Trust and Betrayals: Reputational Payoffs and Behaviors without Commitment

Harry Pei
DOI: https://doi.org/10.48550/arXiv.2006.08071
2020-06-15
Abstract:I study a repeated game in which a patient player (e.g., a seller) wants to win the trust of some myopic opponents (e.g., buyers) but can strictly benefit from betraying them. Her benefit from betrayal is strictly positive and is her persistent private information. I characterize every type of patient player's highest equilibrium payoff. Her persistent private information affects this payoff only through the lowest benefit in the support of her opponents' prior belief. I also show that in every equilibrium which is optimal for the patient player, her on-path behavior is nonstationary, and her long-run action frequencies are pinned down for all except two types. Conceptually, my payoff-type approach incorporates a realistic concern that no type of reputation-building player is immune to reneging temptations. Compared to commitment-type models, the incentive constraints for all types of patient player lead to a sharp characterization of her highest attainable payoff and novel predictions on her behaviors.
Theoretical Economics
What problem does this paper attempt to address?