Behavioral explanations of efficient public good allocations

Theorists have studied decentralized allocation mechanisms by characterizing the Cournot-Nash equilibria of the mechanism. However, when a mechanism is operationalized by an iterative decision process, and agents derive utility only from the final outcome, Cournot behavior may not be an appropriate solution concept. We suggest two alternative behavioral models, reservation behavior and satisficing behavior, which recognize the impact of the iterative process upon incentives, yet maintain some of the critical simplifications that have made Cournot behavior an attractive model.We report on nine public good experiments with a Groves-Ledyard general equilibrium mechanism. Overall, the experiments attain highly efficient allocations, despite widespread violations of Cournot behavior. We suggest the explanation that most subjects initially followed reservation behavior, and then came to adopt satisficing behavior.