Risky rotten kids: an experiment on risk attitudes amongst adolescents in rural Uganda

According to some simple models of the household, parental transfers should equalize measured risk attitudes amongst family members. We explore the theory behind this notion and then use a Holt-Laury mechanism to compare attitudes to risk amongst 412 teenage children and their parents in 38 rural villages in central and eastern Uganda. Although within household risk preferences appear to be weakly correlated, we find large and systematic differences between parental and child attitudes that are not eliminated by controlling for loss-aversion and background variables such as age and education. Children are substantially more risk loving than their parents and conversely, slightly more loss averse than their parents. When we control for individual, household, village and regional factors the results are largely unchanged and in particular age is not an explanation of the differences in attitudes. We find that some aspects of risk and time preferences are linked to poor performance at school, including repeating a year and dropping out. Overall our results are compatible with a world in which parents protect children from risk. Moreover, because measured risk (and discounting attitudes differ between parents and their children, they suggest the value of measuring the attitudes of both generations when considering the relationship between preferences and schooling decisions.