It has been well established that charities often choose to announce past contributions during fundraising drives. However, this type of announcement strategy appears to be inconsistent with the standard prediction that sequential move games result in smaller contribution levels than simultaneous move games. The project outlined in this proposal will develop a theory that explains the use of announcement strategies in fundraising games. In addition, two experiments will be conducted to test the theoretical predictions for contributions in sequential public good games. The proposed project will not only improve understanding of the sequential nature of fundraising games but also our overall understanding of charitable giving. The results of the proposed research will be documented in six publishable papers--four theoretical papers and two experimental papers. In the first paper, a simple model is developed in which contributors are identical and the quality of the charity can be revealed by acquiring costly information. Results indicate that fundraisers for high quality charities prefer to announce past contributions. The reason is that the first contributor, contingent on the announcement, purchases information and reveals the quality of the charity through his first contribution. The second paper l relaxes the assumption that contributors are identical to demonstrate that heterogeneity of contributors implies an optimal solicitation ordering, and this ordering makes it impossible for the fundraiser to cheat and solicit first contributions from multiple contributors. In the first two papers the purchased signal is perfect and the fundraiser has no incentive to make more than one announcement. One reason why fundraisers make continuous announcements may be that the signal is noisy. Such a model is explored in the third paper. To keep the analysis simple, information on the purchasing option is ignored and instead, it is assumed that all contributors receive private and noisy signals. In this model contributors use information about past contributions to update their beliefs about the charity's quality, and an informational cascade is likely to arise. Two different experiments test the predictions of the purely altruistic theory for contributions in sequential public good games. The first experiment tests the predictions of sequential move contribution games by examining the contributions in three public good games: a simultaneous game, a sequential game and finally a sequential game where the first contributor cannot commit to a one-time contribution. The second experiment examines the effect of the public good's production technology on contributions in sequential public good games. In addition to testing the traditional voluntary contribution predictions these two experiments also provide empirical tests of recent models of fairness. An interesting feature of these fairness models is that they predict higher contribution levels in sequential public good games than in simultaneous move games, suggesting that preferences for fairness may be a third explanation for the use of announcement strategies in fundraising games.
