Territorial differences in access to prenatal care and health at birth

We assess the impact of prenatal care on health at birth using birth certificates from the Czech Republic. We use a predictive machine learning algorithm to identify the observables affecting birth health outcomes. We control for those observables in our empirical analysis, which indicates that a more intense use of prenatal care is positively correlated with better health outcomes at birth. Exploiting the Czech adhesion to the EU in 2004, we construct an instrument to capture the geographical heterogeneous access to prenatal care across districts. Differently from the OLS results, the IV results do not capture any significant effect of prenatal care, leaving room for the hidden role of unobservable mothers' characteristics when it comes to health behaviors during pregnancy.