The Carrot or the Stick: Rewards, Punishments, and Cooperation

Many institutional arrangements suggest that punishments and rewards each play a separate role in providing incentives. In New York City’s recent negotiations with its teacher’s union, for instance, the city sought a contract that would strengthen school principals’ ability to assign teachers to be cafeteria monitors, in part because this allows them to punish underperforming teachers. The contract also includes rewards for teachers with good performance. Another example is that some universities now use a combination of raises and differential teaching loads to encourage good performance. Similarly, procurement and production contracts and government regulations in areas ranging from meat inspection to sulphur dioxide pollution often include both bonuses for good performance and various sorts of clawbacks for bad. Punishments and rewards also play an important role in informal relationships. We seem to care about the unwritten social contract that people should be cooperative and kind to others, and we are often willing to pay to enforce these ideals. For instance, every time we take a taxi or sit in a restaurant we entrust our happiness to another person. If that trust is protected then we may reward it with a generous tip, but if not we may leave less than usual, or nothing at all. Likewise we may shun unfriendly colleagues but invite the friendly ones to our homes, and secretaries may perform more promptly for those who are polite or bring gifts and less promptly or poorly for those who are rude or unfriendly. Absent complete contracts, voluntary punishments and rewards are the mechanisms we use to sustain cooperation. Understanding what triggers the demands for each and their effectiveness may help us understand informal relations and to design institutions that can harness cooperation and improve social welfare. The objective of this paper is to begin a systematic look at both punishments and rewards in economic laboratory experiments. Experimental studies have demonstrated substantial demands for both punishments and rewards. In one-shot proposer-responder games people are willing to sacrifice personal payoffs to punish those who are unkind, and reward those who are kind. These studies typically examine either rewards or punishments, but not both. Thus they make it difficult to determine what motivates these demands, how they interact, or how they affect cooperation.