Crowding Out Voluntary Contributions to Public Goods

We test the null hypothesis that involuntary transfers for the provision of a public good will completely crowd out voluntary transfers against the warm-glow hypothesis that crowding-out will be incomplete because individuals care about giving. Our design differs from the related design used by Andreoni in considering two levels of the involuntary transfer and a wider range of contribution possibilities, and in mixing groups every period instead of every four periods. We analyse the data with careful attention to boundary effects. We retain the null hypothesis of complete crowding-out in two of three pairwise comparisions, but reject it in favour of incomplete crowding-out in the comparison most closely akin to Andreoni's design. Thus we confirm the existence of incomplete crowding-out in some environments, but suggest that the warm-glow hypothesis is inadequate in explaining it.