Doe the endowment make a difference in threshold public good games

What is it about?

In a threshold public good game members of a group are independently asked to contribute towards the provision of a public good. The public good would benefit the group but is only provided if total contributions are sufficiently large. There are four basic parameters in any threshold public good game - the number of people in the group, the endowment of money group members have to contribute, the value of the public good, and the threshold above which the public good is provided. Prior work has focused on the value of the public good and threshold. In this paper we report on an experiment in which we systematically vary the endowment in order to see whether this influences efficiency in providing the public good. A-priori there are two factors at work here. First, the larger the endowment the 'cheaper' it is to contribute to the public good. This may increase efficiency. Second, the smaller the endowment the more 'critical' any one group member is to the overall chance of providing the public good. This may also increase efficiency. Consistent with these two competing effects our experimental results show some evidence of a U shaped relationship in which efficiency is highest for a small or large relative endowment. It should be clarified, however, that the differences in efficiency between treatments were small. More noticeable was a much higher variance of contributions for the intermediate level of endowment. This points towards a difficulty of group members coordinating on a way to provide the public good.

Featured Image

Why is it important?

There has been a number of experimental studies looking at whether groups are efficient at providing threshold public goods. None of those studies, however, had explicitly looked at the endowment. So, this was a gap in the literature. In this paper we show that when there is a money back guarantee the level of endowment does not seem to matter much for efficiency, although it can effect the variance in contributions. In a companion paper we show that without a money back guarantee the endowment matters more.