Challenges in global politics like climate change, maritime piracy and fighting highly contagious diseases concern global public goods. The related policy decisions are mostly made by teams. In contrast, economic models of global public goods typically assume a single rational decision-maker. We use a laboratory experiment to compare team decisions to decisions of individuals in a finitely repeated public good game with and without a costly punishment option. Teams of three participants coordinate on decisions either by majority or unanimity rule. We find that in absence of a punishment option teams contribute more to the public good than individuals. With a punishment option subsequently to the contribution decision team treatments exhibit a less frequent use of anti-social punishment and lower levels of social as well as anti-social punishment. Extreme preferences for punishment are eliminated by the majority decision rule. Overall, team decisions are closer to the social optimum and teams yield higher net payoffs when compared to individuals.

Challenges addressed in global politics, such as climate change, maritime piracy and fighting highly contagious diseases, concern global public goods. The related policy decisions are primarily made by teams. In contrast, economic models of global public goods typically assume a single rational decision maker. We use a laboratory experiment to compare team decisions to the decisions of individuals in a finitely repeated public goods game with and without a costly punishment option. Teams of three participants coordinate on decisions either by majority or unanimity rule. We find that, in the absence of a punishment option, teams deciding by the unanimity rule contribute slightly more than individuals. With the punishment option, unanimity teams choose significantly lower levels of punishment and exhibit anti-social punishment less frequently than individuals. A possible explanation is the elimination of extreme preferences for punishment through the coordination process in teams.