Risk and Time Preferences: Linking Experimental and Household Survey Data from Vietnam

Abstract

We conducted experiments in Vietnamese villages to determine the predictors of risk and time preferences. In villages with higher mean income, people are less loss-averse and more patient. Household income is correlated with patience but not with risk. We expand measurements of risk and time preferences beyond expected utility and exponential discounting, replacing those models with prospect theory and a three-parameter hyperbolic discounting model. Comparable risk parameter estimates have been found for Chinese farmers, using our method.

The original article first appeared in the American Economic Review 100(1):557–571, 2010. A newly written addendum has been added to this book chapter.

Appendices

Appendix

Table 1.9 Switching point (question) in Series 1 and 2, and approximations of σ (parameter for the curvature of power value function) and α (probability sensitivity parameter in Prelec’s weighting function)

Table 1.10 Pairwise time discounting choices

Table 1.11 Correlations with present bias and discount rates (OLS) without trusted agents

Addendum: The Impacts of Risk Preferences on Technology Adoption in Agriculture

This addendum has been newly written by Tomomi Tanaka for this book chapter.

In the chapter entitled “Risk and time preferences: Linking experimental and household survey data from Vietnam”, we examined how basic preferences, namely risk and time preferences, are linked to wealth. We hypothesized that (1) risk averse people are reluctant to enter into risky but profitable economic activities, and (2) impatient people do not engage in long-term projects such as educating their children, so thus remain poor. We conducted risk and time discounting experiments in nine villages in Vietnam and investigated whether risk and time preferences correlate with income, relative income within village, and mean income of village. We found mean village income is correlated with risk and time preferences. People living in poor villages are not necessarily risk averse but they are loss averse. They also have higher discount rates, suggesting they are less patient. These results imply economic circumstances are important in shaping people’s preferences. On the other hand, household income is not strongly related to preferences. Lower income is linked to impatience (higher discount rates) but is not correlated with risk preferences. By conducting experiments in multiple villages with various mean income levels, we were able to investigate whether mean village income (economic environments) or absolute income levels are related with wealth. Our contribution was to show how to expand measurement of risk and time preference beyond expected utility and exponential discounting models, by replacing them with prospect theory and quasi-hyperbolic discounting models with present bias. However, we could not link these preferences with economic activities and decision making in productive activities.

Using our experimental design, Elaine M. Liu (2013) examine whether risk preferences can explain the difference in adoption of agricultural technology among Chinese farmers. Liu shows the adoption of genetically modified Bt cotton is slower among risk averse and loss averse farmers. Also, the farmers who overweight small probabilities adopt genetically modified Bt cotton earlier. Elaine M. Liu and JiKun Huang (2013) further examine whether risk preferences explain overuse of pesticides among these farmers. They show risk averse farmers overuse pesticides, but loss averse farmers use less amounts of pesticides. They hypothesize loss averse farmers are more concerned about the impact of pesticides use on health. The two studies extended our study by using the experimental design we developed in our study and linking risk preferences with actual economic activities, i.e., agricultural technology adoption.