The Effect of Voting on Contributions in a Public Goods Game

This paper reports the results of a public good experiment with voting. The standard game in which subjects decide simultaneously on their contributions to a public good is extended by a second stage. In this stage, subjects can express agreement or disagreement with the contributions of their group members and the resulting payoff by voting yes or no. The treatment variable is the voting threshold, which specifies how many votes are at least needed to implement the outcome. We find that average contributions are higher with a voting system, but only if the required number of votes is sufficiently high. The higher average contribution level is mainly realized because subjects manage to avoid the typical pattern of declining contributions across periods. We argue that the higher and rather stable contributions observed under high threshold levels may be related to the fact that voting is seen as a legitimate instrument. Support for this claim is provided by results from a post-experimental questionnaire.