COOPERATION IN THE PRISONERS-DILEMMA WITH ANONYMOUS RANDOM MATCHING

The paper considers the repeated prisoner's dilemma in a large-populat ion random-matching setting where players are unable to recognize thei r opponents. Despite the informational restrictions cooperation is sti ll a sequential equilibrium supported by ''contagious'' punishments. T he equilibrium does not require excessive patience, and contrary to pr evious thought, need not be extraordinarily fragile. It is robust to t he introduction of small amounts of noise and remains nearly efficient . Extensions are discussed to models with heterogeneous rates of time preference and without public randomizations.