WP 2020-15 Estimating Risk Preferences for Individuals: A Bayesian Approach
AUTHORS: Xiaoxue Sherry Gao, Glenn W. Harrison and Rusty Tchernis
ABSTRACT: Risk preferences play a central role in many descriptive and normative inferences about economic behavior. It is natural to expect that risk preference are heterogeneous. We evaluate a series of Bayesian econometric models that allow for different levels of heterogeneity of individual risk preferences. We carefully compare inferences about risk preferences across these models using both simulated and observed data. Using simulated data, we show the effectiveness of the Bayesian approach in recovering the underlying true parameters and evaluate the cost of model mis-specification. Using observed data, we evaluate the heterogeneity present in a typical subject pool. We suggest extensions and applications of the models and illustrate the application to the evaluation of behavioral welfare gains or losses from insurance purchases.