The Optimal Design of Research Contests

University Of Wisconsin-Madison, Madison WI

Investigators

Abstract

The procurement of new products such as weapons systems, transportation systems, or new car models often requires substantial innovative effort by suppliers. Procuring such efforts through bilateral contracts may be difficult, however, since the efforts may be "unverifiable" and "cooperative." That is, the level of effort cannot be verified by a third party, and a substantial fraction of the value generated may accrue to the buyer. Also, bankruptcy concerns and legal restrictions limit the ability of the procurer to charge entry fees (i.e., "limited liability" ). Procurers have often relied on contests to procure innovative effort in such an environment. In particular, research contests have been an important part of R&D procurement competitions sponsored by various government agencies (notably the Department of Defense) as well as private firms. The proposed research will study the optimal method of procurement when unverifiability, cooperativeness and limited liability are important. We will first make the case for using contests in place of bilateral contracts. Second, we will study how the procurer should select the set of firms to participate in the contest. Finally, and most importantly, we will characterize the optimal contest design. In doing so, we allow for a vast array of prize structures, which includes auctions (contestants bid their prizes) and tournaments (the prize is fixed) as special cases. Some aspects of these issues have been studied previously. For instance, the auctions literature has yielded substantial insights into the en post performance of auctions (i.e., how the good or Contract is allocated, and how the surplus is divided), showing that auctions may be an optimal mechanism. The search for an optimal mechanism has not been extended to incorporate en ante investment incentives, however. Recent work on research contests and entry into auctions has incorporated an ex ante dimension, but it still takes as given the form of contest (auction or tournament). The proposed research will be the first to consider a broad class of contests, to endogenize the prize structure, and to identify the optimal contest when ex ante and ex post performance considerations are important. The completed research will be presented in two parts. In the first part, we will provide closed-form expressions for equilibrium investment and bidding strategies, even with asymmetric firms (i.e., firms with different investment technologies). The second part will identify the optimal contest design, using a novel duality argument. (The standard mechanism design approach does not work here.) A first-price auction with just two firms and no reserve price is optimal if the firms are symmetric. If the firms have different technologies, an auction with two firms appears to be optimal, but with the more efficient firm being handicapped. This research will make a signficant contribution to the literature on contests in general, and auctions in particular. It will also contribute to the incomplete contracts literature by investigating the special role of contests in motivating cooperative investments. Finally, it will prove useful to contest designers. The optimality of first-price auctions with just two firms is particularly important, given the ease of administering auctions.