Lotteries as a funding tool for financing public goods

In the Review of Economic Studies, Morgan (2000) proposed that targeted self-funding lotteries could be used as a method of increasing voluntary contributions to public goods. In the same issue, Morgan and Sefton (2000) tested the theoretical predictions in a laboratory experiment and found support for the theory. The theory, and consequently the experiment, both assumed a quasi-linear utility function with one private good and one representative public good. This current research asks the question, does such a lottery work when there are two public goods? In the original case, expected utility maximization causes agents to divert funding away from the private good and towards the public good. Enough resources are diverted to not only fund the lottery prize but also to lead to an overall increase in public good provision thereby increasing social welfare. However, when two public goods are involved, funds are diverted from both the private good and from any out-of-equilibrium voluntary contributions made to the public good that does not involve the lottery. This paper presents the theory and an initial experiment run at CEEL in Trento using PGLottery software designed at McEEL (McMaster) and CEEL. There are two key findings. First, behaviour in a multiple public good experiment seems to differ from behaviour in traditional single public good experiments. Second, opposite to the findings of Morgan and Sefton (2000), the introduction of the lottery decreases efficiency, adding evidence to the argument that lotteries decrease social welfare.