A Bayesian Analysis of Free Rider Metagames

It is well-known that private markets may fail to allocate "nonexcludable" resources efficiently because in the large numbers case individual demanders initially understate their true willingness to pay (i.e., they free ride).' In the conventional analysis of this private market failure only coercive tax-sharing can finance the Pareto optimal output of such a good. Although several challenges to this interventionist logic have arisen,2 the cumulative force of these separate counterarguments has often been missed. The purpose of this paper is to strengthen one such counterargument which challenges the persistence of any false revelation of preference for non-excludable goods. Several previous authors have argued that if free rider prisoner dilemma games are repeated over and over by the same players (in a free rider "supergame," [22; 23]) or if such a sequence of games is merely anticipated (in a free rider "metagame," [20; 22; 17]), then true revelation of preference and the Pareto-optimal cooperative solution will emerge. In