Collective sanction enforcement: New experimental evidence from two societies

Abstract

This paper presents the first experimental study on how higher-order punishment affects third-party sanction enforcement in the presence of multiple third parties. The design varies across treatments the number of third parties witnessing a norm violation and the opportunities available for third parties to costly punish each other after observing their peers’ enforcement actions. To test generalizability of higher-order enforcement effects, the experiment is conducted across two contrasting societies – India and the United Kingdom – using a prisoner’s dilemma game. These societies are selected for their positions at opposite ends of the tight-loose ancestral kinship spectrum. In both societies, third parties punish defectors who exploit their paired cooperators more strongly than any other person, consistent with prior research. Yet, punitive patterns differ. In the UK, third parties punish defectors less frequently and less strongly when other third parties are present; when higher-order punishments are available among third parties, their failure to punish defectors and acts of anti-social punishment invite strong higher-order punishment from their peers, which encourages their pro-social first-order punishments and makes mutual cooperation a Nash equilibrium outcome in the primary cooperation dilemma. However, in India, overall punishment levels are lower, group size and incentive structure changes have no discernible effects, and higher-order punishments are not better disciplined. These findings support a model of norm conformity for the UK and do not contradict such a model for India.