Rank dependent expected utility in the pelota betting system: an experiment

Abstract

We theoretically and experimentally study a zero sum betting market: the pelota betting system, but with commonly known objective probabilities and without commissions. We know that risk-averse expected utility maximizers with identical objective probabilities cannot agree on a bet. Nevertheless, the rank dependent expected utility model allows us to explain the existence of such betting markets even assuming individuals are all identical even in utilities. We focus on behaviour in a given period in a pelota betting market and we aim to explain the volume of bets assuming that all individuals are equal and their marginal utility on wealth is decreasing. We do this in two stages. First, subjects are asked to take betting decisions and the power utility function and probability weighting function are estimated. Once the underlying utility and probability weighting function are known, in a second stage subjects interact in a betting market and we test whether the volume of bets differs from proposed theoretical predictions.