A game-theoretic model for the role of reputation feedback systems in peer-to-peer commerce

We propose a dynamic game-theoretic model to study how a reputation feedback system affects a peer-to-peer (P2P) trading platform. Continuums of buyers and sellers with bounded rationalities are paired to trade periodically. Sellers have private knowledge on their own types that encode tendencies to cheat. Their reputation scores are, on the other hand, visible to all and periodically updated by their trading partners. A buyer can use a seller's score to assess the latter's characters and decide whether to trade. Using a continuity-based fixed point theorem, we first establish the existence of equilibria that convey information on traders' behaviors as well as the type-score composition on the market. We then look into a special case which, for model tractability, includes two types of sellers with two score levels. Besides the proportion of prone-to-cheating sellers, it indicates that how much sellers value their future payoffs also plays a major role in determining the prevalence of online frauds. Therefore, reputation feedback systems are especially beneficial under high trading volumes, which can in turn, for instance, be facilitated by attracting more buyers to the online marketplace.