Non-binding agreements in public goods experiments

This article presents experimental evidence on the effects of non-binding agreements on co-operation in public goods games. In particular, it compares first-best agreements that require full co-operation by all players and second-best agreements that require only a minimum contribution level rather than full co-operation. The results show that when there is no punishment opportunity, second-best agreements work better than first-best agreements because they are more likely to be formed and kept. First-best agreements form at the beginning of the game, but non-compliant behaviour by some players causes co-operation to collapse. This result is reversed when there is a punishment opportunity. In this case, first-best agreements work better than second-best agreements as they allow groups to establish a common behavioural standard, deter violations, and achieve very high levels of co-operation.