On the Development of Risk Preferences: Experimental Evidence

Abstract

Using a field experiment eliciting the risk preferences of 490 9th and 11th grade students from a variety of school environments, we examine various factors influencing the development of these risk preferences. In addition to factors previously considered by economists (gender, ethnicity, height, and parental education) we also evaluate cognitive (proxied by math literacy) and emotional (proxied by patience and planning ability) development theories. There is substantially stronger support for the variables typically considered by economists; consistent with prior work we find that girls are more risk averse while tall and nonwhites individuals are more risk tolerant. Next, the impact of school characteristics on the development of risk preferences is examined. Two effects emerge: a peer effect and a quality effect. For the peer effect, individuals who are in a school with a higher percentage of students on free or reduced lunches are significantly more risk averse than those with a lower percentage of students from low-income families. For the quality effect, individuals in schools that have smaller class sizes and a higher percentage of educators with advanced degrees have more moderate levels of risk aversion. Further, school characteristics only influence the 11th graders in the sample-lending some support to a causal argument.