Working Paper: CEPR ID: DP16349
Authors: Joo Montez; Vincent Laferrire; Christian Thoeni; Catherine Roux
Abstract: Many strategic situations involve players repeatedly interacting across multiple games. While there is an extensive experimental literature on the determinants of cooperation in indefinitely repeated games, there is no evidence on how multigame contact affects behavior in these games. In theory, multigame contact should foster cooperation by improving players' ability to punish. We study the effect of multigame contact on cooperation in two preregistered laboratory experiments, where the main treatment consists of subjects playing a pair of indefinitely repeated prisoner's dilemmas with either the same partner or with different partners, i.e., respectively with and without multigame contact. In contrast to the theoretical predictions, we find no evidence that multigame contact increases average cooperation rates. Yet, and now in line with theory, we observe that multigame contact has systematic effects on behavior: subjects often link the strategies in the two games when they play with the same partner. Such linkage turns out to be a double-edged sword: while it helps subjects reaching cooperation in both games, it makes cooperation in only one of the games less likely.
Keywords: cooperation; repeated games; prisoners dilemma; multigame contact; experiment
JEL Codes: C72; C73; C91; C92; L41
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
Cause | Effect |
---|---|
multigame contact (C72) | cooperation (P13) |
multigame contact (C72) | punishment effectiveness (K42) |
multigame contact facilitates cooperation in both games (C72) | defection in one game (C72) |
linkage of strategies across games (C73) | cooperation in one game (C71) |
linkage of strategies across games (C73) | defection in the other game (C72) |
multigame contact does not enhance overall cooperation (C72) | average cooperation rates remain statistically indistinguishable (C71) |