Effect of education on health: The Czech Republic case

Previous research has uncovered a large, positive and causal link between education and health. This paper is devoted to examining the topic in the former Czechoslovakia. My analysis is conducted on a data set pooled from the Survey of Health, Ageing and Retirement in Europe (SHARE). I utilize a continuum of ages at school entry, caused by the use of a single school cut-off, to identify the effect of education on health, which is uniquely created from the PCA method and using 30 questions of the SHARE. Therefore, I apply instrumental variable approach with a month of birth as an instrument for education. The results from the first-stage suggest that the instrument is not valid, since a correlation between the instrumental (Month of birth) and the instrumented variable (education) is very low and insignificant. The results remain insignificant even after adjusting for different measures of education, health, institutional changes or heterogeneous effects. As the most probable cause, I state the inability to control for non-compliers in my instrumental variable regressions. As a consequence, all the results regarding the link between education and health are inconclusive. Powered by TCPDF (www.tcpdf.org