Inferring Risk Attitudes from Certainty Equivalents : Some Lessons from an Experimental Study

This paper investigates experimentally whether certainty equivalents (CE) can be useful indicators for an individual's risk attitude. It is found that the reliability of this indicator (i.e. of a CE elicited by a Vickrey auction) is rather low. The possibility that the Vickrey auction mechanism causes considerable distortions can be ruled out, but noise in observing one's CE may explain the low degree of reliability in the data. This casts doubt on the usefulness of CE indices for the measurement of individual risk attitudes