The effects of knowledge network characteristics on R&D alliance formation

Firms' knowledge characteristics shape their research and development (R&D alliance activities). When denoting the characteristics, the alliance literature has exclusively focused on firms' internal knowledge characteristics, leaving structural properties of their knowledge relatively unexplored. Therefore, this study aims to investigate how firms' knowledge network characteristics are associated with the propensity to ally. By using 235,024 unique alliance dyads between U.S. semiconductor firms and their potential partners, we examine the effects of two knowledge network characteristics – degree centrality and structural holes – on alliance formation. We further examine the joint effects of the knowledge network characteristics. Our findings reveal the curvilinear relationship between degree centrality and the likelihood of alliance formation. We also find support for the positive joint effects on the likelihood of alliance formation. These findings provide implications for the varying propensity to ally associated with the knowledge network characteristics, i.e. structural positions of a firm's knowledge elements in the sector-level knowledge network.