Repeated interaction and the evolution of preferences for reciprocity

This paper models self-enforcing agreements and the evolution of preferences for reciprocity when all agents are rational, Bayesian optimisers. In this 'indirect' evolutionary model, player types are defined not by their strategies, but by their preferences. The paper studies the evolution of a community consisting of 'opportunists', who maximise material payoffs, and 'reciprocators', who prefer joint co-operation to exploiting their opponents. Players are randomly matched to play a finitely repeated Prisoner's Dilemma game. In the unique evolutionary equilibrium, the reciprocators co-operate throughout their careers, and the opportunists co-operate up to, but not including, the last stage of their careers.