Human Capital and fertility in Germany after 1990: Evidence from a Multi-Spell Model

Abstract

We analyze the timing of birth of the first three children based on German panel data (GSOEP) within a hazard rate framework. A random effects estimator is used to accommodate correlation across spells. We consider the role of human capital – approximated by a Mincer-type regression – and its gender-specific effects on postponement of parenthood and possible recuperation at higherorder births. An advantage of the use of panel data in this context consists in its prospective nature, so that determinants of fertility can be measured when at risk rather than ex-post, thus helping to reduce the risk of reverse causality. The analysis finds evidence for strong recuperation effects, i.e., women with greater human capital endowments follow, on average, a different birth history trajectory, but with negligible curtailment of completed fertility.