Cooperation in Continuous Dilemma and Uncertain Reaction Lag

This paper shows that cooperation can be sustained until close to the end of a finite-horizon, continuous-time prisoners’ dilemma when there is informational asymmetry in how quickly players can respond. The simulated equilibrium closely replicates recent experimental results (Friedman and Oprea, 2012, AER). The core argument is extended to a class of canonical preemption games with private information on the player’s payoff margin of preempting relative to being preempted, that can be applied to other well-known examples of conflict such as the centipede game. (JEL Classification Codes: C72, C73, D82)