Essays on the public goods game

Justin D Krieg, Purdue University

Abstract

This dissertation consists of three essays on the public goods game. Each essay is an experimental examination of the game and investigates whether behavior in the lab environment is affected by changes to the basic structure of the game. The first essay investigates if differences in information within a group impacts behavior. In each group there are two subject types, a manager who observes the individual contribution choices of each group member and donors who can only view the total amount contributed. The difference in information does not lead to a significant difference in average contributions for the two types. It is suggested that no effect is found because both information sets allow subjects to compute the average contributions of others. The second essay investigates how agency problems at public goods producers, such as non-profit organizations, impact the voluntary provision of public goods. One subject is assigned as the manager, as in the first essay, but in certain treatments the manager is given the ability to exploit the contributions of others. These managers often take advantage of the opportunity to exploit the contributions of others but do so strategically. This strategic behavior reduces the impact of a change in the marginal per capita return and results in contribution behavior that is no longer conditional on the behavior of others. Interestingly, there is little effect on donors, who behave in a manner consistent with the existing literature on voluntary contribution games. There are two notable exceptions. First, donors contribute significantly lower amounts when they can be exploited and the exploitation is unobservable. Second, when exploitation is observable managers are successful at maintaining contributions at intermediate levels until end periods when contributions fall to zero. This results in average contribution levels that are not different from treatments where exploitation is not possible. These results suggests one potential impact agency has on the provision of public goods is through the transparency offered by the organizations producing the public good. The final essay asks the basic question of whether behavior commonly observed in linear public goods games differs when the game is played in a two-game ensemble. This is an important question because individuals face multiple giving opportunities or multiple options for giving within a single organization. The results suggest that even when the two games are identical there is a spillover of information between the two games. Certain treatments fail to be effective at raising average contributions when the games are played in a two-game ensemble; this includes non-monetary sanctioning and the display of digital photos. In contrast, lowering the cost of contribution in one game of the ensemble is effective at raising average contributions for that game. Together the results support the position that games should be considered in ensembles instead of in isolation.

Degree

Ph.D.

Advisors

Cason, Purdue University.

Subject Area

Economics

COinS