We use miscarriage as a biological shock to fertility in order to estimate the causal impact of motherhood on labor market outcomes. The number of instruments is increased by exploiting the response-heterogeneity to miscarriage along three dimensions: time, age, and birth order. This allows us to separately identify the effect of the first, second and third child as well as the effects of pregnancy and caretaking for small children. We find each child reduces female earnings by around 18%, only part of it due to reduced work hours. We find no evidence of an adverse health effect of having children.
female labor supply health motherhood fertility shock