Time-stability of risk preferences: A new approach with evidence from developed and developing countries

Abstract

Time-stability of preferences is a crucial assumption in economics. We develop a novel test-retest method to examine the stability of risk preferences over time, while quantifying the importance of both idiosyncratic shocks and measurement error. Using eight large, representative datasets from developing and developed countries, we find risk preferences to be unstable in developing countries. In contrast, they are very stable in developed countries, except for low-income individuals in the U.S.. We discuss the important implications of these findings for policies and research.