On disclosure of participation in innovation contests: a dominance result

International audienceThis paper examines the effects of disclosing the actual number of participants in innovation contests with endogenous stochastic entry. We model innovation contests as a two-bidder allpay auction with complete information, but in which each bidder has to incur a private cost to participate. The contest organizer observes solvers' participation decisions ex post and can commit ex ante to either fully disclosing or concealing the number of participating solvers. We characterize the equilibrium behavior of the solvers and compare the performances of the disclosure policies by four criteria. We find that full concealment dominates full disclosure in terms of expected total bid and expected winner's bid. Full concealment is dominated by full disclosure in terms of prize allocation efficiency and solvers' welfare. These findings are in sharp contrast to those under exogenous entry