Fairness and limited information: Are people Bayesian meritocrats?

This paper studies how limited information about the source of inequality affects inequality acceptance. We show theoretically that non-Bayesian belief updating can be as important as heterogeneity in fairness views for understanding disagreements about inequality: in a large class of economic environments, signal-neglecting meritocrats consider the egalitarian distribution to be fair, while base-rate neglecting meritocrats consider the libertarian distribution to be fair. We provide experimental evidence showing significant heterogeneity in the updating behavior of meritocratic spectators, and, as a result, polarization in inequality acceptance. Finally, we show that the distributive decisions of third-party spectators are best explained by a structural model that allows for heterogeneity both in fairness views and belief updating, with heterogeneity in belief updating causing twice as much polarization in inequality acceptance as heterogeneity in fairness views. The findings contribute to better understanding of the foundations of inequality acceptance in society.