Top Contributors as Punishers

We experimentally investigate cooperation within a finitely repeated public goods game framework where peer punishment is possible but, unlike previous work, in each round access to sanctioning power is exclusively awarded to the group’s top contributor. We compare this mechanism with a treatment where the right to punish is assigned to one randomly selected subject (O’Gorman et al., 2009), as well as with classic discretionary punishment (Fehr and Gächter, 2000) and with ‘legitimate punishment’ (Faillo et al., 2013). We show that the “Top Contributors as Punishers” mechanism is extremely effective in both raising cooperation and welfare, compared to the randomly selected punisher treatment and to discretionary punishment. This interestingly occurs despite the fact that the (first and second-order) free riding problem may lead subjects to perceive the new institution as an excessively demanding one: in fact, the lure of the top contributor role induces many subjects to significantly contribute and many top contributors to incur relevant costs to sanction others.