Renegotiation-proof equilibria in repeated prisoners' dilemma (Q1264101): Difference between revisions

The usual proof of the ``Perfect Folk Theorem'' uses punishments that are rational for the punisher only if the original deviator will enforce their use by punishing the punisher for not punishing the original deviation. In other words, in the continuation game following a deviation the usual subgame perfect continuations are often Pareto dominated. \textit{J. P. Farrell} [``Credible repeated game equilibrium'', mimeo (1983)] proposed the concept of ``renegotiation-proof equilibrium'' to avoid this difficulty and conjectured that the only subgame perfect equilibria of the Prisoner's Dilemma to survive the renogotiation- proofness requirement would induce the repeated one-shot Nash outcome. In this note, the author demonstrates that Farrell's conjecture is spectacularly wrong: when players are sufficiently patient, any feasible and individually rational outcome can be sustained by means of an equilibrium in which the punishing player profits by implementing the punishment. That is (for the Prisoner's Dilemma), there is a ``Renegotiation-proof Folk Theorem''.