Sponsored by:
This study examines an assumption frequently invoked in economic models that individuals form preferences early, including with respect to risk, and those preferences remain stable. The challenge explores the idea that preferences might be shaped by situations individuals experience that cause them to reexamine and modify their beliefs. Risk preference is widely involved in the individual's decision-making process through influencing the willingness for risk-taking to get expected returns. The possible disclosure of this instability of risk preference may help both the individual and the social planner to acknowledge the possibility of manipulation of risk preference, and to begin to explore the method to maintain or modify the risk preference of the individual.
The study explores the endogenous risk preference with two stages. First, the study uses comparison of risk preference between twins to explore the possibility of unstable risk preference. If identical twins have statistically significant different risk preferences, and they depend on the rearing status, risk preference becomes a trait not solely dependent on genetic factor. Second, the study uses education as the proxy of experience and tests the hypothesis that risk preferences vary systematically (and causally) with differences in education. The study starts with the association between genetic factor and risk preference uncovered by Karlsson Linnér et al. (2019). To distinguish the genetic effect from the environmental effect of education, a mediation analysis is conducted. An instrumental variable design is adopted to address the endogeneity between risk preference and education, that risk preference may influence education decision.
The data come from the Health and Retirement Study (HRS), Wisconsin Longitudinal Study (WLS), and The National Longitudinal Study of Adolescent to Adult Health (Add Health). For twin study, both Add Health and Swedish Adoption/Twin Study on Aging (SATSA) are used. To measure risk preference, HRS and WLS use a set of risk aversion questions asking respondents to choose between different asset allocation with different expected values and probabilities. The study uses these sets of questions to measure risk preference. For Add Health and SATSA, one question is available asking respondents to subjectively report estimates of their risk preference, and the study adopts these measures. The education attainment measures include both the level of education and the year of education; for Add Health, more detailed educational characteristics are available and adopted. For the genetic data, the study uses the Single Nucleotide Polymorphisms (SNPs) collected in the three surveys, which are the types of genes on a specific position in the genome varying between different species and different individuals. In addition, apart from the respondents themselves, WLS and Add Health provide extra observations from siblings. The researcher includes these observations from siblings as the fixed effect in the analysis.