Do employment subsidies reduce early apprenticeship dropout?

Abstract

We evaluate the effect of the Apprenticeship Bonus, an employment subsidy programme, on early apprenticeship dropout. Eligibility for the programme is restricted to school leavers who have actively looked for an apprenticeship to start immediately after leaving school, but were unsuccessful in finding one. Our analysis is based on rich survey data that were collected specifically for this study. Using this data, we describe the characteristics of school leavers who looked unsuccessfully for an apprenticeship directly after leaving school and analyse the effect of the subsidy on the probability of apprenticeship dropout. Even though the subsidy provides strong incentives to discourage dropout, we do not find significant effects of the programme. Our findings suggest that financial incentives are not effective in increasing the probability of people finishing vocational in-firm training successfully.

Keywords:

• apprenticeship
• vocational training
• subsidised employment
• dropout

JEL classification:

• I21
• J08
• J38

Acknowledgements

This paper is based on the evaluation study ‘Effects of the Apprenticeship Bonus on the apprenticeship market and on public finance’, which was conducted on behalf of the German Federal Parliament. We gratefully acknowledge the provision of data by the German Federal Employment Agency and the Institute for Employment Research (IAB). We have benefited from discussions in the ZEW seminar, at the ESPE Annual Conference, and IZA workshop on Youth Unemployment and Labour Market Integration. We are also grateful to Bernd Fitzenberger, Thomas Zwick and three anonymous referees for helpful comments, and to Annette Hillerich and Markus Zimmermann for excellent research assistance.

Notes

1. Apprenticeship Bonus denotes the German programme Ausbildungsbonus.

2. Young people who cannot finish their apprenticeship because the firm goes bankrupt are eligible for the Apprenticeship Bonus, too. This case is not studied here.

3. See Franz and Soskice (1995) for an overview. Some other countries like Switzerland, Austria and Denmark have apprenticeship systems that are similar to Germany’s.

4. In 2009, 64.8% of German school leavers joined the apprenticeship system (BMBF 2010).

5. The duration of an apprenticeship depends on occupation and the previously obtained school qualification or vocational qualification of the apprentice. A higher school qualification reduces the duration of an apprenticeship by about one year. Formally, apprenticeship duration lasts between 18 and 42 months.

6. We distinguish three types of apprenticeships in the German apprenticeship system: a combination of schooling and in-firm training in the private sector, a combination of schooling and in-firm training in the public sector, and training that is provided full-time by vocational schools. Since the Apprenticeship Bonus is only provided for the private sector, we focus on the first scheme. To start an apprenticeship in a private-sector firm, people have to apply for an apprenticeship directly with the firm.

7. The Ausbildungsbonus is regulated by § 421r of the German Social Security Code III.

8. The Apprenticeship Bonus can also be provided for slow learners, the socially deprived, and apprentices whose firm has closed down. This study, however, focuses only on workers who have unsuccessfully looked for an apprenticeship after leaving school.

9. Since many firms fulfil this criterion anyway, it is apparent that it cannot ensure that the subsidised apprenticeship is created as a result of the Apprenticeship Bonus.

10. The subsidy amount is 4000 € if first-year monthly wage is not more than 500 €; 5000 € if first-year monthly wage is between 500 and 750 €; and 6000 € if first-year monthly wage is equal to or more than 750 €.

11. If the apprentice drops out, he or she cannot take the Apprenticeship Bonus to another firm. If a new firm is willing to hire him or her, the application procedure has to be gone through again.

12. Firms can pay higher wages by paying bonuses or extra hours; compare the argumentation and evidence in Mohrenweiser, Zwick, and Backes-Gellner (2013).

13. An alternative explanation for the fact that fewer subsidised apprentices mention low wages as a reason for dropout could be differences in wage expectations. It could be the case that subsidised apprentices have lower wage expectations than the group of unsubsidised apprentices.

14. In fact, empirical evidence regarding net costs of an apprenticeship is ambiguous. Based on a questionnaire, Wolter and Schweri (2002) report varying cost–benefit structures during an apprenticeship for Switzerland. Mohrenweiser and Zwick (2009) find evidence for net costs in German firms only in manufacturing occupations. Schönfeld et al. (2010) also report substantial net costs mainly in manufacturing occupations.

15. Another shortcoming of apprenticeship subsidies can be stigmatisation of the apprentices, making it harder to find employment after the end of the subsidised training period.

16. As general information on the start and duration of apprenticeships are also available in register data, they come with a time lag of several years.

17. The data were drawn from the population with respect to selected stratification variables: sex, migration status, firm size and region. All results in this study are weighted in order to obtain results for the underlying population.

18. Lower secondary education contains 9 or 10 years of schooling.

19. The German school grade system is represented by a scale reaching from 1 (very good) to 6 (unsatisfactory).

20. Vocational schools or Berufsfachschulen in Germany are typically full time, where pupils have the possibility to obtain vocational qualification.

21. See figures for basic vocational year, basic training, job preparing training, job application training, occupational orientation courses and temporary extra courses in Table 3.

22. This information was obtained in November 2009, i.e. about 14 months after the start of the apprenticeship.

23. Data on late dropout are not yet available.

24. The question is taken from the Programme for International Student Assessment (PISA) study; compare e.g. Q22 in the 2009 questionnaire: http://pisa2009.acer.edu.au/downloads/PISA09_Student_questionnaire.pdf.

25. The Epanechnikov kernel function is optimal with respect to minimising the asymptotic mean integrated squared error. The bandwidth is chosen by the rule of Härdle (1991). Robustness checks, which apply the doubled and quadrupled value of the chosen bandwidth, do not show different results.

26. Detailed results for each variable are available on request from the authors.

27. These interviews have been conducted for the evaluation study for the German Federal Ministry of Labour and Social Affairs (compare BMAS 2010) and are described in Appendix 1.

28. As the provision of an apprenticeship is of public interest, support is often provided by public facilities.

29. This robustness check is explained in detail in Becker and Caliendo (2007). For binary outcomes, the procedure builds on the Mantel and Haenszel (1959) statistic.