Intermediation Reduces Punishment and Reward∗

Most economic theories of punishment assume the moral decision-maker has in mind a model of fairness based on outcomes, intentions and procedures and punishes deviations from fairness. In contrast, this paper shows that moral decision-making is better modeled as an unreasoned, intuitive reaction, correlated but not tethered to the punisher’s ideals of right and wrong. Specifically, the experiments in this paper demonstrate that misbehavior is punished substantially less when its consequences do not directly follow; moral judgment is immoderately focused on the immediate outcomes of an action. Moreover, this narrow judgment is shown to greatly reduce the effectiveness of punishment and reward, which allows anti-social behavior to go unsanctioned and prosocial behavior unrewarded. In Study 1, in the laboratory, the action of keeping money at the expense of a poorer player is punished less when done through an intermediary, even when the intermediary is not involved in selecting the outcome. In the game, the intermediary can only make the poorest player worse off, so current fairness models would predict an increase in punishment, the opposite of these results. Many social preference papers find a discontinuity between positive and negative actions, so in Study 2, I test the robustness of the psychology behind the main result from Study 1 in a charity/reward domain. In Study 2, a framed field experiment, reward of charitable behavior (donating mosquito nets) decreases when the saliency of an intermediary (the charity) is increased. Together the results show that moral decision-making is not wellpredicted solely by normative ideals. As a result punishment and reward are unreliable alone in enforcing pro-social equilibria. JEL classification: C92, D63