The rich, the poor and strength of Inequality: evidence from a meta-dataset of public good games experiments

Rational: Cooperation is at the root of a great many economic activities. At the level of a country, social capital, the ease with which strangers cooperate, has a sizable economic payoff (Knack and Keefer, 1997). At the micro level, cooperation is often necessary to achieve economic efficiency: for instance, communities must cooperate to avoid the exhaustion of common resources. Given the well-documented recent increase in inequalities across the world (e.g., Piketty 2014), a pressing question is whether cooperation can be sustained between agents with unequal resources. The effect of inequality on cooperation is theoretically uncertain, empirically hard to identify with observational data, and has important policy implications. Method: To contribute to this question, we identify all the published economics experiments in which heterogeneous endowments are introduced into a linear public good game. We collected the original dataset from most of these studies to build a meta-dataset. Our final dataset includes 23 papers, more than 100 treatments, and 60000 observations at the individual level. We coded a wealth of relevant variables at the paper, treatment and group level. Notably, for each observation, we computed the Gini index of endowments. The Gini index is a well-known measure of inequality. It is continuous, allowing us to measure not only the effect of the presence of inequality, but also of the strength of inequality. Results and conclusion: We run two complementary analyses. First, we investigate the effect of inequality on contributions at the group level, i.e., we relate our measure of inequality to the share of the sum of endowments contributed within a group. We document a significant negative effect of the strength of inequality on cooperation that cannot be fully due to publication bias. Second, we identify that the negative effect exists at the intensive margin: it is not only a matter of the presence of inequality but also of its strength. Second, we investigate individual contribution decisions. We identify contribution gaps between the rich and the poor: the poor contribute a higher share of their endowments than the rich or the members of equal groups. We also find that both these gaps increase as the strength of inequality (measured by the Gini coefficient of endowment) increases. Therefore, we conclude that the negative marginal effect of inequality on cooperation is driven mainly by the "rich" who reduce their participation in the public good as inequality increases.