Three essays in health econometrics

This thesis consists of three essays that address questions in health economics using different datasets and econometric approaches. In the first essay, I apply novel non-parametric econometric techniques to estimate the causal effect of retirement on health using the U.S. Health and Retirement Study (HRS) survey. I use a non-parametric Fuzzy Regression Discontinuity Design (RDD) technique for my analysis to avoid restrictive assumptions on a particular functional form and to capture the potential reverse causality from health to retirement (endogeneity issue) by exploiting the exogenous variation in retirement decisions induced by U.S. pension eligibility ages at 62 and 65. The results show that retirement is associated with an 8% decline in the cognitive functioning score of retirees, and 0.42 points increase in the CESD depression scale. Retirees also are 13 percentage points less likely to report good general health status, 8.8 percentage points less likely to be drinkers, and they are 4 percentage points less likely to consume alcohol more than three times per week. In the second essay, I use an administrative database to investigate the impact of the timing of first exposure to maternal depression on a comprehensive measure of children’s school readiness that incorporates multidimensional developmental domains that underlie school class adaptation and later success. I find that exposure to maternal depression is associated with developmental vulnerability in emotional, physical, social, and cognitive domains. The strongest adverse effects on development are from exposure to depression during pregnancy, followed by exposure during the preschool period. In the third essay, I examine the impact of Type I Diabetes Mellitus (TIDM) during childhood on educational attainment and labor market outcomes in adulthood using the National Health Interview Survey (NHIS). The results show that individuals who developed Type I diabetes early in life are 7 to 17 percentage poin