The Poor, the Rich and the Middle Class: Experimental evidence from heterogeneous public goods games

We present the results of one-shot and repeated public good experiments that seek to understand the interaction between the endowment and marginal return in heterogeneous groups. Our focus is on treatments in which the endowment and the marginal return are either inverse or proportionally related to each other. While two normatively appealing contribution rules are aligned in the proportional treatment, a conflict between the two exists in the inverse treatment. In the one-shot experiments, we do not find significant differences across treatments. Contributions increase when the endowment, the marginal return or both increase. This is observed in all treatments except when the endowment and the marginal return are inversely related. In this case, the 'middle class' participants contribute more than both the high and low endowment types, mirroring real world observations with regards to a 'squeezed middle'. This suggests the presence of a conflict between the highly endowed subjects (but with low marginal return) and those with a high marginal return (but with low endowment). This pattern is similar when we elicit beliefs about others' contributions, whereby the two conflicting types expect others to contribute more than they do for themselves. In the long-run, however, when allowing for repeated interaction, the differences across types vanish in the inverse but not in the proportional treatment. This suggests that over time the conflicting interests arising from the interplay between the endowment and the marginal return can be overcome. Our findings have welfare implications indicating that the inverse treatment reduces inequality measured by the Gini coefficient but this is not the case for the proportional treatment, where inequality remains the same.