Stability of risk preference measures: results from a field experiment on french farmers

We compare three different elicitation methods for measuring risk attitudes of French farmers in a field experiment setting. We consider two experiments based on the lottery choices initially proposed by Holt and Laury (2002) and by Eckel and Grossman (2002,2008), a risk-taking psychological questionnaire and a self-reporting of perceived risk attitudes for different domains.The main empirical results from this within-subject study are the following. First, within theclass of lottery choices, risk preference measures are affected by the type of mechanism used. In particular, farmers appear to be more risk averse using the Eckel and Grossman lottery than using the Holt and Laury one. However attitudes towards risk are significantly correlatedacross lotteries which means that the ranking of risk preferences seems to be preserved. Second,risk preferences appear to be context-dependent. French farmers are highly risk averse for decisions belonging to financial and ethical domains. They report a higher willingness to take risk for professional decisions. Lastly, using the psychological questionnaire, we find that the risk attitude elicited through lottery choices often correlates with risk attitude toward investments. These findings contribute to the literature which addresses the stability of risk preferences across elicitation methods.