50 Pages Posted: 14 Nov 2009 Last revised: 22 Nov 2009
Date Written: April 1, 2009
Educational attainment is strongly associated with better health outcomes. It is less clear whether this reflects a direct causal effect of schooling on health, reverse causality, or the influence of other, unobserved variables. In the past few years researchers have increasingly relied upon natural experiments to examine this question. This paper uses variations in state-specific compulsory schooling laws in the U.S. as instruments for education. I examine a wide variety of outcomes related to lifestyle diseases and general health status. These variables, together with detailed information on schooling and other demographics, were taken from the first two waves of the National Health and Nutrition Examination Survey (NHANES). While associations between schooling and health outcomes are found to be very strong in this sample, the instrumental variable results are inconclusive due to large standard errors in the presence of weak instruments. Using estimators and tests that are robust to weak instruments yields similar results.
Keywords: Compulsory school law, Health, Education, NHANES, Weak instrument
JEL Classification: I10, I12, I20, C01
Suggested Citation: Suggested Citation
By Thomas Dee