Abstract

Three public good mechanisms, all sharing the characteristics of collective excludability, unanimity and budget balance, are compared: The mechanisms differ in ways that are hypothesized to effect free-riding behavior with the Auction mechanism expected to show the least, and the Free-Rider and Quasi Free-Rider mechanisms showing the greatest such behavior. All three mechanisms yield mean quantities of a public good that are significantly greater than the free-rider quantity. However, the Auction mechanism provides a mean quantity of the public good which is significantly larger than that of the other two procedures, and closer to the Lindahl optimal quantity.

Introduction

Coinciding with the recent theoretical contributions of incentive compatible mechanisms for the provision of public goods, e.g. Clarke (1971), Groves (1973), Groves & Ledyard (1977), several experimental studies have tested various propositions in public goods theory; see Bohm (1972), Sweeny (1973), Marwell & Ames (1977), Ferejohn, Forsythe & Noll (1979), and Smith (1979a, b). All of these widely differing studies tend to support the proposition that decentralized mechanisms exist that allow a collective to choose and finance optimal or near-optimal quantities of a public good. Three of these studies, Ferejohn, Forsythe & Noll (1979), and Smith (1979a, b), have assumed implicitly, if not explicitly, that collectives cannot be expected to provide optimal quantities of a public good unless a decision mechanism is used that provides explicit individual incentives that favor the optimal provision of the public good.