No abstract

No abstract

No abstract

Research sharing is an important objective of many research joint ventures. When partners share R&D but do not maximize joint profits, large consortia are more profitable than small ones, and joint ventures prefer dispersed rivals. For much of the spillover space, a coalition formation game that permits limited membership predicts that at most, three joint ventures form. Research‐sharing joint ventures improve welfare when spillovers are low, and banning research sharing joint ventures is beneficial for high spillovers. With imperfect research sharing and low spillovers, allowing only research sharing is the best industry‐wide joint venture alternative for consumer surplus.

In this paper we extend the existing literature on research and development (R&D) investments and research joint ventures (RJVs) in two important ways. First, we analyze and compare the case where firms collude in the product market to the benchmark case of competition in the output market. Second, we allow firms to form coalitions endogenously as a separate stage in the game. We develop profit functions that depend on the partition of firms into joint ventures and the nature of product competition between venture partners. Our results illustrate the restrictive nature of some assumptions made in the literature. Typically multiple RJVs of different sizes form in equilibrium. In general, RJVs should not be promoted if they entail product market collusion. Given the information available to policy-makers, it is unlikely that an R&D policy more refined than analyzing and allowing RJVs on a case-by-case basis is feasible.