When their own payoff is unknown, all subjects prefer a more unequal distribution provided it makes everyone weakly better off; 25% burn money to make things less unequal, even when it does not make anyone better off

Preferences over income distribution: Evidence from a choice experiment. Sophie Cetre et al. Journal of Economic Psychology, August 28 2019, 102202. https://doi.org/10.1016/j.joep.2019.102202

Highlights
•    We elicit subjects preferences over pairs of payoff distributions within small groups, in a firm-like setting.
•    When they do not know what their own payoff will be, all subjects prefer the more unequal distribution provided it makes everyone weakly better off. This is true, no matter whether income positions will be based on merit or luck.
•    People change their choice when they learn about their own payoff rank.
•    Even then, 75% of subjects prefer Pareto-dominant distributions.
•    However, 25% engage into money burning at the top in order to reduce inequality, even when it does not make anyone better off.
•    Even when their own money is at stake, 20% of subjects are willing to pay just to reduce the degree of inequality.

Abstract: Using a choice experiment in the lab, we assess the relative importance of different attitudes to income inequality. We elicit subjects’ preferences regarding pairs of payoff distributions within small groups, in a firm-like setting. We find that distributions that satisfy the Pareto-dominance criterion attract unanimous suffrage: all subjects prefer larger inequality provided it makes everyone weakly better off. This is true no matter whether payoffs are based on merit or luck. Unanimity only breaks once subjects’ positions within the income distribution are fixed and known ex-ante. Even then, 75% of subjects prefer Pareto-dominant distributions, but 25% of subjects engage in money burning at the top in order to reduce inequality, even when it does not make anyone better off. A majority of subjects embrace a more equal distribution if their own income or overall efficiency is not at stake. When their own income is at stake and the sum of payoffs remains unaffected, 20% of subjects are willing to pay for a lower degree of inequality.