Efficient Non-Cooperative Provision of Costly Positive Externalities via Conditional Commitments

Abstract

We consider games where individual contributions are costly but beneficial to other players, so that contributing nothing is a dominant strategy. Considering that players may be unable to write binding agreements but may make binding unilateral commitments that are conditional on others' actions, we study a mechanism based on conditional commitment functions (CCFs). If players must choose their CCFs once and simultaneously, the mechanism contributes to the Nash program since its strong (or coalition-proof) equilibria realize precisely the core outcomes of the corresponding bargaining problem. If players can communicate, the outcome can thus be expected to be Pareto-efficient. Even without communication, the core outcomes may be found by simple individual learning rules. We motivate the idea in a Cournot duopoly and a public good problem and then derive our results in a very general decision-theoretic framework and give further examples from different areas of economics.