Other-regarding preferences and altruistic punishment: A Darwinian perspective

Altruistic punishment - the punishment of norm violators at one's own cost without material benefit - is frequently observed in experimental economics, field studies and in people's everyday life. The existence of this ostensibly irrational behavior is often linked to other-regarding preferences in humans, which relate a person's decision to her social environment. Economists have begun incorporating other-regarding preferences in their utility frameworks to better capture empirical evidence. While these theoretical additions seem intuitive, their evolutionary foundation is poorly documented and the origin of altruistic punishment remains puzzling. Using a new approach, that closely integrates empirical results from a public goods experiment with punishment together with an evolutionary agent-based simulation model, enables us to identify the underlying key mechanism of this puzzling behavior. We design the agent-based model by means of a limited set of behavioral patterns observed in the experimental data. We validate our results by comparing our quantitative predictions of the punishment behavior and its effect on the level of cooperation with the corresponding empirical observations. Unlike descriptive theories, our approach provides a broad explanatory solution for the emergence and existence of altruistic punishment behavior in public goods game experiments: Altruistic punishment emerges spontaneously from a population of agents who are initially non-punishers, if variants of inequality or inequity aversion are the predominant preference sets. In particular we find, that a weaker, asymmetric variant of inequity aversion quantitatively explains the level of punishment observed in contemporary experiments: Our evolutionary model shows that disadvantageous inequity aversion cause altruistic punishment behavior to evolve to a level that precisely matches the empirical observations in public goods experiments.