Knowledge sourcing by multidivisional firms

Abstract

While the real options perspective recognizes that firms are complex, non- monolithic entities, its application to the study of R&D investments under uncertainty has largely been at the firm level. This study examines how the internal distribution of capabilities across divisions affects knowledge-sourcing strategies of multidivisional firms. Our formal model shows that, under low uncertainty, firms commit resources to strong divisions to maximize returns. Under high uncertainty, however, they invest more in weak divisions for fear of omitting a potential winner. These choices are interdependent in that higher returns from the revealed winner when sourcing late-stage knowledge (low uncertainty) gives the firm incentive to cast a wider net in the early stage (high uncertainty). Consequently, at the firm level, more unequal distribution of capabilities across divisions leads to more early- stage sourcing but less late-stage sourcing. We find empirical support for these predictions using data on knowledge-sourcing activities of large pharmaceutical companies. This study extends the real options literature by arguing that capabilities of a firm’s constituents influence its strategy over and above aggregate firm-level capabilities. Thus, attention to heterogeneity across divisions within firms helps unpack differences in knowledge sourcing across firms.