Reputation in Contests
研究了有限数量、不同技能和声誉的异质性参与者在多阶段平行竞赛中的行为，发现声誉可能导致不稳定环境，并给出了解决不稳定的充分条件，通过算法逼近子博弈完美均衡。
Abstract We investigate two-stage parallel contests with a finite number of heterogeneous agents with various skills (cost of efforts) and reputations, as well as a finite number of heterogeneous contests. Each agent’s value of winning is a combination of a nominal value and the reputation of the competing agents. We demonstrate that reputation in contests may lead to an unstable environment. We present a sufficient condition for resolving this instability, which results in the existence of a subgame perfect equilibrium with pure strategies which cannot be explicitly calculated. However, we arrive at this equilibrium by introducing algorithms that describe dynamic processes in which the agents’ strategies approach the equilibrium point. We demonstrate that our equilibrium is far more likely to hold in an environment where the agents’ skills and reputations are correlated.