Abstract

This study considers the influence of relaxing the widely held assumption that investors operate according to monotonically declining marginal utility as proposed by Bernoulli in 1738. The analysis is conducted by examining the assumption change has on the performance of optimal portfolio theory, the accuracy of the capital asset pricing model (CAPM), and the magnitude of the equity risk premium puzzle. The data used includes the Ibbotson monthly frequency series used by Mehra and Prescott (1984) as well as the CRSP real return series for the SP500 companies that survived on the index through the 1990's. Although the study is preliminary, it suggests that investors do indeed behave in a manner unlike Bernoulli's solution. The suggested value function of Kahneman and Tversky (1979) appears to minimize the magnitude of the premium puzzle, produces a portfolio process that offers significantly positive ex ante return for risk borne and also allows for the theoretical existence of a two-beta CAPM that better predicts asset returns. Conclusions indicate that there are both empirical and theoretical failings with the Bernoulli solution and that further refinement and study of investor utility functions may result in derived models that are superior both in estimating positive investor behavior and in prescribing normative investor behavior.