Determinants of firm boundaries: Empirical analysis of the Japanese auto industry from 1984 to 2002

Abstract

Using the framework of contract economics, we assess the determinants of the choice of vertical integration, relational contracting (keiretsu sourcing) or market sourcing by seven Japanese automobile manufacturers (OEMs) with respect to 54 components. Our major findings are as follows. First, the specificity of a component significantly promotes keiretsu over market sourcing. This effect declines with the testability of the component, consistent with transaction cost economics. Second, the interdependency of a component makes vertical integration significantly preferred to keiretsu sourcing, while the specificity of the component does not. These results suggest that the hold-up risk due to specific investment can often be effectively controlled by relational contracting based on keiretsu sourcing, while accommodating non-contractible design changes may often require vertical integration. Third, high testability of a component significantly promotes the choice of keiretsu over market sourcing. One interpretation of this result is that a keiretsu supplier may have both the incentive to gather information for quality improvement in an environment of high testability, and the incentive to share the information with the OEM for its implementation. J. Japanese Int. Economies 22 (2) (2008) 187–206.