Women's labor supply reactions to divorce and childbirth in Germany

Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)




Jan Ondrich


Women workers, Working mothers, Maternity leave, Child care policies, Childbirth, Divorce

Subject Categories

Labor Economics


This dissertation examines female labor supply reactions to childbirth and marital separation. Since 1979 German federal maternity leave and benefit policy has, for some period after childbirth, protected the jobs of working mothers who stay home to take care of their newborns and youngest children, and since 1986 this leave and benefit policy was expanded in several ways. Chapter one takes advantage of this natural experiment to examine the effect of maternity leave and child care policies on mothers' returning to work after childbirth. A theoretical model is presented which demonstrates that hazard rates initially decline as protection is lengthened and that cumulative return probabilities at the end of protection can not decline unless the mother re-evaluates the parameters of the process. Using a flexible duration dependence estimation technique for proportional hazards due to Prentice and Gloeckler and applying grouped unemployment durations by Meyer (1990), we estimate post childbirth return to work hazards for women bearing children in Germany in the period 1984-1991. As potential duration increases, the hazard rates confirm the theoretical predictions, while the cumulative return rates suggest that some mothers re-evaluate the relative benefits of day care and mother care as the leave progresses.

Previous studies using American data have found a significant effect of divorce probability on the labor supply of married women. Chapter two of this dissertation examines effect of anticipated divorce on female labor supply using a German panel data set, seeking to both fill the gap in Germany, where similar empirical research on this issue is scarce and to enable a comparison between the United States and a European country with a historically low divorce rate. A duration model with controls for unobserved heterogeneity, rather than only cross-sectional snapshots is utilized to better understand the dynamic process of divorce and infer a probability of separation which is used as a regressor in the ordered probit estimation of labor force participation. Our results strongly support the model that women do anticipate a future separation probability and make labor supply adjustments in the current period.


Surface provides description only. Full text is available to ProQuest subscribers. Ask your Librarian for assistance.