| A broadly supported general finding from collaboration research is that organizations involved in inter-firm collaboration usually achieve above average performance and survival, although with substantial variation in outcomes (Burt, 1983; Hagedoorn & Schakeraad, 1994; Mitchel & Singh, 1996; Owen-Smith & Powell, 2003; Singh & Mitchell, 2005; Stuart & Podolny, 1999; Uzzi, 1997). Explanations for those large variations in outcomes (e.g., structural and capability based approaches) offer conflicting predictions and findings as to their effect on alliance performance (Ahuja, 2000; Bae & Gargiulo, 2004; Burt, 1992; Chung, Singh, & Lee, 2000; Coleman, 1988, 1990; Gulati & Singh, 1998; Hargadon & Sutton, 1997; Stuart, 2000; Walker, Kogut & Shan, 1997). I identify two distinct theoretical dimensions of strategic alliances: collaboration (i.e., value creation) and competition (i.e., value appropriation) (Burt, 1991; Hamel, Doz, & Prahalad, 1989; Inkpen, 2001; Lax & Sebenius, 1986; Teece, 1986). In the literature these dimensions are combined in the construct of performance and yet often have opposite effects on an organization's ability to derive value from an alliance. By formulating and testing models for these two dimensions of performance I provide an explanation to resolve the conflicting guidance of both the structural and the capability based approaches about alliances effect on performance. For example, I theoretically develop and test the idea that network closure may result in value creation within alliances while structural holes may facilitate value appropriation by individual partners.; Similarly, I study how a firms partners' resources may predict value creation but differential resource endowments between partners may result in dissimilar value appropriation potential by individual partners.; I test these models on a large longitudinal sample of strategic alliances with sequential cross-nested multilevel techniques. This analytical approach addresses methodological difficulties implicit in the study of networks data (e.g., dependencies among observations, cross-level interactions) that are not addressed in previous strategic alliances research and is capable of dealing with the interdependencies between firms in the sample. My reformulation of theory and methods to study alliance performance contributes to realizing the potential of resource based and structural approaches as powerful explanations and sources of guidance for firms' strategic alliance choices. |
