Stability of Experimental and Survey Measures of Risk, Time, and Social Preferences: A Review and Some New Results

Underlying preferences are often considered to be persistent, and are important inputs into economic models. We first conduct an extensive review of the disparate literature studying the stability of preferences measured in experiments. Then, we test the stability of individuals' choices in panel data from rural Paraguay over almost a decade. Answers to social preference survey questions are quite stable. Experimental measures of risk, time, and social preferences do not exhibit much stability. Correlations between experimental measures of risk aversion are a more precisely estimated zero, whereas correlations for time and social preferences are larger and noisier. We also find no systematic evidence that real world shocks influence play in games. We suggest that in a developing country context researchers should explore designing simpler experiments and including survey questions in addition to experiments to measure preferences.