Optimal Design of Research Contests

Many new products (e.g., weapons systems) require substantial innovative effort by suppliers. Procurement of effort through bilateral contracting is often problematic, however, because the effort may be unverifiable and cooperative. That is, a third party cannot verify the level of effort, and the procurer receives a substantial fraction of the value created by the effort. In addition, bankruptcy concerns and legal restrictions limit the ability of the procurer to charge entry fees (we refer to this as limited liability). Procurers often rely on contests in such an environment. For example, the U.S. Federal Communications Commission sponsored a competition to develop the technology for high-definition television. This paper finds the optimal procurement mechanism when unverifiability, cooperativeness and limited liability are present. We accomplish this through the use of a novel duality argument. (The standard mechanism design methodology cannot be used because two mechanisms that are ex post efficient may induce different investment decisions, hence different distributions of types.) Holding an auction with two firms and no reserve price is optimal for the procurer, when the firms have the same technology. When the firms have different technologies, an auction involving the two most efficient firms is optimal, with the more efficient firm being handicapped. This paper contributes to mechanism design theory by incorporating ex ante investment incentives into the mechanism design problem. That is, it finds the optimal mechanism with endogenous types. Recent work on research contests and entry into auctions has incorporated an ex ante dimension, but that work takes the form of contest (auction or tournament) as given.