Voluntary financing of a public good: an experimental analysis of cooperation.

Summary The principle of voluntary contribution for the financing of a public good constitutes an appropriate alternative to the impossibility of resorting to a market mechanism by financing at marginal cost, because of the principles of non-rivalry and non-exclusion. This principle can be likened to a prisoner's dilemma game. The Nash equilibrium of such a game corresponds to the choice of a zero contribution to the financing of the public good on the part of the agents (free rider behavior), whereas the Pareto optimum assumes that all contribute their full endowment (complete cooperation). The satisfaction of individual interest is thus opposed to the satisfaction of collective interest. Experimental analyses of such a voluntary contribution game, however, reveal differences in behavior that lead to deviations between theoretical predictions and observed results, in the sense of over-contribution relative to equilibrium (Davis and Holt, 1993; Ledyard, 1995). Our study focuses on the importance of remuneration methods in the emergence of cooperation in voluntary contribution behaviors, with a focus on two directions. A first explanatory model of divergent contribution behavior introduces an endogenous asymmetry of individuals' resources in the form of reinvestment of the gains made during the game. In theory, while the asymmetry of resources reinforces the choice of free rider behavior, the introduction of the reinvestment mechanism increases the individual interest in cooperation. This model is the subject of an experiment whose results do not refute the theoretical predictions. A second model experimentally studies suboptimal overcontributions in order to decide between two hypotheses. The first hypothesis assumes that a lack of understanding of the rules of the game is at the origin of cooperation. The second hypothesis points to an aversion to financing due to the level of commitment required to finance the public good efficiently. The introduction of a level of efficient financing internal to the contribution space tests these hypotheses. The experimental evidence supports the second hypothesis.