Optimal incentives with state-dependent preferences

Abstract In both experimental and natural settings incentives sometimes under-perform, generating smaller effects on the targeted behaviors than would be predicted for entirely self-regarding agents. A parsimonious explanation is that incentives that appeal to self-regarding economic motives may crowd out non-economic motives such as altruism, reciprocity, intrinsic motivation, ethical values and other social preferences, leading to disappointing and sometimes even counter-productive incentive effects. We present evidence from behavioral experiments that crowding may take two forms: categorical (the effect on preferences depends only on the presence or absence of the incentive) or marginal (the effect depends on the extent of the incentive). We extend an earlier contribution (Bowles and Hwang, 2008) to include categorical crowding, thus providing a more general framework for the study of optimal incentives and as a result, an expanded range of situations for which the sophisticated planner will (surprisingly) make greater use of incentives when incentives crowd out social preferences than when motivational crowding is absent