BEER Lab Meeting - Vincent Laferrière

Multigame contact refers to cases where an individual engages in two or more strategic interactions with the same other individual. By improving a player’s ability to punish, multigame contact is thought to foster cooperation (Bernheim & Whinston, 1990). Previous experimental studies find mixed evidence on the positive effect of multigame contact on cooperation. Using results from a laboratory experiment, we investigate if and how multigame contact affects cooperation. We compare cooperation rates and subjects’ behavior when they play two indefinitely repeated prisoner’s dilemmas either with the same partner, or with two different partners. Our novel experimental design allows us to compare cooperation rates and behaviors varying the number of partners and the expected length of a repeated game. In agreement with theoretical accounts, we find that multigame contact has systematic effects on behavior: subjects link the strategies in the two games when they play with the same partner. However, in contrast to the theoretical accounts we find no systematic effect on average cooperation rates. Multigame contact is a double-edged sword: Although it helps subjects reaching simultaneous cooperation in both games (full cooperation), it also makes cooperation only in one of the two games less likely.