Why punish cheaters? Those who withdraw cooperation enjoy better reputations than punishers, but both are viewed as difficult to exploit

Negatively sanctioning cheaters promotes cooperation. But do all negative sanctions have the same consequences? In dyadic cooperation, there are two ways that cooperators can sanction failures to reciprocate: by inflicting punishment or withdrawing cooperation. Although punishment can be costly, it has been proposed that this cost can be recouped if punishers acquire better reputations than non-punishers and, therefore, are favored as cooperation partners. But the evidence so far is mixed, and nothing is known about the reputations of those who sanction by withdrawing cooperation. Here, we test two novel hypotheses about how inflicting negative sanctions affects the reputation of the sanctioner: (i) Those who withdraw cooperation are evaluated more favorably than punishers, and (ii) both sanctioners are viewed as less exploitable than non-sanctioners. Observers (US online convenience sample, n = 246) evaluated withdrawers as more cooperative and less vengeful than punishers and preferred withdrawers as a partner. Sanctioners were also viewed as more difficult to exploit than non-sanctioners, with no difference between punishers and withdrawers. The results were the same when punishment was costly (US college sample, n = 203) with one exception: Costly punishers, who lost their payoffs by punishing, were viewed as more exploitable than withdrawers. Our results indicate that withdrawing cooperation has advantages over punishing: Withdrawers are favored as cooperative partners while gaining a reputation as difficult to exploit. The reputational consequences of the three responses to defectors—punishing, withdrawing cooperation, and not sanctioning at all—were opposite to those predicted by group selection models.