Cooperation in Multiplayer Dilemmas
Title: Cooperation in Multiplayer Dilemmas (with Ismael Martínez-Martínez)
Abstract: We analyze multiplayer prisoner’s dilemmas in continuous-time experiments. We control for strategic incentives when the number of players changes (that is, we keep constant the payoffs from all defect, all cooper-ate, and unilateral defection and cooperation). By studying three different prisoner’s dilemma variants for various numbers of players, we can disentangle strategic incentives and strategic uncertainty. In a further treatment variation, we allow for the explicit use of mixed strategies. We find that cooperation rates decay in the number of players, suggesting that strategic uncertainty is a prominent driver of cooperation. Strategic incentives also matter, even in treatments with a higher number of players. Allowing for the explicit use of mixed strategies strongly reduces cooperation rates. Strategy frequency estimations suggest that the strategies adopted in larger groups are not so dissimilar from those in two-player experiments. At an aggregate level, we find that proportional Tit-For-Tat well explains cooperation decisions.