Trends in gender disparities at the transition from school to work: labour market entries of young men and women between 1984 and 2005 in West Germany

Abstract

This paper examines trends in school-to-work transitions of young men and women with lower and higher secondary education in West Germany between 1984 and 2005. This period was marked by an increase in young women’s educational attainment and a continuous growth of the service sector. We assume that both developments have benefited women more than men in terms of smooth labour market integration. Results from discrete event history analyses show that in recent years women indeed found their first job faster than men. However, this is not mainly due to an improvement in young women’s chances to enter employment, but to deteriorating employment prospects of men, in particular of unskilled men.

Keywords:

• gender
• labour market entry
• West Germany

Notes

1. Most young men in Germany enter military service (or alternative civilian service) for about a year after leaving the educational system for the first time. Before determining the school-to-work transition we deleted all months in service from the activity calendar, so that these were neither considered as part of the 12 months’ period mentioned above nor of any job search spell.

2. In the SOEP data, there is no information on the type of vocational education, i.e. if it was firm based or school based. Since the participation rates in vocational schools did not change systematically in the time period under view, this differentiation can be neglected in our analyses.

3. The vacancy ratio equals the number of vacant positions per unemployed worker.

4. In our sample, 8.6% of job search episodes are right censored, with women’s spells being censored more often than men’s.

5. Multiple observations per person are likely to lead to serial error correlation, so that robust standard errors have to be used in order to obtain valid significance tests.

6. The interpretation of interaction terms in non-linear models in the form of log-odds or odds ratios is associated with a number of problems (Ai and Norton 2003). To solve these problems we followed the statistical literature and calculated discrete change effects and absolute probabilities with their respective confidence intervals.

7. For details, see Model 1 in Table A1 in the Appendix.

8. For details, see Model 2 in Table A1 in the Appendix.