Inconsistent choices over prospect theory lottery games: Evidence from field experiments

Abstract

Expected utility theory and classroom experiments have been commonly used to test the consistency and/or stability of elicited risk preferences among individuals. Usually conducted in developed countries, these studies have shown that risk preferences are inconsistent or unstable. However, the literature falls short when assessing the consistency of risk preferences under the Prospect Theory (PT) paradigm. This study reports the results of lottery games, played during two consecutive experimental sessions, to test the consistency of PT risk preferences among smallholder farmers in Ecuador, a developing country. I find evidence of consistency in the farmers’ risk preferences at the sample level. Nonetheless, I find inconsistency in risk preferences (risk aversion, probability distortion, and loss-aversion) at the individual level. I assess whether there is heterogeneity in the results across farmers’ demographic and socioeconomic attributes. Evidence suggests that relative consistency in the degree of loss aversion is higher among farmers who are female, own larger farms, and are more educated.