Title: "Divide to Unite: How Does Competition in a Firm's Alliance Portfolio Affect the Structure and Outcome of its Knowledge Production?"
Abstract: While organizations can capture substantial knowledge complementarities across the portfolios of partnerships they engage in, research suggests that a firm can extract greater benefits when its alliance partners are in direct competition with each other. However, we argue that when a firm maintains simultaneous collaborative ties to rivaling partners, it is unlikely to be able to orchestrate a free flow of knowledge between its own inventors, which is a prerequisite for cross-pollinating and recombining ideas it obtains from these partners. Because permitting such unrestricted exchanges poses leakage concerns for partners whose rivals also ally with the focal firm, we argue that the focal firm will be pressured to impose credible knowledge safeguards that will be reflected in the increased disconnectedness of its inventor collaboration network. Moreover, the extent to which a firm can resist such pressures will depend on the relative technological bargaining power it holds over its partners. Importantly, these safeguards are not costless, and the disconnectedness induced will affect the focal firm's innovation performance by throttling its ability to produce higher-quality inventions through knowledge recombination and undermining its ability to appropriate value by spawning cumulative inventions. We demonstrate consistent support for these arguments on a longitudinal sample of pharmaceutical firms using a robust research design that mitigates dual endogeneity concerns in firms' partner choices and their inventor collaboration patterns. Our findings contribute to scholarship at the intersection of inter-firm alliances, intra-organizational networks, and innovation while offering practical insights to managers.
