Coupled lotteries : a new method to analyze inequality aversion

We develop and implement a new measure for inequality aversion: two peers are endowed with identical binary lotteries and the only choice they make is whether they want to play out the lotteries independently or with perfect positive correlation (coupling). Coupling has no other e ect than preventing outcome inequality. We implement the method in a survey in rural Thailand as well as a supplemental sample in a lab in Germany. As theoretically expected, coupling is related to being more risk averse, to having social status concerns, and to relying more often on formal and informal insurance. However, coupling is not related to giving in the dictator game.