Essays on Institutions for Facilitating Cooperation in the Provision of Public Goods

Abstract

Three essays approach the problem of methods to facilitate cooperation in the provision of public goods by expanding opportunities for reciprocity using the traditional voluntary contributions mechanism (VCM). In the first two essays, mechanisms that allow subjects to sanction or reward other group members are studied. The third essay examines how the opportunity to make binding cooperative commitments affects behavior in the VCM. In the first essay, rewards and sanctions are examined in a one-shot VCM setting that so far has been unexplored in the literature. The research finds that while some subjects are willing to reward and sanction others at a personal cost, the opportunity to reward or sanction is ineffective in facilitating cooperation relative to previous experiments in which a repeated game environment is employed. The essay also compares behavior in decision situations in which the imposition of rewards and sanctions is certain to decision situations in which imposition is uncertain. Uncertainty does not change behavior in a significant way, either in the level of cooperation or the willingness of individuals to impose rewards or sanctions. The second essay expands on the first essay by examining rewards and sanctions that vary in relative size in relation to the cost of their imposition. Each type of reward or sanction is examined in a one-shot voluntary contribution mechanism setting. In every environment, some subjects are willing to reward or sanction other subjects at a personal cost. Evidence is found that contributions are significantly increased in the environment in which the cost of sanctioning is the least relative to the size of the sanction. Finally, the third essay examines the effect of allowing binding multi-round commitments to the group account in a repeated voluntary contributions mechanism game. Subjects are found to make commitments averaging between 25% and 35% of their endowments. However, total group-account allocations are not systematically greater on average in the commitment experiments than in otherwise identical control experiments without commitments. Further analysis reveals that subjects respond to commitments in a reciprocal manner, and that the variance of outcomes across groups is larger in the commitment experiments than the control experiments.