Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Contracting for research is often infeasible because research inputs are unobservable and research outcomes cannot be verified by a court. Sponsoring a research tournament can resolve these problems. A model is presented in which contestants compete to find the innovation of highest value to the tournament sponsor. The winner receives a prespecified prize. The tournament game has a unique subgame-perfect equilibrium. Free entry is not optimal because equilibrium effort by each researcher decreases in the number of contestants. An optimally designed research tournament balances the probability of overshooting the first-best quality level against the probability of falling short. Copyright 1995 by American Economic Association.