Stochastic Expected Utility and Prospect Theory in a Horse Race: A Finite Mixture Approach

This study compares the performance of Prospect Theory versus Stochastic Expected Utility Theory at fitting data on decision making under risk. Both theories incorporate well-known deviations from Expected Utility Maximization such as the Allais paradox or the fourfold pattern of risk attitudes. Stochastic Expected Utility Theory parsimoniously extends the standard microeconomic model, whereas Prospect Theory, the benchmark for aggregate choice so far, is based on psychological findings. First, the two theories' fit to representative choice is assessed for two experimental data sets, one Swiss and one Chinese. In a second step, finite mixture regressions reveal a consistent mix of two different behavioral types suggesting that researchers may take individual heterogeneity into account in order to avoid aggregation bias.