Partner Selections in the Formation of Multi-Partner Alliances:
A Group Approach
Xiaoli Yin
Department of Management
BaruchCollege
The CityUniversity of New York
New York, NY10010-5585
Tel: 646- 312- 3679
Email:
Jianfeng Wu
BusinessSchool
University of International Business and Economics
Beijing, China 100029
Tel: +86 10 6449-4303
Email:
The authors are indebted toArnie Cooper, Joe Galaskiewicz, Steve Green,Tim Pollock,Mark Shanley, Wei Shen, and Lihua Wang, Kathy Williams for comments on earlier versions of this paper.
Partner Selections in the Formation of Multi-Partner Alliances:
A Group Approach
Abstract
This study examines focal firms’ choices of one group of partners over others in the formation of multi-partner alliances. We consider the entire multi-partner alliance as a group and explore factors influencing a focal firm's decision to form the group. With a sample of industry-sponsored e-marketplaces (ISMs), we find that while social ties and relational heterogeneity significantly influence a focal firm’s choice of partners, their effects are also moderated by the size of the multi-partner alliance.
Prior research suggests that organizations are subject to high uncertainty when forming exchange relations with others (e.g., Gulati & Singh, 1998; Pfeffer & Salancik, 1978; Podolny, 1994; Simon, 1957; Thompson, 1967; Williamson, 1975, 1985). One way to avoid market uncertainty is to adopt a social orientation by adhering to a principal of exclusivity in selecting exchange partners (Podolny, 1994). For instance, investment banks tend to select partners with whom they have transacted in the past and partners of similar status to reduce uncertainty (Podolny, 1994). This principal of exclusivity is commonly applied in strategic alliances. Based on social network theory, prior research on dyadic alliances suggests that firms manage to reduce uncertainty by partnering with their prior social ties (Gulati, 1998; Gulati & Gargiulo, 1999). Multi-partner alliances or alliance constellations formed by three or more partners, however, face new challenges in responding to uncertainty during the partner selection process. As three or more partners collaborate, there is no direct reciprocity among them and social exchanges become generalized. Subsequently, uncertainty and risks of free-riding are higher in multi-partner alliances than in dyadic alliances (Das & Teng, 2002). More important, uncertainty increases as the number of participants of a constellation (i.e., group size) increases (Kanter, 1977; O’Reilly, Caldwell & Barnett, 1989). Given the high level of uncertainty and the complex nature of multi-party exchanges, it is no longer sufficient to merely apply social network theory to examine how firms choose a group of partners in multi-partner alliance formation. Multi-partner alliances should be examined as a distinctive type of alliance both for theoretical development and empirical testing.
We focus explicitly on multi-partner alliances for two reasons. First, multi-partner alliances provide a rare opportunity to test the relationship between uncertainty and exclusivity. This study investigates a unique type of uncertainty that stems from variations in alliance size. We do not focus on external environmental or market uncertainty. Instead, uncertainty in this study refers to the difficulty and complexity of collaborating with a group of partners. As the number of alliance participants increases, uncertainty increases, and it becomes more difficult for focal firms to manage the potential conflicts and free-riding behaviors of their alliance partners. Specifically, we examine whether focal firms become more exclusive in their partner selection process as the size of alliance constellations increases. Second, multi-partner alliances have increasingly gained attention in the past decade (e.g., Das & Teng, 2002; Doz & Hamel, 1998; Gomes-Casseres, 1996; Makino & Beamish, 1999; Sakakibara, 1997; Zeng & Chen, 2003). Among strategic alliances, it is not uncommon to observe that three or more partners collaborate with each other to achieve the same goal. For example, Makino and Beamish (1999) found that in their sample of 737 joint ventures, 55 percent had more than 2 partners. Popular types of alliance constellations include R&D consortia, joint bidding, product bundling, horizontal keiretsu, and industry-sponsored e-marketplaces (ISMs) (Das & Teng, 2002). In spite of the growing consensus about its importance, theoretical and empirical investigation of multi-partner alliances is rather limited. The few existing studies assume that multi-partner alliances already exist, and none of them address the issue of with whom to form multiple-partner alliances. The main objective of this study is to explore: When a focal firm decides to form a new multi-partner alliance, why will it choose one group of partners over others? This question is quite significant as the behavior of alliance partners is a key factor in the success of an alliance (Gulati, 1995). Firms face high uncertainty and high risks of opportunistic behaviors in inter-organizational collaborations such as those seen in strategic alliances and R&D consortia (Browning, Janice, & Shetler, 1995; Gulati & Gargiulo, 1999). In fact, the normal life span for most alliances is usually no more than five years (Kogut, 1988). The considerable uncertainty associated with entering cooperative ties makes it imperative to decide with whom to build those ties.
To answer this question, social network theory should be coupled with group dynamics theory to address the unique challenges and uncertainties in multi-partner alliances. In other words, social network theory only provides a partial explanation for partner selections in constellation formation. As alliance constellations involve a group of partners, group dynamics theory is also relevant. One of the key contributions of this study is to introduce group dynamics theory to examine the direct and moderating effects of group factors on partner selections in multi-partner alliances.
This study also contributes to alliance research by adopting a group approach to examine a focal firm’s collaboration with a group of partners in forming a multi-partner alliance. Most previous alliance research focuses on the dyadic level of analysis. We adopt a group approach to uncover important group attributes of multi-partner alliances that are neglected or masked by a dyadic approach. Instead of examining each pair of firms in a multi-partner alliance, this study takes a holistic view considering the entire multi-partner alliance as a group and explores factors influencing a focal firm's decision to form the group. By adopting a group approach to study partner selections in multi-partner alliances, we advance our understanding of these important yet underdeveloped topics.
This study relies on social network theory, group dynamics theory, and a group approach to investigate what determines a focal firm’s choice of partners in multi-partner alliances. Specifically, we examine to what extent prior social ties, relational heterogeneity, and interactions with group size contribute to a focal firm’ propensity to cooperate with one group of partners over others in forming multi-partner alliances. The study suggests that focal firms are more likely to choose a group of partners with whom they share more prior ties and who are more similar to themselves. The study also suggests a contingency perspective, that the size of multi-partner alliances (i.e., anticipated group size) will moderate the effect of prior social ties and relational heterogeneity on focal firms’ choice of partners. As the alliances have more partners, indirect ties have diminishing effect on partner selections, and the effect of relational heterogeneity becomes more pronounced. We study a unique type of collaborative venture in the new information economy: industry-sponsored e-marketplaces (ISMs). Specifically, we examine 19 ISMs (with a total of 189 founding firms) established in the year 2000 in four different industries (i.e., aerospace, chemical, energy, and metal). Empirical results show considerable support for our hypotheses, that both social networks and group dynamics factors have main and interacting effects on partner selections in multi-partner alliances.
Theoretical Framework and Hypotheses
Multi-partner Alliances and Uncertainty
Most previous research on strategic alliances has focused on alliances between two firms. Research attention on multi-partner alliances has been sporadic, because most researchers have not regarded multiple-partner alliances as distinctive from dyadic alliances (Das & Teng, 2002). As Das and Teng (2002) point out, however, multiple-partner alliances are a particular, yet complex form of strategic alliances. Recently some researchers begin to look into collaborative organizations involving three or more members and examine issues of collective strategies among members, membership continuity, motives for cooperation, social control and management challenges, and cooperation of alliance constellations (Barnett, Mischke, & Ocasio, 2000; Das & Teng, 2002; Olk & Young, 1997; Sakakibara, 1997; Zeng & Chen, 2003).
While previous research contributes to our understanding of multi-partner alliances, it does not address the issue of how a focal firm chooses one group of partners over others in forming a multiple-partner alliance. There are some basic questions remained to be answered: How would a focal firm choose one group of partners over others in the formation of multi-partner alliances?If we treat an alliance constellation as a group, how will the group attributes affect an individual member’s decision to jointly form that constellation? The distinctiveness and complexity of partner selections in multi-partner alliances should be explored further in the research of strategic alliances.
For instance, one important distinction of multi-partner alliances is that uncertainty is higher in alliance constellations than in dyadic alliances. When there are two partners in an alliance, social exchanges are restricted and reciprocal as there are direct exchanges of favors between the two partners. As there are three and more partners involved in an alliance, social exchanges become generalized as there is no direct reciprocity among multiple partners (Das & Teng, 2002). The key feature of generalized exchanges is the lack of one-to-one correspondence between the giver and the receiver (Das & Teng, 2002). There are major differences between restricted and generalized social exchanges (Ekeh, 1974). First, individual members have more incentives to free-ride in generalized exchanges. In restricted social exchanges, two parties directly exchange favors with each other and thus it is relatively easy to observe and remedy free-riding. In generalized exchanges, a group-based exchange relationship makes it harder to detect and thus prevent free-riding. Second, the need for trust is especially high in generalized social exchanges as exchanges are carried out by multiple members who do not directly reciprocate with each other. Trust among exchange parties not only helps reduce uncertainty and anxiety associated with potential free-riding, but also encourages cooperation and enhances the overall benefits of the group. Dyadic alliances are characterized by restricted exchanges. Alliance constellations are characterized by generalized social exchanges and thus experience a higher risk of free-riding and higher need for trust (Das & Teng, 2002). Given the higher uncertainty in multi-partner alliances as compared to dyadic alliances, it would be important to examine if similar predictors of alliances between two firms (i.e., prior ties and homophily) would also lead to a focal firm’s choice of a group of partners in multiple-partner alliances.
Meanwhile, the other key distinction of multi-partner alliances is that uncertainty increases as the number of alliance participants increases. From a group perspective, multi-partner alliances vary in group size or number of alliance participants. Group research found that group size has significant implications to firm behaviors. As group size increases, face-to-face interactions become more difficult, and group cohesiveness and performance start to deteriorate (House & Miner, 1969). Further, increase in group size also makes group members more prone to free-ride (Albanese & Van Fleet, 1985). In the case of multi-partner alliances, as there are more firms participating in an alliance constellation, uncertainty increases and solidarity and trust become increasingly important in a focal firm’s partner selection process. The unique group dynamics of multi-partner alliances provide an appropriate setting to study whether firms become more exclusive in partner selections as uncertainty increases.
Put together, even though multiple-partner alliances share many similarities with dyadic alliances, “greater numbers of participants also complicate alliance design and governance” (Doz & Hamel, 1998: 224). Compared to dyadic alliances, multiple-partner alliances face higher uncertainty and have higher needs for trust. Meanwhile, as the number of participants increases, uncertainty increases in alliance constellations. Multi-partner alliances should be treated as a distinctive type of alliances for theoretical development and empirical testing. To address the unique challenges and uncertainties in multi-partner alliances, social network theory should be coupled with group dynamics theory to study partner selections. Based on social network perspective and group dynamics theory, the study tests whether the most common determinants of partner selections in dyadic alliances (i.e., prior ties and homophily) would replicate themselves in the context of alliance constellations. More important, the study investigates how a focal firm’s choice of a group of partners might be moderated by the anticipated size of the alliance.
This study adopts a group approach to examine a focal firm’s collaboration with a group of partners in multi-partner alliance formation. When three or more partners collaborate, a focal firm will not only consider its relationship with individual partners, but also will evaluate the characteristics of the other partners as a group in constructing an alliance. In this study, we treat each multi-partner alliance as a group and examine how group level characteristics such as social network ties and group heterogeneity affect a focal firm’s decision to jointly found that group. We know of only one prior study that assesses partner selections directly from a group approach.[1] There may be significant heterogeneity in the overall attributes of each group of partners, and this important source of variation is neglected in research that simply examines each pair of firms within multi-partner alliances. Based on social network theory, group dynamics literature, and a group approach, we hope to uncover important distinctions of partner selections in multi-partner alliances.
We examine our research questions in the setting of ISMs. Industry analysts define ISM as a separate legal entity formed by three or more established companies to create a B2B net market (Temkin, Kafka, King, Freiden, & Hurd, 2001). Building an ISM is extremely expensive and may cost more than 100 million US dollars. Therefore, most ISMs are endorsed by large industry incumbents. There are five major functions targeted by an ISM: (1) support buy and sell transactions; (2) supply chain optimization; (3) developing industry-wide e-business standards; (4) strategic sourcing; and (5) collaborative product development (Temkin et al., 2001). For example, in the aerospace industry, two of the major ISMs are Exostar and Aeroxchange. Exostar was established on March 28, 2000, by four major leaders in the industry: BAE Systems, Boeing Company, Lockheed Martin Corporation, and Raytheon. Each founding member contributes equal equity for this new entity. Its main objective is to bring both industry customers and suppliers together to streamline supply chain management. Aeroxchange was established in October 2000 by 13 member airlines from Asia-Pacific region, North America, and Europe. This global e-commerce marketplace is dedicated to transform the complex aviation industry supply chain by integrating buyers and sellers, reducing transaction costs, and facilitating information transparency.
Compared to traditional R&D consortia or collective organizations, the prospect of ISMs is quite uncertain. It remains unclear whether this new collaborative platform is a viable organizational form. Most ISMs emerged within a few months in 2000 and thereafter experienced a lot of restructuring and consolidations. Even some of the most successful ISMs, such as Metalsite and Scrapsite which were once rated as best B2B exchange sites by Forbes, ceased operation in June 2001, due to the lack of funding and were later acquired by Metal Supply Alliances. In general, most ISMs haven’t yet generated profits. Cash insufficiency is still a big problem plaguing managers of ISMs. Although these ISMs promise many solutions for supply chain management, new product development, and distribution, there is still no solid evidence indicating that these benefits have been realized. We suggest that when there is high uncertainty about the viability of collective ventures, industry players become more exclusive in selecting partners to jointly found ISMs. Therefore, studying the unique challenges and uncertainties faced by individual firms in ISMs might present a rare opportunity to examine the relationship between uncertainty and exclusivity in inter-organizational cooperation.
Social Network Theory
Previous studies on alliance formation have applied various theoretical perspectives including social network theory (Gulati, 1998). Social network theory not only helps us explain whether firms would choose to form ties with other firms, it also predicts with whom to form such alliances (Gulati &Gargiulo, 1999). Among all interdependent organizations, some firms are more familiar with each other due to previous interactions or affiliations. To avoid uncertainty of potential partnerships, firms tend to choose partners they find trustworthy. Social networks can provide differential information advantages through relational embeddedness and structural embeddedness (Granovetter, 1992). Relational embeddedness suggests that direct cohesive ties between organizations help firms gain common information of each other. Cohesive ties also help create trust and solidarities which tend to lead to subsequent cooperation between organizations (Gulati & Gargiulo, 1999).Structural embeddedness goes beyond direct ties and instead focuses on the structure of relations around organizations (Granovetter, 1992). The next section explores the effects of direct ties and indirect ties on a focal firm’s choice of one group of partners over others in founding an ISM. We take a group approach to study social network effects on partner selections and consider the entire ISM as a group. Instead of examining prior ties between each pair of firms, we study how the total number of prior ties between a focal firm and a group of partners affects the focal firm’s decision to form an ISM with that group.