Interfirm innovation under uncertainty: Empirical evidence for strategic knowledge partitioning

This paper analyzes how uncertainty and life-cycle effects condition the knowledge boundary between assemblers and suppliers in interfirm product development. Patents associated with automotive emission control technologies for both assemblers and suppliers are categorized as architectural or component innovations, and technology-forcing regulations imposed by the government on the auto industry from 1970 to 1998 are used to define periods of high and low uncertainty. Results confirm that suppliers dominate component innovation whereas assemblers lead on architectural innovation. More importantly, when facing uncertainty firms adjust their knowledge boundary by increasing the knowledge overlap with their supply-chain collaborators. Suppliers clearly expand their knowledge base relatively more into architectural knowledge during such periods. But assemblers' greater emphasis on component innovation in periods of greater uncertainty is only true as a relative deviation from an overall trend toward increasing component innovation over time. This trend results from an observed life-cycle effect, whereby architectural innovation dominates before the emergence of a dominant design, with component innovation taking the lead afterward. Thus, for assemblers life-cycle effects may dominate over task uncertainty in determining relative effort in component versus architectural innovation. This work extends research on strategic interfirm knowledge partitioning as well as on the information-processing view of product development. First, it provides a large-scale empirical justification for the claim that firms' knowledge boundaries need to extend beyond their task boundaries. Further, it implies that overlaps in knowledge domains between an assembler and suppliers are particularly important for projects involving new technologies. Second, it offers a dynamic view of knowledge partitioning, showing how architectural knowledge prevails in the early phase of the product life cycle whereas component knowledge dominates the later stages. Yet the importance of life-cycle effects versus task uncertainty in conditioning knowledge boundaries is different for assemblers and suppliers, with the former dominating for assemblers and the latter more influential for suppliers. Finally, it supports the idea that architectural and component knowledge are critical elements in the alignment of cognitive frameworks between assemblers and suppliers and thus are key for information-exchange effectiveness and resolution of task uncertainties in interfirm innovation.