The expansion of early childcare and transitions to first and second birth in Germany

We use quasi-experimental expansion of publicly funded childcare slots for children under the age of three from Germany and exploit regional variations of this large-scale expansion to account for endogenous and selective fertility decisions. To account for left and right censoring, we implement this quasi-experimental framework into the setting of the semiparametric Cox hazard model. By using spatial data on childcare provision at the level of counties and microdata from the German Socio-Economic Panel (SOEP) from 1998 to 2012, we find a significant increase in the transition probability to first birth by 11.9% for native childless couples who were in the labor force before childbearing. With regard to transition to the second birth, however, no significant effect is found from the increase in childcare slots. With a particular focus on the transition to first birth, the effects are demonstrated not to be driven by selective residency choices and internal migration patterns. Furthermore, a large set of robustness checks is applied to show that highly educated mothers react the most, while effects are not attributable to the upper decile of income distribution.