On the Merits of Separate Spaces: Why Institutions Isolate Cooperation and Division Tasks

Do institutions shape the possibility of sustaining cooperation when the same individuals must first divide resources and then attempt to cooperate? It could be that simply having received an inequitable division undermines cooperative behavior, reducing aggregate welfare. Alternatively, it might be that only when interacting with the same individual or group does this spillover occur, in which case separating tasks across institutions may prevent this negative spillover. To test these arguments, we designed a two-stage incentivized experiment in which participants interact in a division task and then in a task in which cooperation improves aggregate welfare. In two experiments, individuals were randomly assigned to interact either with the same individual for both tasks or with a different individual for each task. In the second experiment, individuals could also interact with a person who was in the same arbitrary group as their partner in the division task. Holding constant both past history and past partner behavior, the results of these experiments provide support for a Partner History effect in which the mechanism that produces spillover is interacting with the same individual in both decisions. We also find evidence for a weaker Group History effect in which negative spillover occurs when the partner in the cooperative task is a member of the same group as the partner from the division task.