Ex interim voting in public good provision

We report the results of an experimental study that compares voting mechanisms in the provision of public goods. Subjects can freely decide how much they want to contribute. Whether the public good is ¯nally provided is decided by a referendum under full information about all contributions. If provision is rejected, contributions are reduced by a fee and reimbursed. We compare unanimity with majority voting and both to the baseline of cheap talk. Contributions are highest under unanimity. Yet, results concerning overall e±ciency are mixed. When provision occurs, only unanimity enhances e±ciency. Overall, however, unanimity leads to too many rejections.