Family structure is usually believed to affect children's human capital. Is it possible that causality goes in the opposite direction? This paper shows that the behavior of family structure variables over the life cycle dramatically changes when women have babies in their forties. These data regularities align with a significant increase in the risk of having a child with health problems when women enter the last decade of their reproductive life. I present a simple theoretical model that provides a common underlying explanation for the data patterns and generates additional testable implications. I estimate the model predictions using ACS data.
child health family structure advanced maternal age