Standard prisoners' dilemma games offer players the binary choice between cooperating and defecting, but in a related game there is the third possibility of leaving the game altogether. We conceptualize exiting as taking the individual beyond the reach of externalities generated in the original group, and on that basis—together with the assumption of self-interested (dollar-maximizing) behavior on the part of all players—we derive the prediction that the exit option will drain the community or group more of cooperators than of defectors.
But experimental data do not support this prediction; cooperators do not leave more frequently than defectors and, in fact, there is evidence that defectors are more prone to leave than cooperators. We consider and reject the possibility that this failure of prediction results from the (admitted) greater optimism of cooperators about the incidence of cooperation “here,” and present data supporting the hypothesis that cooperators often stay when their personal interest is with exiting because of the same ethical or group-regarding impulse that (presumably) led them to cooperate in the first place. Cooperation can be produced for a group or community either by inducing people to cooperate or by inducing those who are going to cooperate to stay in the game, and ethical considerations seem to underlie the decision to stay as well as the decision to cooperate while staying.