Folk theorems for Bayesian (public good) games
Résumé
Two characterization results are behind the intimate relationshipbetween "repetition" and "cooperation" under complete information:the standard Folk theorem for infinitely repeated games and the "commitment Folk theorem" for one-shot games. We propose extensionsof the previous characterization results in Bayesian games, with independent private values, which satisfy a further property, "uniformpunishment strategies". Public good games fall in this class. We showthat the Nash equilibria of the Bayesian infinitely repeated game arepayoff equivalent to separating (i.e., completely revealing) equilibriaand can be achieved as interim cooperative solutions of the Bayesiangame. We also show that the reverse of the latter result is not true:unlike the set of interim cooperative solutions of the Bayesian game,the set of Nash equilibrium payoffs of the infinitely repeated game canbe empty.
Origine : Fichiers produits par l'(les) auteur(s)
Loading...