Heterogeneity in Risk Preferences: Evidence from a Real-World Betting Market

We develop a structural model of behavior that accounts for individual heterogeneity within and across utility 'types' characterized by different features of risk preferences, and we estimate it using a unique dataset of individual activity in a sports wagering market. We estimate the population distribution of utility curvature, loss aversion, and probability weighting, and we evaluate their importance in explaining behavior. We find — for the first time using individual-level field data — that all features of prospect theory are present outside the lab: Individuals have heterogeneous risk attitudes, but on average they are risk averse (loving) over gains (losses), exhibit loss aversion, and overweight the probabilities of extreme outcomes. Furthermore, we find that utility curvature alone does not explain choices and that, while loss aversion is important, probability weighting is the most prevalent behavioral feature of risk attitudes: Two thirds of individuals exhibit loss aversion, but all exhibit probability weighting.