Abstract:
This paper studies n-player collective-action games in which a public good is produced if and only if m or more volunteers contribute to it. Quantal-response strategy revisions allow play to move between equilibria in which a team of m players successfully provide, and an equilibrium in which the collective action fails. A full characterisation of long-run play reveals the determinants of success. These include the correlation between players` costs of provision and their valuations for the good. The addition of an extra "bad apple" player can "spoil the barrel" by destabilising successful teams. A contemporary application is the team-based provision of open-source software. The analysis reveals the features of successful open-source projects, and suggests a rationale for limiting the pool of possible contributors.