A New Method of Estimating Risk Aversion

摘要

Expected utility is the canonical theory of choice under uncertainty in economics. In the expected utility model, risk aversion arises from the curvature of the utility function, typically measured by the coefficient of relative risk aversion (γ). Despite its importance in many micro-economic and macroeconomic models, the value of γ remains disputed, largely because of limitations in estimating risk aversion empirically. This paper develops a new method of estimating γ using data on labor supply behavior. In particular, I show that existing evidence of the effects of wage changes on labor supply imposes a tight upper bound on the curvature of utility over wealth (γ < 2). Hence, the standard expected utility model cannot generate high levels of risk aversion without contradicting established facts about labor supply.