in a position to monitor and manipulate the flow of resources (Burt, 1992, 2000). Prior research has contrasted these advantages in different contexts, such as innovation (Ahuja, 2000; Burt, 2000; Lomi & Pattison, 2006), acquisition (Yang et al., 2011), and global expansion (Guler & Guillén).

A number of studies show that a firm’s alliance centrality and brokerage provide important sources of social power (e.g., Bae & Gargiulo, 2004; Baum, Calabrese, & Silverman, 2000; Koka & Prescott, 2008; Rosenkopf & Schilling, 2007). From this perspective, on one hand, power is associated with the degree of centrality of actors’ network positions (Bonacich, 1987; Brass, 1984). Central firms are powerful as a result of the volume and speed of resource flows from a large number of alliance partners (Gnyawali & Madhavan, 2001). Moreover, as the number of partners in the same business domain increases, a firm’s power will be enhanced because of its enhanced reputation and influence, increased social trust and support, strong protection from coalition, as well as the availability of more resources and opportunities. In contrast, firms at the periphery of the alliance network are often associated with low bargaining power as a result of the lack of such attributes (Gulati & Gargiulo, 1999; Shipilov, 2009).

On the other hand, power is associated with a firm’s network position to exploit structural holes (Burt, 1992, 2000), which is built on the assumption that “power accrues to those who are between two others” (Krackhardt, 1995: 351). Firms in this position often serve as an “intermediate actor” (Marsden, 1982) between otherwise disconnected firms. Moreover, structural holes enhance a firm’s autonomy (i.e., lack of constraint) to make strategic decisions; thus, being in the structurally autonomous position holds advantages (Burt, 1992). Specifically, firms occupying a strategic brokerage position can access complementary resources from nonredundant ties that are at the two ends of structural holes (Shipilov, 2009) or exploit emerging opportunities embedded in a wide range of disconnected firms that provide nonredundant information (Gnyawali & Madhavan, 2001). In contrast, to the extent that there are no holes, a firm’s opportunities are constrained (Krackhardt).

Reproducing Patterns of Network Advantages and Logic Compatibility

To exploit the network advantages continuously, firms have to maintain their network positions over time (Burt, 1992; Coleman, 1988). As alliance formation is part of a firm’s alliance evolution, it is useful to understand not only how firms leverage the power advantage but also how they reproduce the pattern of interconnection to maintain the power advantage. As Bourdieu stated, “The existence of a network of connections is not a natural given” (1985: 249). Instead, it is strategically constructed to “produce and reproduce lasting, useful relationships that can secure material or symbolic profits” (Bourdieu: 249). Decisions regarding alliance formation have important impacts on the direction of a firm’s network evolution through reproduction. We argue that centrality and brokerage advantages (Burt, 1992, 2000; Guler & Guillén, 2010) can be reproduced in different ways to maintain firms’ respective power positions, which may serve as critical contingencies of the dependence logic.

Central firms are more likely to seek alliance partners within the primary industry to reproduce the pattern of interconnection and maintain their advantageous positions. The basic reason is as follows. In an industry with many players, established alliance networks become a coalition against other firms outside the networks (cf. Emerson, 1962). Centrality reflects the relative power obtained by a firm as a result of its coalition with many network ties (Burt, 2000; Gargiulo & Benassi, 2000; Rowley, 1997; Shipilov, 2009), which may undermine the power of other firms in the same industry (Gimeno, 2004). Moreover, firms centrally located in the alliance network have greater control over relevant resources unavailable to those on the periphery of the network (Bonacich, 1987; Freeman, 1979; Ibarra & Andrews, 1993; Mizruchi & Blyden, 1998). As such, the reproducing pattern of centrality within the primary industry is incompatible with the dependence logic of alliance formation between industries.

In contrast, brokers maintaining their power positions tend to follow a “tertius gaudens” strategy (i.e., “the third who benefits”; Burt, 1992; Obstfeld, 2005). For brokers, it is inherently costly and difficult to sustain a lasting relationship with dissimilar actors at the two ends of structural holes (Burt, 2002). As a result, the brokerage advantage declines at an alarming rate over time (Soda, Usai, & Zaheer, 2004). To deal with the problem of the rapid decay of structural holes, brokerage firms are constantly under pressure to maintain their network advantage by seeking interorganizational relationships with other disconnected firms (Burt, 2002), the amount of which is abundant in a different industry setting. In this way, the reproducing pattern of structural holes is compatible with the dependence logic.

What remains unclear is how a firm’s power position in its alliance networks may affect its response to the dependence logic. As centrality and brokerage positions are reproduced in different ways, we propose that a firm’s centrality advantage may reduce the effect of market dependence on alliance formation, whereas its brokerage advantage may enhance the effect.

Market Dependence Effect Moderated by Network Centrality

Firms often face conflicting environmental demands (Pfeffer & Salancik, 1978). On one hand, the dependence logic suggests that it is useful for a firm to form alliances in order to stabilize resource flows between industries. On the other hand, a central firm has to maintain its power position within the primary industry. In this situation, the firm has to identify its strategic priority to cope with these conflicting demands due to limited organizational resource and attention (Cyert & March, 1963). We argue that central firms will be less responsive than peripheral forms to the dependence logic of alliance formation.

First of all, powerful actors tend to use the power to their advantages (Pfeffer & Salancik, 1978). As noted, while central firms tend to reproduce the pattern of interconnections to dominate the primary industry, peripheral firms face a higher level of uncertainty than central firms because of the lack of protection from the alliance network. Forming alliances with firms in other industries can be used as a strategy to enhance a firm’s power position in its primary industry (Emerson, 1962; Pfeffer & Salancik). In particular, controlling resources through vertical integration outside the primary industry allows the firm to increase its power (Lavie, 2007; Provan, Beyer, & Kruytbosch, 1980). As a result of the constraint from the lack of opportunities within the primary industry as compared to central firms (Ibarra & Andrews, 1993), peripheral actors may reach out to other industries to alter the power-disadvantaged position in the primary industry. In such a manner, peripheral firms are more likely than central firms to follow the dependence logic of alliance formation not only to stabilize the revenue flow to the target industry but also to enhance their market power in the primary industry.

Conversely, the market dependence effect will diminish when the degree of a firm’s network centrality is high. Network centrality also demonstrates a firm’s strategic focus on managing uncertainty in the primary industry. One of the strategic benefits of building alliance networks is to maintain competitive advantages within the firm’s primary industry. According to Emerson’s (1962) study on coalition, firms allying with certain competitors are able to consolidate their power in the primary industry because the network, as a form of coalition, allows them to deal with competitors outside of the coalition. As central firms enjoy a confluence of information and a broad array of benefits unavailable to peripheral firms, they tend to produce the pattern of interconnection to maintain the power position to protect their interests (Ahuja, 2000; Gargiulo & Benassi, 2000; Perry-Smith, 2006) and, thus, will be less induced by opportunities across industries.

For example, this pattern of strategic behavior is exemplified by a leading producer of PC games—Electronic Arts (EA) Sports. This company consolidates its central position in the PC gaming sector by forming many exclusive alliances with its partners, including the National Football League (NFL), National Collegiate Athletic Association (NCAA), and International Federation of Association Football (FIFA), with the purpose to jointly develop sports game series, such as Madden NFL, NCAA Football, and FIFA 2005. However, EA Sports is reluctant to form special ties with any single downstream platform, such as Microsoft Xbox, Nintendo Wii, or Sony PlayStation, although these platforms play a critical role in promoting the sales of EA Sports games. Taken together, we predict that a central firm will be less responsive to market dependence in other industries but, instead, will tend to maintain its power advantage through reproducing more connections in its primary industry.

Hypothesis 2: A firm’s centrality in the primary industry network weakens the effect of market dependence on the formation of its joint ventures in other industries.

Market Dependence Effect Moderated by Network Brokerage

We further argue that brokerage firms are more responsive to the dependence logic of alliance formation. Brokerage relies on unexplored market opportunities and diverse information. As noted, a firm’s brokerage advantage is difficult to sustain because actors connected through structural holes are likely to be dissimilar, and the arbitraging opportunities will dissipate rapidly (Soda et al., 2004). Burt (2002) finds that 9 out of 10 brokerage relationships decay within a year. Thus, brokerage firms must constantly search for disconnected partners to maintain and enhance this unique network position. At the same time, brokerage firms are often capable of pursuing a tertius gaudens strategy as a result of their brokering skills to establish relationships with distant actors (Shipilov, 2009), such as those in other industries.

In addition, brokerage firms connected to sparse or loosely integrated partners are more capable of accessing new information beyond industry boundaries. Prior research reveals that firms located in brokerage positions will have a higher rate of entry into emerging product markets (Lee, 2007). Nonredundant information flows via structural holes from existing alliance networks with peer firms allow brokerage firms to capture emerging opportunities in other (e.g., dependent) industries, particularly when the target industry constitutes an important source of resources for the brokerage firms and their alliance partners in the primary industry.

Taken together, on one hand, brokerage firms face declining opportunities from intermediate relationships because other firms in the primary industry may become increasingly connected via networks. On the other hand, firms in other industries, which represent viable sources of nonredundant resources, are full of structural holes. Hence, forming alliances with firms in other industries is more likely to increase their structural holes and, thus, help maintain or increase their power. Since market dependence provides guidance for alliance formation (Pfeffer & Nowak, 1976), brokerage firms are more likely to follow the dependence logic by reaching out to new nonredundant ties in dependent industries. In doing so, they are able to maintain their power position on one hand and stabilize their revenue flows on the other.

A notable example of brokerage firms’ behavior is manifested by an independent video game developer—Activision Publishing Inc. By developing diverse alliances with distinct studios, such as Infinity Ward, Pandemic Studios, and Raven Software, Activision was able to come up with creative games such as Call of Duty, Dark Reign 2, and Quake 4. Later on, with the acquisition of some alliance partners, including Raven Software and Neversoft, Activision sought to find new brokerage opportunities, especially with downstream platform providers. For instance, it formed an alliance with Sega in 2003 to publish PC versions of Sonic Adventure DX and an alliance with Glu Mobile Inc. in 2008 to bring Call of Duty to mobile phones. These alliances in the downstream value chain not only help Activision increase its power in the gaming sector but also enhance its sales in other industries. We thus predict that brokerage firms like Activision will be more responsive to market dependence in other industries. Formally,

Hypothesis 3: A firm’s brokerage in the primary industry network strengthens the effect of market dependence on the formation of its joint ventures in other industries.

Method

Research Design

The classic RDT model maintains that firms in Industry A that depend on Industry B for critical resources will take actions, such as forming alliances to stabilize flows of resources between the two industries (Pfeffer & Nowak, 1976). From this perspective, scholars have examined a set of firm behaviors corresponding to industry-to-industry transactions (Burt, 1980; Casciaro & Piskorski, 2005; Finkelstein, 1997). This industry-level approach, though helpful, falls short of suggesting direct implications for the strategies of individual firms (G. F. Davis & Powell, 1992; Frooman, 1999). To avoid any potential aggregation bias, we argue that resource dependence can be better understood at the firm level (G. F. Davis & Cobb, 2010; Finkelstein).

Building on this insight, we use a firm-to-industry approach that is regarded as more effective to test RDT at the firm level (Pfeffer, 1987). For example, consider a situation where Hewlett-Packard Company (HP) depends on a given industry, such as a film industry consisting of only two firms, Eastman Kodak Company (Kodak) and Fujifilm Holdings Company (Fuji), for selling its inkjet digital photo processing machines; HP sells its outputs through Kodak but not through Fuji. However, we may not conclude that HP depends only on Kodak and not on Fuji because HP actually depends on the film industry, or both Kodak and Fuji that constitute the film industry. The fundamental reason is that allying with either Kodak or Fuji has the same effect in terms of solving the problem of HP’s dependence on or constraint from the film industry for outputs (cf. Casciaro & Piskorski, 2005; Pfeffer & Nowak, 1976). HP did form a joint venture with Kodak to manufacture and sell inkjet digital photo processing machines in the year of 2000. Alternatively, HP can also form a joint venture with Fuji as the joint venture can also help HP to manage its dependence on the same film industry.

This conceptualization is consistent with Jacobs’s original study in which alternatives and substitutability are viewed as “two components of dependence” (1974: 50-51). In our example above, HP is able to reduce its dependence uncertainty only when its alternative sources of resources (Kodak and Fuji) are substitutable. Arguably, the firm-to-industry approach is also more appropriate than a firm-to-firm (or dyadic) approach to test RDT. In their study of environmental linkages and power from an RDT perspective, Provan et al. (1980) observe that a firm can gain power over exchange partners by entering into alliances with its competitors. A dyadic approach focusing on exchange between two partners may ignore the fact that a firm is also likely to form alliances with the exchange partner’s competitors instead of the partner itself because they are substitutable (Lavie, 2007). Thus, our study contributes to research design by developing a firm-to-industry dependence approach, which is more fine-tuned and appropriate than prior approaches, such as the industry-to-industry or firm-to-firm approaches, to test the market logic of RDT.

Data and Sample

We examined our hypotheses by using a sample of firms from the U.S. computer industry, including both the hardware (Standard Industrial Classification, SIC, codes 3571, 3572, 3575, and 3577) and the software (SIC codes 7371, 7372, 7373, 7374, and 7375) sectors, between 1994 and 2007. We chose this industry because stabilizing market sales are essential for firms to compete in the high-tech industry where alliances are often used to access external resources. The data on alliance formation and partner industries were retrieved from the Securities Data Corporation (SDC) Platinum database. The SDC database is a comprehensive source of alliance activities, particularly for U.S. firms (Schilling, 2009). We verified our alliance data by using LexisNexis and the Dow Jones News Retrieval Service.

Following Rowley, Behrens, and Krackhardt (2000), we constructed the alliance network by using computer firms defined by the above SIC codes. To be included in the network, the sample firms should also have formed at least one alliance with other computer firms during the sample period. We constructed yearly alliance matrices by obtaining a roster of 3,433 computer firms as network actors and 5,177 alliance events (ties) formed by these firms. Since the SDC database provides very limited information on alliance terminations, a common practice in alliance research is based on a 4- or 5-year moving window to account for the duration of each alliance in their samples (e.g., Bae & Gargiulo, 2004; Gulati & Gargiulo, 1999; Podolny, Stuart, & Hannan, 1996). In this study, we used a 5-year moving window for constructing the time-varying network matrices since the duration of an alliance is often about 5 years¹ (Kogut, 1988; Lavie & Rosenkopf, 2006).

We determined the alliance formation for computer firms in our observation window by first identifying within our alliance matrices 261 focal (public) firms for which financial information was available in Compustat because market sales in each industry for each firm are a primary predictor in our study. We then traced their alliance activities in all other industries. Specifically, we constructed the opportunity set of alliance activities by including not only those target industries that reported alliances with computer firms but also all other target industries that did not have such alliances. This produced a total of 735,523 observations at the firm-industry-year level during the period from 1994 to 2007. This approach is considered the most conservative in assessing effects on alliance formation (Gulati, 1995; Guler & Guillén, 2010; Rothaermel & Boeker, 2008).

Dependent Variable

The dependent variable, joint venture formation, was measured as follows. First, we used Compustat data to derive all possible four-digit SIC industries outside of all the SIC codes that comprise the computer industry with which the focal firm might form a joint venture. The event of joint venture formation was proxied as the probability of a joint venture being formed by a computer firm with other firms in a noncomputer target industry. We then counted the number of joint ventures established by the focal firm in each firm-industry-year segment during the period from 1994 to 2007. Using a count variable is consistent with the practice in previous studies examining alliance formation (e.g., Park et al., 2002; Rothaermel & Boeker, 2008).

Independent Variables

Market dependence

In line with the firm-to-industry dependence approach, this study measured market dependence as the ratio of a firm’s sales in a given target market to total sales (Carpenter & Fredrickson, 2001; Pfeffer, 1972a). It captured the extent to which a focal firm’s outputs (sales) relied on a particular industry outside of its primary industry. Accordingly, we retrieved firm revenue data from the Compustat Business Segment database for each four-digit SIC industry and calculated a firm’s market dependence on each SIC industry as the percentage of its segment data within that SIC industry over its total sales in each year. In testing the sensitivity of the results by measuring market dependence using either 3- or 5-year segment data, the results remained qualitatively the same.

Firm centrality

In this study, we used closeness centrality since it captures both direct and indirect relationships (Gulati & Gargiulo, 1999). Closeness centrality represents the ability to reach a large number of industry firms while being able to rely on a minimum number of intermediaries (Mizruchi & Blyden, 1998). We first constructed the yearly network matrix by using alliances reported during the prior 5 years and then used the command of “closeness centrality” in UCINET 6 to calculate the index for each firm within the network (Borgatti, Everett, & Freeman, 2002). The formula is as follows:

C_{c}^{'} (p_{k}) = \frac{n - 1}{\sum_{i = 1}^{n} d (p_{i}, p_{k})}

where C’_c(p_k) is the closeness centrality for firm k, d(p_i, p_k) is the path distance between firm i and k, and n is the total number of firms within the network. A firm with a high closeness centrality is able to reach all other network partners in the least number of steps.

Firm brokerage

We measured firm brokerage by using the constraint measurement of structural holes in the yearly network matrix (Burt, 1992). This measurement captures the extent to which a firm’s network is either directly or indirectly concentrated via a single contact. If a firm’s alliance partners all have formed partnerships with one another in the same industry, then the firm is highly constrained and has few structural holes. Following Soda et al. (2004), we multiplied the constraint value by −1 to capture either structural holes or a brokerage advantage (the “opposite” of constraint).

Control Variables

In keeping with existing research on alliance formation, we included several control variables to exclude possible confounding effects in our analyses. The data for these control variables were drawn from the Compustat, SDC, and National Bureau of Economic Research (NBER) patent databases.

Industry distance

Following Wang and Zajac (2007), we measured the industry distance between two industries by using SIC codes to assess the overlap between the two industries. Particularly, if the focal firm’s industry has the same four-digit SIC code as the target industry, then the distance is 0; if the first three digits of SIC codes are the same, then the distance is 0.25; if the first two digits of SIC codes are the same, then 0.5; if the first digit of SIC codes is the same, then 0.75; and the distance is 1 if none of the SIC codes is the same.

Firm size

As large firms have more resources to form alliances (Ahuja, Polidoro, & Mitchell, 2009), we controlled for firm size as measured by a firm’s annual assets during the previous year. Since the firm size distribution was skewed, we took the natural log of the variable as the firm size measurement.

Market share

To control for a firm’s market power arising from its market share in its primary industry, we controlled for its market share, measured by its sales divided by the total industry sales at the four-digit SIC level. The data of industry sales for each industry were the summed sales of public firms in Compustat with the assumption that public firms take on the lion’s share of the industry sales.

Firm knowledge stock

Knowledge assimilation is also a factor attracting alliance formation (Rothaermel & Boeker, 2008). We measured a firm’s knowledge base as the total number of patents filed by the firm during the previous 5 years in order to further control for the effects of patent propensities on alliance formation.

Industry network density

We measured industry network density by the density index of annual network in the primary industry of the focal firm. Network density captures the presence (or absence) of alliances among firms in a given industry (Bae & Gargiulo, 2004).

Alliance experience

We measured each firm’s alliance experience by counting the number of alliances formed by the focal firm with partners in each other industry outside its primary industry during the previous 5 years. In order to remedy the significant positive skewness evident in the pretransformed count measurement, alliance experience was defined as the log of 1 plus the number of alliances.

R&D intensity

A firm often relies on its absorptive capacity to make alliance formation decisions because this capacity enables it to identify, assimilate, and integrate external knowledge whenever it becomes available (Cohen & Levinthal, 1990). We controlled for this type of capacity as measured by a firm’s R&D intensity (i.e., R&D expenditure/total sales) in the previous year.

Firm diversification

We adopted the entropy measurement of corporate diversification (Hitt, Hoskisson, & Kim, 1997) to control for the potential diversification effect and to measure the dispersion level of a firm’s business in the previous year, as defined below:

C o r p o r a t e d i v e r s i f i c a t i o n = \sum (P_{i} * L n (1 / P_{i}))

where P_i is the proportion of sales during a particular product segment i, and 1/P_i is the weight given to each segment.

Prior performance

Since past financial performance affects the resources a firm can use for alliance formation as well as its managerial tendency for risk taking and future expansion (Ahuja et al., 2009; Cyert & March, 1963), we controlled for prior performance as measured using the return on assets in the previous year.

Target industry growth

We also included the target industry growth rate to control for the potential of economic activity in other industries. It is likely that computer firms will be motivated to form alliances in an industry that has high growth potential. Following prior research (P. Davis & Duhaime, 1992; Rosenkopf & Schilling, 2007), we measured this growth by using the growth rate of industry sales for each four-digit SIC industry during the previous year.

Knowledge connection

To control for the possible knowledge-seeking behavior in other industries, we control the degree of knowledge connection, which was measured as the extent to which a focal firm relies on knowledge inputs (i.e., patent citations) from a particular industry outside of its primary industry. We took several steps to construct this variable. First, we used the updated NBER (Hall, Jaffe, & Traijtenberg, 2001) and U.S. Patent and Trademark Office databases to identify all patents filed by our sampled firms during the period from 1989 to 2006 to secure a 5-year observation window. Second, we tracked all non-self-citations made by these firms. Third, we used Silverman’s (1999) U.S. Patent Class–U.S. SIC concordance to derive probability-weighted assignments for the four-digit SICs of all non-self-citations the firm filed during the previous 5 years. And fourth, we aggregated the probability-weighted SIC assignments over the firm’s entire patent portfolio to determine the knowledge connection of the firm on each four-digit SIC industry.

Finally, we included year and industry dummy variables in our analyses to control for any unobserved heterogeneities arising from changing macroeconomic conditions and industry differences, respectively.

Analysis

Since our dependent variable is a count variable with high variance relative to the mean, we used negative binomial models to analyze the data as ordinary least squares regression techniques are inappropriate. This method has been widely used in prior studies to overcome the overdispersion problem associated with count dependent variables (Haunschild & Beckman, 1998). In this study, we used the Stata command “xtnbreg” with fixed effects to address the potential interdependence concern in our longitudinal data set. As discussed in the Post Hoc Analyses section below, we also tested the robustness of the results by using the zero-inflated negative binomial regression model (ZINB) to handle the potential “excess zeros” issues in the data (Vuong, 1989). As noted above, we have lagged all the time-varying independent and control variables by 1 year in all the regression analyses so as to avoid potential simultaneity problems.

In addition, since we chose our focal firms on the basis of the availability of their financial information in Compustat and many firms that were not public were thus excluded from our sample, we conducted a Heckman two-stage selection procedure to address this concern. Specifically, we followed prior research on the antecedents of initial public offering (IPO) and identified several key predictors, such as prior performance, firm size (Walters, Kroll, & Wright, 2010), intellectual property position in knowledge stock (Lerner, 1994), industry growth rate (Aslan & Kumar, 2011), and industry trend of IPO, to model the likelihood of being a public firm in the first stage (1 = public, 0 = nonpublic). For public firms, we used only the year in which they went public as the event date and measured the predictor variables 1 year ahead. We compiled a sample of nonpublic firms with a similar firm size in the same event date to represent the nonevent data. The variable of industry trend of IPO works as an instrumental variable, which is theoretically related to the first-stage dependent variable but not to the second-stage dependent variable of JV formation. We then generated the inverse Mills ratio from the first stage and controlled it in the second-stage model for predicting alliance formation.

Results

Table 1 presents the correlations and descriptive statistics of the study variables, and Table 2 presents the findings of the hierarchical regression models. Following Aiken and West (1991), we mean centered the predictor variables before generating the interaction terms. After assessing the potential multicollinearity threat by estimating the variance inflation factors (VIFs), we found that no variable has a VIF greater than 1.58, well below the recommended ceiling of 10 (Kleinbaum, Kupper, & Muller, 1988).

Table 1

Correlations and Descriptive Statistics

Variables	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	M	SD
1. Joint Venture Formation																0.00	0.02
2. Industry Distance	.01															0.06	0.13
3. Firm Size	.02	−.02														0.39	2.03
4. Market Share	.02	.03	.59													0.07	0.17
5. Firm Knowledge Stock	.02	.06	.72	.55												2.88	2.40
6. Industry Network Density	.01	.09	.02	.05	.12											0.00	0.00
7. Alliance Experience	.04	.00	.56	.45	.57	.21										8.09	17.83
8. R&D Intensity	.00	.00	−.08	−.04	−.05	.04	−.05									0.27	2.16
9. Firm Diversification	.02	−.02	.50	.43	.49	−.06	.30	−.03								0.26	0.45
10. Prior Performance	.00	.01	−.02	−.08	.00	.23	.02	−.12	−.02							0.43	0.42
11. Target Industry Growth	.00	.00	−.01	.00	−.01	.02	.00	.00	.00	.01						0.14	6.97
12. Knowledge Connection	.13	.04	.06	.06	.06	.03	.10	.00	.03	.00	.00					0.03	32.21
13. Inverse Mills Ratio	.00	.01	.08	.07	.09	.07	.06	.00	.01	.00	.00	.01				0.07	0.52
14. Market Dependence	.03	.14	.01	.02	.01	.02	.02	.00	.02	.00	.00	.06	.00			0.00	0.02
15. Firm Centrality	.01	.02	.19	.14	.17	.57	.30	.00	.03	.14	.01	.04	.05	.00		3.01	3.72
16. Firm Brokerage	.00	−.04	.19	.09	.10	−.15	.13	.00	.03	−.07	−.01	.00	.03	.00	−.14	−0.32	0.38

Note: N = 735,523. Correlations above |.01| are significant at the .01 level.

Table 2

Negative Binomial Regression on Alliance Formation

Variables	Model 1	Model 2	Model 3	Model 4
Industry Distance	2.01***	1.75***	1.71***	1.87***
	(5.93)	(5.17)	(5.04)	(5.66)
Firm Size	0.67***	0.74***	0.74***	0.74***
	(5.04)	(5.25)	(5.27)	(5.26)
Market Share	−0.07	0.37	0.35	0.37
	(–0.09)	(0.46)	(0.42)	(0.45)
Firm Knowledge Stock	0.04	−0.04	−0.04	−0.04
	(0.46)	(–0.45)	(–0.47)	(–0.43)
Industry Network Density	7,992.62***	6,114.93***	5,919.11***	5,825.89***
	(6.68)	(4.12)	(3.98)	(3.91)
Alliance Experience	−0.00	−0.00	−0.00	−0.00
	(−1.14)	(−1.21)	(−1.19)	(−1.23)
R&D Intensity	0.11	0.11	0.11	0.12
	(0.34)	(0.31)	(0.30)	(0.33)
Firm Diversification	−0.12	−0.16	−0.16	−0.17
	(−0.81)	(−1.05)	(−1.07)	(−1.13)
Prior Performance	0.36	0.12	0.08	0.13
	(0.85)	(0.21)	(0.15)	(0.24)
Target Industry Growth	0.08***	0.08**	0.08**	0.08**
	(3.37)	(3.28)	(3.15)	(3.12)
Knowledge Connection	0.00***	0.00***	0.00***	0.00***
	(10.27)	(8.15)	(8.29)	(8.96)
Inverse Mills Ratio	−0.09	−0.10	−0.10	−0.09
	(−1.03)	(−1.17)	(−1.17)	(−1.14)
Market Dependence		3.04***	3.36***	1.75*
		(6.57)	(7.20)	(2.30)
Firm Centrality		0.07**	0.08***	0.07***
		(3.26)	(3.50)	(3.39)
Firm Brokerage		0.51*	0.51*	0.35
		(2.41)	(2.42)	(1.62)
Firm Centrality × Market Dependence			−0.17†
			(−1.66)
Firm Brokerage × Market Dependence				6.21**
				(2.72)
N	735,523	735,523	735,523	735,523
Log likelihood	−2,139	−2,119	−2,118	−2,114

Note: Unstandardized coefficients are reported with z values in parentheses.

†

p < .10.

p < .05.

p < .01.

***

p < .001.

Hypothesis 1 predicts that a firm’s market dependence on a given industry for sales will increase the firm’s likelihood of forming joint ventures in that industry. Model 2 in Table 2 indicates that the coefficient of market dependence is both positive and significant (at p < .001), thus supporting Hypothesis 1.

Hypothesis 2 investigates the negative moderating effect of a firm’s centrality on the relationship between market dependence and alliance formation. Our findings show that the coefficient for the interaction between firm centrality and market dependence is negative and marginally significant (at p < .10) in Model 3, thus providing marginal support to Hypothesis 2 that a firm’s centrality reduces the effect of market dependence on joint venture formation.

Model 4 examines Hypothesis 3. The interaction between brokerage and market dependence is positive and significant (at p < .01) in Model 4, thus supporting Hypothesis 3 that brokerage strengthens the effect of market dependence on joint venture formation.

The interaction plots in Figure 1 provide additional insights to show patterns that are consistent with the findings reported above. In particular, Figure 1a suggests that peripheral firms are more likely than central firms to engage in joint ventures when these firms have a high market dependence on the target industry; this confirms Hypothesis 2. Figure 1b confirms our finding for Hypothesis 3, suggesting that firms with a higher degree of brokerage are more likely than firms with a lower degree of brokerage to engage in joint ventures when they have high market dependence on the target industry.

Figure 1

Interaction Plots

We supplemented the above analyses by conducting several sensitivity tests as follows. First, we followed Penner-Hahn and Shaver (2005) to test the robustness of our interaction effects by dichotomizing the moderator variable. This is regarded as a more fine-grained examination of the nonlinear dependent variable. Specifically, we examined the marginal effect of the moderated variables (i.e., firm centrality and brokerage) on the likelihood of alliance formation by splitting the sample by the means of these moderators, respectively. As Penner-Hahn and Shaver suggest, this approach helps reconcile the nonlinearity issue in the model by focusing on differences in the marginal effect. By comparing the high and low of the marginal effects of the moderators in each subsample, we derived the same interpretation as the interaction terms reported in Table 2.

Second, we experimented with a ZINB to deal with the potential problem of “excess zeros” in the data (Vuong, 1989). This analysis first runs a logit model to demonstrate the probability of forming alliances between each SIC industry and then uses a negative binomial model to estimate the probability of positive outcomes. The ZINB estimation remains qualitatively the same as our reported findings.

Finally, since we used only a 1-year lag for our predicator variables in the above analyses, we also experimented with a 2-year lag between the predictor variables and the alliance formation to see whether a lasting effect exists. We found that all patterns remained consistent, while in the new analyses the level of significance was weaker for market dependence (p < .01), compared with the results in Table 2; this indicates that the dependence effect is robust although weakening over time. These findings parallel Finkelstein’s (1997) findings on acquisitions, suggesting that the dependence effect on alliance formation in a given industry diminishes over time. One important reason is that once a firm has established joint ventures in a target industry, the firm can be less motivated to establish more subsequent joint ventures in the same industry. This is because the established joint ventures are able to help the focal firm to stabilize the resource flow to a certain degree. As such, the market dependence effect on firm strategy may diminish over time (Finkelstein).

Discussion

Although researchers have approached the question of alliance formation from diverse theoretical angels, our understanding of this phenomenon is constrained by two divided streams of research on either market forces or social forces in driving alliance formation (Chung et al., 2000; Eisenhardt & Schoonhoven, 1996; Gimeno, 2004; Gulati, 1995, 1999; Rosenkopf & Padula, 2008; Walker, Kogut, & Shan, 1997). Our study extends this stream of research by integrating insights on market power and social power originating from two different theories, namely, RDT and SNT, and investigating whether they complement or substitute each other in predicting alliance formation. Our results show that the power derived from a firm’s alliance networks in its primary industry cannot be ignored, indicating that the interplay of resource dependence and social network theories on power provides useful and additional insight on the complex decision of alliance formation.

Specifically, our findings support the dependence logic by showing that firms tend to focus on their dependent market and form alliances to reduce uncertainties associated with the market. More importantly, we extend the baseline logic and find that firms with power from a central network position are less responsive to the market dependence effect because they tend to reproduce the pattern of interconnection focusing on the primary industry. In contrast, firms with power from a brokerage position are more responsive to the dependence logic by forming alliances in other industries.

Theoretically, these findings have important implications for the long-running debate on whether environmental determinism and strategic choice are incommensurate (Child, 1972) or can be integrated (Hrebiniak & Joyce, 1985). Since the phenomenon of alliance formation is complex, it is difficult to suggest that the phenomenon is predictable within a single choice or deterministic framework. Instead, there is a trade-off as choice and determinism can jointly affect organizational decisions (Lawless & Finch, 1989). Our integrative approach represents a unique blend of determinism and choice, which recognizes both the importance of dependence environments that determine firm decisions (Pfeffer & Salancik, 1978) and the importance of network positions that permit strategic choice (Gulati, 1995; Rothaermel & Boeker, 2008) in a longitudinal setting. The longitudinal setting allows us to capture the dynamic process in which network advantages can alter firms’ decisions shaped by certain environmental conditions. The evidence that supports Hypotheses 2 and 3 indicates that the choice perspective is more applicable to firms with network advantages. In contrast, firms without such advantages have to forfeit certain strategic choices. Thus, the deterministic perspective is more appropriate to explain decisions made by firms without such advantages.

Implications for Resource Dependence Research

Our study provides an interactive framework that combines insights from both RDT and SNT for a better understanding of strategic decisions. One of the ways to facilitate theory building at the intersection of different theories is to explore the interactive relationships of these theories (Okhuysen & Bonardi, 2011; Zahra & Newey, 2009). As noted by Fry and Smith, “Being more explicit about defining a theory’s boundaries would help us look for conditions that would lead to modification of our theories” (1987: 124). RDT is one of the primary theories to explain interorganizational relationships (Barringer & Harrison, 2000; Hillman et al., 2009; Oliver, 1990). The focus on power has been one of its main characteristics (Pfeffer, 2003). Our study enriches the RDT literature by showing that power derived from alliance networks affects firms’ responses to the dependence logic.

Our findings imply that resource dependence and social network research may complement in meaningful ways. As Pfeffer noted RDT “is filled with network and relationship imagery” (2003: xii), and it can be extended by using network approaches. Since the 1980s, researchers have integrated network methodologies to capture the complexity of resource dependence. The industry structural autonomy model initially introduced by Burt (1980) has enriched our understanding of relative power embedded in industry input-output networks. Scholars have since incorporated this network approach into the resource dependence analysis of interorganizational activities in different settings (Casciaro & Piskorski, 2005; Finkelstein, 1997; Palmer, Barber, Zhou, & Soysal, 1995), revealing the promise of integrating RDT and SNT. However, the lack of firm-level constructs has limited our ability to fully understand the impacts of resource dependence on firm strategies (Finkelstein).

Social network research complements resource dependence research because the social network literature is rich in firm-level network constructs (Borgatti & Foster, 2003), and some of them, such as network centrality and brokerage, have been widely conceptualized as indicators of power in network analyses (e.g., Bae & Gargiulo, 2004; Bonacich, 1987; Brass & Burkhardt, 1993; Ibarra & Andrews, 1993; Mizruchi & Blyden, 1998). Our study offers additional insights to combine resource dependence and social network perspectives, as advocated by Borgatti and Foster, Hillman et al. (2009), and Pfeffer (2003), among others. Since Finkelstein’s (1997) study on the boundary conditions of the resource dependence effect on corporate acquisitions, it has become a research trend in the RDT literature (Casciaro & Piskorski, 2005; Pfeffer). However, the progress has been slow, as indicated by Hillman et al. Our findings add to this body of research by showing that firms with centrality and brokerage advantages respond differently to market dependence.

Implications for Network Research and Alliance Formation

Our resource dependence approach also complements alliance network research by offering a power explanation of network evolution. Because of its dual implications, a firm’s alliance decision that follows the dependence logic to stabilize the flow of resources is also subject to the reproducing mechanism of interconnection. That is, alliance formation as part of a firm’s network evolution can be influenced by the reproduction of different patterns of interorganizational connections to maintain its social power. Prior network research predominantly focuses on alliance formation within a given network boundary, but it stops short of investigating alliance formation across industries (Ahuja, 2000; Dyer & Nobeoka, 2000; Rowley et al., 2000). Our study echoes recent research (Guler & Guillén, 2010), suggesting that a firm’s network advantage can also influence its alliance activity outside the embedded network.

Our integrative approach also advances network research on alliance formation from an RDT perspective, which provides an externally focused theory to explain why firms ally with other firms (Barringer & Harrison, 2000; Oliver, 1990), complementing earlier studies on alliance formation in different research streams and disciplines (e.g., Ahuja, 2000; Ahuja et al., 2009; Gulati, 1995, 1999; Gulati & Gargiulo, 1999; Rosenkopf & Padula, 2008; Rothaermel & Boeker, 2008). Although scholars have used network centrality and structural holes to capture a firm’s advantaged network position in different settings (Ahuja; Burt, 2000; Dyer & Nobeoka, 2000; Lomi & Pattison, 2006; Podolny & Stuart, 1995; Powell et al., 1996), the potential synergy between resource dependence and SNTs, which offers new insights, has not been fully explored. Building on existing studies that have focused on the main effect of network advantages (Ahuja; Guler & Guillén, 2010; Podolny & Stuart), we introduce a firm’s network advantages as contingencies to explore the boundary conditions of the dependence logic of alliance formation and provide additional insights.

Our study also provides insight to bridge the network approaches that have been widely divided into “structuralist vs. connectionist” perspectives (Borgatti & Foster, 2003: 1002). The structuralist approach focuses on elaborating patterns of interconnections but does not explicitly address resource flows (e.g., Burt, 1992; Coleman, 1990). In contrast, the connectionist approach views the same connections as “pipes” over which resource flows (Borgatti & Foster). In this sense, RDT reflects the connectionist perspective that emphasizes the flow of resources through alliance formation, whereas the network approach reflects the structuralist perspective that highlights the importance of exploiting a firm’s existing alliance network position. A notable research gap in the literature is under what conditions these two perspectives are complements or substitutes. Our study provides an answer by differentiating the centrality and structural holes as a result of their different patterns of reproducing interconnections.

Managerial Implications

Our results also provide a message for managers. Firms often utilize a network of interorganizational relationships to gain power (Bae & Gargiulo, 2004; Hillman et al., 2009; Provan et al., 1980). Managers need a better understanding of the multiple perspectives given that alliance formation as an important strategic choice consists of complex considerations. Dependence pressures originate not only from firms’ primary industries but also from other industries upon which they depend for resources. An appropriate alliance strategy may enhance a firm’s power in its primary industry. In addition, when forming alliances, managers should be aware of their network advantages that may have potential influences on the resource dependence effect.

Our findings suggest that firms with different network advantages will face different challenges when leveraging alliances to enhance their positions within their primary industries. Different types of alliance network building are not equally important to a given firm. Firms with a centrality advantage tend to exploit their power advantage within the primary industry. These findings are particularly useful for peripheral firms. Such firms can actively engage in alliance activities in response to market dependence pressures so as to remedy their unfavorable positions within their primary industries.

In addition, firms with a brokerage advantage can enhance their power position by forming alliances when managing market dependence. Brokerage firms with the skill to bridge disconnected firms in the primary industry may also leverage the skill and capability to bridge disconnected actors across industries. Managers concerned about their firms’ power positions should be aware of this complexity when making appropriate strategic responses.

Limitations and Directions for Future Research

These findings should be viewed in light of the limitations of this study, which also provide opportunities for future research. First, we tested our hypotheses on the basis of one knowledge-intensive industry, the U.S. computer industry; this raises questions regarding the generalizability of our findings. Arguably, our conceptualization can be applied to explain market dependence and its contingency conditions in other high-tech industries. We believe that as a first step to extend RDT in this direction by focusing on its interplay with SNT on power, this conceptualization may trigger future studies in different research settings.

Second, the archival data in this study do not allow us to capture the decision-making process of alliance formation. Future research may combine both qualitative and quantitative research methodologies to advance our knowledge about the process (Jick, 1979). For example, this process can be revealed by thoroughly interviewing top managers who have been involved in the decision-making process. Moreover, case studies can be used to generate a pictorial image of the process. The qualitative data may help enhance the validity of findings based on archival data.

Third, it is important to consider when and how network advantage changes over time. It will be interesting to go beyond the structural aspects of alliance networks in future studies to examine how the idiosyncratic nature of each alliance, such as power dynamics within an alliance, may evolve (Mitchell, Agle, & Wood, 1997) and affect firms’ decisions for subsequent alliance formation. Meanwhile, information on termination dates of alliance is very limited as firms often opt not to announce their alliance terminations (Bae & Gargiulo, 2004). Future research may use fine-grained information to measure alliance duration. It is also important for future research to incorporate when and how a network advantage changes over time by drawing on the notion of stakeholder salience when integrating insights from RDT and SNT to understand the salience of strategic decision making at different points in time.

Fourth, among the various boundary conditions of the market dependence effect, this study focuses on a firm’s power position within the network. However, the network is by no means the only boundary condition that either enhances or weakens RDT predictions. Future studies may coalesce with other theoretical perspectives, such as stakeholder theory or institutional theory, to advance our understanding of this important subject. Moreover, given that the data were not available for the termination date of alliances, we followed prior practice in using a 5-year moving window for constructing the time-varying network matrices (Kogut, 1988; Lavie & Rosenkopf, 2006). Future research may use more fine-grained measures to address this issue.

Fifth, in this study, we examined only how a firm’s market dependence may jointly interact with its network advantages in deciding its decisions for interindustry alliance formation. Since a firm is dependent on both input and output markets, future research may further explore how a firm’s dependence on the factor market may drive its decisions for alliance formation.

Finally, whereas prior RDT studies have largely focused on mergers and acquisitions (M&As), other interorganizational relationships, such as alliances, have been relatively understudied (Barringer & Harrison, 2000). Unlike M&As, alliances like joint ventures do not completely absorb dependencies; however, an alliance strategy gives a firm greater flexibility, particularly when M&As are not viable. Our focus on joint venture formation does not preclude the possibility of M&As as another effective way to reduce dependence. It would be worthwhile to explore how firms choose between alliances and acquisitions in this respect and to compare the relative implications in response to different magnitudes of dependence pressures.

Conclusion

This study has shown that the interplay of SNT with RDT reveals how a firm’s network advantage may constrain or promote the dependence logic of alliance formation, allowing us to probe into the conversation between the two theories. The results show that it is worthwhile to investigate whether firms with centrality and brokerage advantages respond differently to the dependence logic of alliance formation, given that the firms with these network advantages may try to retain their social power positions through different reproducing mechanisms. Nonetheless, our understanding of the complex alliance activity is still incomplete. Future research in other settings is warranted to validate the usefulness of this integrative approach for further understanding firm decisions on alliance formation.

Footnotes

Acknowledgements

This article was accepted under the editorship of Patrick M. Wright. We greatly appreciate the constructive comments from Sucheta Nadkarni and two anonymous reviewers. This research was supported in part by research grants from City University of Hong Kong (CityU 7004133), the National Natural Science Foundation of China (71428005 and 71332007), the National Social Science Foundation of China (12&ZD205), and the Research Project of the Ministry of Education of China (13JZD017).

Notes

References

Ahuja

2000. The duality of collaboration: Inducements and opportunities in the formation of interfirm linkages. Strategic Management Journal, 21: 317-343.

Ahuja

Polidoro

Jr. Mitchell

2009. Structural homophily or social asymmetry? The formation of alliances by poorly embedded firms. Strategic Management Journal, 30: 941-958.

Aiken

L. S.

West

S. G.

1991. Multiple regression: Testing and interpreting interactions. London: Sage.

Aslan

Kumar

2011. Lemons or cherries? Growth opportunities and market temptations in going public and private. Journal of Financial and Quantitative Analysis, 46: 489-526.

Bae

Gargiulo

2004. Partner substitutability, alliance network structure, and firm profitability in the telecommunications industry. Academy of Management Journal, 47: 843-859.

Barringer

B. R.

Harrison

J. S.

2000. Walking a tightrope: Creating value through interorganizational relationships. Journal of Management, 26: 367-403.

Baum

J. A. C.

Calabrese

Silverman

B. S.

2000. Don’t go it alone: Alliance networks and startups’ performance in Canadian biotechnology. Strategic Management Journal, 21: 267-294.

Bonacich

1987. Power and centrality: A family of measures. American Journal of Sociology, 5: 1170-1182.

Borgatti

S. P.

Everett

M. G.

Freeman

L. C.

2002. UCINET for Windows: Software for social network analysis. Harvard, MA: Analytic Technologies.

10.

Borgatti

S. P.

Foster

P. C.

2003. The network paradigm in organizational research: A review and typology. Journal of Management, 29: 991-1013.

11.

Bourdieu

1985. The forms of capital. In Rich Pressdson

J. G.

(Ed.), Handbook of theory and research for the sociology of education: 241-258. New York: Greenwood Press.

12.

Brass

D. J.

1984. Being in the right place: A structural analysis of individual influence in an organization. Administrative Science Quarterly, 29: 518-539.

13.

Brass

D. J.

Burkhardt

M. E.

1993. Potential power and power use: An investigation of structure and behavior. Academy of Management Journal, 36: 440-470.

14.

Burt

R. S.

1980. Autonomy in a social topology. American Journal of Sociology, 85: 892-925.

15.

Burt

R. S.

1992. Structural holes: The social structure of competition. Cambridge, MA: Harvard University Press.

16.

Burt

R. S.

2000. The network structure of social capital. In Sutton

R. I.

Staw

B. M.

(Eds.), Research in organizational behavior, vol. 22: 345-423. Greenwich, CT: JAI Press.

17.

Burt

R. S.

2002. Bridge decay. Social Networks, 24: 333-363.

18.

Carpenter

M. A.

Fredrickson

J. W.

2001. Top management teams, global strategic posture, and the moderating role of uncertainty. Academy of Management Journal, 44: 533-545.

19.

Casciaro

Piskorski

M. J.

2005. Power imbalance, mutual dependence, and constraint absorption: A close look at resource dependence theory. Administrative Science Quarterly, 50: 167-199.

20.

Child

1972. Organizational structure, environment and performance: The role of strategic choice. Sociology, 6: 1-22.

21.

Chung

Singh

Lee

2000. Complementarity, status similarity, and social capital as drivers of alliance formation. Strategic Management Journal, 21: 1-22.

22.

Cohen

W. M.

Levinthal

D. A.

1990. Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly, 35: 128-152.

23.

Coleman

1988. Social capital in the creation of human capital. American Journal of Sociology, 94: 95-120.

24.

Coleman

1990. Foundations of social theory. Cambridge, MA: Belknap Press.

25.

Cyert

R. M.

March

J. G

. 1963. A behavioral theory of the firm. Englewood Cliffs, NJ: Prentice Hall.

26.

Dacin

Ventresca

Beal

1999. The embeddedness of organizations: Dialogue & directions. Journal of Management, 25: 317-356.

27.

Davis

G. F.

Cobb

2010. Resource dependence theory: Past and future. In Schoonhoven

C. B.

Dobbin

(Eds.), Research in the sociology of organizations, vol. 28: Stanford’s organization theory renaissance, 1970–2000: 21-42. Bradford, UK: Emerald Group.

28.

Davis

G. F.

Powell

W. W.

1992. Organization-environment relations. In Dunette

M. D.

Hough

L. M.

(Eds.), Handbook of industrial and organizational psychology: 315-376. Palo Alto, CA: Consulting Psychologists Press.

29.

Davis

Duhaime

1992. Diversification, vertical integration, and industry analysis: New perspectives and measurement. Strategic Management Journal, 13: 511-524.

30.

Dyer

J. H.

Nobeoka

2000. Creating and managing a high-performance knowledge-sharing network: The Toyota case. Strategic Management Journal, 21: 345-367.

31.

Eisenhardt

Schoonhoven

1996. Resource-based view of strategic alliance formation: Strategic and social effects in entrepreneurial firms. Organization Science, 7: 136-150.

32.

Elg

2000. Firms’ home–market relationships: Their role when selecting international alliance partners. Journal of International Business Studies, 31: 169-177.

33.

Emerson

R. M.

1962. Power-dependence relations. American Sociological Review, 27: 31-41.

34.

Finkelstein

1997. Interindustry merger patterns and resource dependence: A replication and extension of Pfeffer (1972). Strategic Management Journal, 18: 787-810.

35.

Freeman

L. C.

1979. Centrality in social networks: Conceptual clarification. Social Networks, 1: 215-239.

36.

Frooman

1999. Stakeholder influence strategies. Academy of Management Review, 24: 191-205.

37.

Fry

L. W.

Smith

D. A.

1987. Congruence, contingency, and theory building. Academy of Management Review, 12: 117-132.

38.

Galbraith

Stiles

1984. Merger strategies as a response to bilateral market power. Academy of Management Journal, 27: 511-524.

39.

Gargiulo

Benassi

2000. Trapped in your own net? Network cohesion, structural holes, and the adaptation of social capital. Organization Science, 11: 183-196.

40.

Gimeno

2004. Competition within and between networks: The contingent effect of competitive embeddedness on alliance formation. Academy of Management Journal, 47: 820-842.

41.

Gnyawali

D. R.

Madhavan

2001. Cooperative networks and competitive dynamics: A structural embeddedness perspective. Academy of Management Review, 26: 431-445.

42.

Granovetter

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

43.

Gulati

1995. Social structure and alliance formation patterns: A longitudinal analysis. Administrative Science Quarterly, 40: 619-652.

44.

Gulati

1998. Alliances and networks. Strategic Management Journal, 19: 293-317.

45.

Gulati

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

46.

Gulati

Gargiulo

1999. Where do interorganizational networks come from? American Journal of Sociology, 104: 1439-1493.

47.

Guler

Guillén

M. F.

2010. Home-country networks and foreign expansion. Academy of Management Journal, 53: 390-410.

48.

Hagedoorn

1993. Understanding the rationale of strategic technology partnering: Interorganizational modes of cooperation and sectorial differences. Strategic Management Journal, 14: 371-385.

49.

Hall

B. H.

Jaffe

A. B.

Traijtenberg

2001. The NBER patent citations data file: Lessons, insights and methodological tools. NBER working paper 8498, National Bureau of Economic Research, Cambridge, MA.

50.

Haunschild

P. R.

Beckman

C. M.

1998. When do interlocks matter? Alternate sources of information and interlock influence. Administrative Science Quarterly, 43: 815-844.

51.

Hennart

J. F.

1988. A transaction cost theory of equity joint ventures. Strategic Management Journal, 9: 36-74.

52.

Hillman

A. J.

Withers

M. C.

Collins

B. J.

2009. Resource dependence theory: A review. Journal of Management, 35: 1404-1427.

53.

Hitt

M. A.

Hoskisson

R. E.

Kim

1997. International diversification: Effects on innovation and firm performance in product-diversified firms. Academy of Management Journal, 40: 767-798.

54.

Holmqvist

2004. Experiential learning processes of exploitation and exploration within and between organizations: An empirical study of product development. Organization Science, 15: 70-81.

55.

Hrebeniak

Joyce

1985. Organizational adaptation: Strategic choice and environmental determinism. Administrative Science Quarterly, 30: 336-349.

56.

Ibarra

Andrews

S. B.

1993. Power, social influence, and sense making: Effects of network centrality and proximity on employee perceptions. Administrative Science Quarterly, 38: 277-303.

57.

Jacobs

1974. Dependency and vulnerability: An exchange approach to the control of organizations. Administrative Science Quarterly, 19: 45-59.

58.

Jick

T. D.

1979. Mixing qualitative and quantitative methods: Triangulation in action. Administrative Science Quarterly, 24: 602-611.

59.

Kim

E. H.

Singal

1993. Mergers and market power: Evidence from the airline industry. American Economic Review, 83: 549-569.

60.

Kleinbaum

D. G.

Kupper

L. L.

Muller

K. E.

1988. Applied regression analysis and other multivariate methods (2nd ed.). Boston: PWS-Kent.

61.

Kogut

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

62.

Kogut

Shan

Walker

1992. The make-or-cooperate decision in the context of an industry network. In Nohria

Eccles

R. G.

(Eds.), Networks and organizations: Structure, form, and action: 348-365. Boston: Harvard Business School Press.

63.

Koka

B. R.

Prescott

J. E.

2008. Designing alliance networks: The influence of network position, environmental change, and strategy on firm performance. Strategic Management Journal, 29: 639-661.

64.

Krackhardt

1995. Review symposium: Structural holes. Administrative Science Quarterly, 40: 350-354.

65.

Lavie

2007. Alliance portfolios and firm performance: A study of value creation and appropriation in the U.S. software industry. Strategic Management Journal, 28: 1187-1212.

66.

Lavie

Rosenkopf

2006. Balancing exploration and exploitation in alliance formation. Academy of Management Journal, 49: 797-818.

67.

Lawless

M. W.

Finch

L. K.

1989. Choice and determinism: A test of Hrebiniak and Joyce’s framework on strategy-environment fit. Strategic Management Journal, 10: 351-365.

68.

Lee

G. K.

2007. The significance of network resources in the race to enter emerging product markets: The convergence of telephony communications and computer networking, 1989-2001. Strategic Management Journal, 28: 17-37.

69.

Lerner

1994. Venture capitalists and the decision to go public. Journal of Financial Economics, 35: 293-316.

70.

Lomi

Pattison

2006. Manufacturing relations: An empirical study of the organization of production across multiple networks. Organization Science, 17: 313-332.

71.

Marsden

P. V.

1982. Brokerage behavior in restricted exchange networks. In Marsden

(Ed.), Social structure and network analysis: 201-218. Beverly Hills, CA: Sage.

72.

Mayer , 2013. Integrating theories in AMJ articles. Academy of Management Journal, 56: 917-922.

73.

Mitchell

R. K.

Agle

B. R.

Wood

D. J.

1997. Toward a theory of stakeholder identification and salience: Defining the principle of who and what really counts. Academy of Management Review, 22: 853-886.

74.

Mizruchi

M. S.

Blyden

B. P.

1998. Centrality and power revisited: Actor success in group decision-making. Social Networks, 20: 353-387.

75.

Obstfeld

2005. Social networks, the tertius iungens orientation, and involvement in innovation. Administrative Science Quarterly, 50: 100-130.

76.

Okhuysen

Bonardi

2011. The challenges of building theory by combining lenses. Academy of Management Review, 36: 6-11.

77.

Oliver

1990. Determinants of interorganizational relationships: Integration and future directions. Academy of Management Review, 15: 241-265.

78.

Oliver

1997. Sustainable competitive advantage: Combining institutional and resource-based views. Strategic Management Journal, 18: 697-713.

79.

Palmer

Barber

B. M.

Zhou

Soysal

1995. The friendly and predatory acquisition of large U.S. corporations in the 1960s: The other contested terrain. American Sociological Review, 60: 469-499.

80.

Park

S. H.

Chen

Gallagher

2002. Firm resources as moderators of the relationship between market growth and strategic alliances in semiconductor start-ups. Academy of Management Journal, 45: 527-545.

81.

Penner-Hahn

Shaver

J. M.

2005. Does international research and development increase patent output? An analysis of Japanese pharmaceutical firms. Strategic Management Journal, 26: 121-140.

82.

Perry-Smith

J. E.

2006. Social yet creative: The role of social relationships in facilitating individual creativity. Academy of Management Journal, 49: 85-101.

83.

Pfeffer

1972a. Interorganizational influence and managerial attitudes. Academy of Management Journal, 15: 317-330.

84.

Pfeffer

1972b. Merger as a response to organizational interdependence. Administrative Science Quarterly, 17: 382-394.

85.

Pfeffer

1976. Beyond management and the worker: The institutional function of management. Academy of Management Review, 1: 36-46.

86.

Pfeffer

1987. A resource dependence perspective on interorganizational relations. In Mizruchi

M. S.

Schwartz

(Eds.), Intercorporate relations: The structural analysis of business: 25-55. Cambridge, England: Cambridge University Press.

87.

Pfeffer

2003. Introduction to the classic edition. In Pfeffer

Salancik

G. R.

(Eds.), The external control of organizations: A resource dependence perspective: xi-xx. Stanford, CA: Stanford University Press.

88.

Pfeffer

Nowak

1976. Joint ventures and interorganizational interdependence. Administrative Science Quarterly, 21: 398-418.

89.

Pfeffer

Salancik

G. R.

1978. The external control of organizations: A resource dependence perspective. New York: Harper.

90.

Podolny

J. M.

2001. Networks as the pipes and prisms of the market. American Journal of Sociology, 107: 33-60.

91.

Podolny

J. M.

Stuart

1995. A role-based ecology of technological change. American Journal of Sociology, 100: 1224-1260.

92.

Podolny

Stuart

Hannan

1996. Networks, knowledge, and niches: Competition in the worldwide semiconductor industry, 1984-1991. American Journal of Sociology, 102: 659-689.

93.

Powell

W. W.

1990. Neither market nor hierarchy: Network forms of organization. In Staw

B. M.

Cummings

L. L.

(Eds.), Research in organizational behavior, vol. 12: 295-336. Greenwich, CT: JAI Press.

94.

Powell

W. W.

Koput

K. W.

Smith-Doerr

1996. Interorganizational collaboration and the locus of innovation: Networks of learning in biotechnology. Administrative Science Quarterly, 41: 116-145.

95.

Provan

K. G.

Beyer

J. M.

Kruytbosch

1980. Environmental linkages and power in resource-dependence relations between organizations. Administrative Science Quarterly, 25: 200-225.

96.

Rosenkopf

Padula

2008. Investigating the microstructure of network evolution: Alliance formation in the mobile communications industry. Organization Science, 19: 669-687.

97.

Rosenkopf

Schilling

2007. Comparing alliance network structure across industries: Observations and explanations. Strategic Entrepreneurship Journal, 1: 191-209.

98.

Rothaermel

F. T.

Boeker

2008. Old technology meets new technology: Complementarities, similarities, and alliance formation. Strategic Management Journal, 29: 47-77.

99.

Rothaermel

F. T.

Deeds

D. L.

2004. Exploration and exploitation alliances in biotechnology: A system of new product development. Strategic Management Journal, 25: 201-222.

100.

Rowley

T. J.

1997. Moving beyond dyadic ties: A network theory of stakeholder influences. Academy of Management Review, 22: 887-910.

101.

Rowley

Behrens

Krackhardt

2000. Redundant governance structures: An analysis of structural and relational embeddedness in the steel and semiconductor industries. Strategic Management Journal, 21: 369-386.

102.

Schilling

M. A.

2009. Understanding the alliance data. Strategic Management Journal, 30: 233-260.

103.

Shipilov

A. V.

2009. Firm scope experience, historic multimarket contact with partners, centrality, and the relationship between structural holes and performance. Organization Science, 20: 85-106.

104.

Silverman

B. S.

1999. Technological resources and the direction of corporate diversification: Toward an integration of the resource-based view and transaction cost economics. Management Science, 45: 1109-1124.

105.

Soda

Usai

Zaheer

2004. Network memory: The influence of past and current networks on performance. Academy of Management Journal, 47: 893-906.

106.

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: 668-698.

107.

Suddaby

Hardy

Huy

2011. Where are the new theories of organization? Academy of Management Review, 36: 236-246.

108.

Thompson

1967. Organizations in action. New York: McGraw-Hill.

109.

Uzzi

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

110.

Vuong

1989. Likelihood ratio tests for model selection and non-nested hypotheses. Econometrica, 57: 307-333.

111.

Walker

Kogut

Shan

1997. Social capital, structural holes, and the formation of an industry network. Organization Science, 8: 109-125.

112.

Walters

B. A.

Kroll

Wright

2010. The impact of TMT board member control and environment on post-IPO performance. Academy of Management Journal, 53: 572-595.

113.

Wang

Zajac

2007. Alliance or acquisition? A dyadic perspective on interfirm resource combinations. Strategic Management Journal, 28: 1291-1317.

114.

Yang

Lin

Z. J.

Peng

M. W.

2011. Behind acquisitions of alliance partners: Exploratory learning and network embeddedness. Academy of Management Journal, 54: 1069-1080.

115.

Zahra

S. A.

Newey

L. R.

2009. Maximizing the impact of organization science: Theory-building at the intersection of disciplines and/or fields. Journal of Management Studies, 46: 1059-1075.