An empirical investigation of relative risk aversion

A single-attribute utility function models a decision-maker's preferences as revealed by choices among risky alternatives. A measurable value function models a decision-maker's strength of preference for various levels of the attributes. An experimental investigation of the relationship between the utility function and measurable value function for 29 cases is described. The utility function was the same as the measurable value function in only three cases. Further, the recently developed concept of a relative risk attitude was used to categorize the observed preferences. Although relative risk aversion occurred in 10 cases, it was not found to be the universal relative risk attitude. In addition, a person's relative risk attitude differed across attributes.