We investigate an Educational Attainment Polygenic Score (EA PGS), an index that predicts years of formal education based on genetic data. In our analysis of the National Longitudinal Study of Adolescent to Adult Health we find that the EA PGS affects a number of health-related outcomes. Moreover, the EA PGS interacts with parental socioeconomic status (SES) in childhood: for a number of health outcomes we observe that the effect of the EA PGS is more beneficial for high-SES subjects. We decompose the total effects of the EA PGS into the indirect effect (through education) and the direct effect. We also decompose both the direct and the total effect with respect to potential mechanisms. The mechanisms that partially explain the effects of EA PGS include early skills, early health, education support in the family, and education. As a result of our discovery of a strong direct effect we cast our doubts on the validity of the EA PGS used as an instrumental variable for education affecting health, a case of an increasingly utilized technique called Mendelian Randomization. Finally, after controlling for the EA PGS, genetic health endowments, and unobserved heterogeneity in addition to more traditional controls, we still find that education is associated with better health outcomes, which adds evidence to the ongoing debate about the causal link between education and health.
First version, May 27, 2021
I12: Health Production
I14: Health and Inequality
I24: Education and Inequality
J24: Human Capital; Skills; Occupational Choice; Labor Productivity