Costly punishment

In order for cooperation to be maintained in human groups, there must be a way to discourage cheaters. When each individual’s cooperative history is publicly available, free-riding doesn’t pay because it results in a lousy reputation and no access to reputational benefits. However, in an anonymous setting, or when reputations are hard to track, free-riding may easily spread. In such circumstances, another way in which cooperation could be enforced is by punishment.

In behavioural economic experiments, punishment is expressed by paying a fee in order to reduce the payoff of another person by a larger amount (typically at a 1:3 ratio). Therefore, while punishment is costly to the punisher, it is even more costly to the punished. Researchers working on costly punishment have distinguished two types of sanctions: punishment of free-riders, known as altruistic punishment, and punishment of those who behave cooperatively, known as anti-social punishment.

After early empirical tests, altruistic punishment appeared to be an ideal mechanism for enhancing cooperation. Experiments in the lab suggested that individuals are willing to use altruistic punishment even when they, themselves, cannot benefit from their group’s increased cooperation.

However, a later cross-cultural study on costly punishment revealed that, in some subject pools, a significant number of individuals were willing to use anti-social punishment and sanction cooperators. Anti-social punishment was frequent in subject pools from Muscat (Oman), Athens (Greece), Riyadh (Saudi Arabia), Samara (Russia), Minsk (Belarus), Istanbul (Turkey) and Seoul (South Korea). The occurrence of anti-social punishment  correlated negatively with the country’s GDP per capita, democracy, norms of civic cooperation and the rule of law.

During my post-doc in Bath, I re-analysed this cross-cultural dataset and obtained results that made me question the current definitions of altruistic and anti-social punishment. My colleagues and I have recently submitted a paper based on these results, in which we propose that costly punishment is best understood as a form of aggressive behaviour.