Collective action as an agreeable n-prisoners' dilemma

Abstract

The problem of collective action to produce a group collective good is analyzed as the game of Individual vs. Collective and then as an n-person game to show that, under the constraints of Mancur Olson's analysis, it is an n-prisoners' dilemma in the cases of latent and intermediate groups. The usual analysis according to which noncooperation is considered the rational strategy for classical 2-prisoners' dilemma is logically similar to Olson's analysis, which suggests that rational members of a latent group should not contribute toward the purchase of the group collective good. However, in the game analysis it is clear that the latent and intermediate groups are not logically different, but rather are distinguishable only statistically. Some prisoners' dilemma experimental results are used to suggest how the difference might arise and how the vast prisoners' dilemma literature can be related to the problem of collective action. The game of collective action is then analyzed not from the view of strategies but of outcomes. There is presented a theorem which states that the outcome in which all player-members of a group pay and all benefit is a Condorcet choice from the set of realizable outcomes for the game. Hence the cooperative outcome in such a game would prevail in election against all other outcomes.