The transition from school to work : empirical evidence from Flemish technical and vocational secondary education
Mike Smet
Department of Business Studies, Lessius University College and Center for Economic Studies, KULeuven.
Corresponding address : Korte Nieuwstraat 33, BE-2000 Antwerpen, Belgium. Email :
In this paper we investigate the transition from school to work in Flanders, more specifically the transition of students from technical and vocational education. We were merged administrative databases from the Flemish Department of Education and the Flemish Employment agency in order to construct a longitudinal dataset, enabling to track individual students from secondary (to higher education and) to the labour market. We obtained data from the entire population of school leavers from ten similar study fields in technical and vocational secondary education (n=9991). The data include individual student characteristics, detailed individual enrolment information in secondary education and higher education for four school years, a number of school characteristics and monthly information on the individual employment or unemployment status. These data allow us explain differences in the duration of the first unemployment spell of school leavers and to quantify the impact of the various explanatory variables. Since some of the observed unemployment spells were censored, we estimated a Cox proportional hazards model to account for censoring. The results show that a number of individual characteristics as well as the socio-economic background of the school leavers are significant covariates in explaining the probability of leaving unemployment. In addition, the estimation results also show that some study fields guarantee faster access to the labour market than others.
Keywords: Transition from school to work, Cox regression.
Acknowledgements
This paper is based on the results from a research project which was financed by the Flemish Department of Education and Training (Bogaerts & Smet, 2008). The Flemish Department of Education Training and the Flemish Employment Agency kindly provided the necessary data for this study. However, they are not responsible for the analyses and conclusions presented in this article.
1 Introduction
Mainstream secondary education in Flanders (i.e. the Dutch speaking part of Belgium) is divided into four major education forms : general secondary education (GEN), technical secondary education (TEC), vocational secondary education (VOC) and secondary arts education (ART). The focus of this paper is on technical and vocational education. Technical education focuses on general and technical-theoretical subjects. After finishing technical education, students may either continue their educational career to higher education (college or university) or they can go to the labour market and exercise a profession. Approximately 30% of all secondary school students attend technical education. Vocational education is more practice-oriented and prepares pupils for a specific occupation. In general, these students are not allowed to higher education. Approximately 25% of all secondary students are enrolled in vocational education (Department of Education and Training, 2008).
Despite the distinction between technical (more oriented towards higher education) and vocational education (more labour market oriented) a number of similar study fields coexist both in the technical and the vocational form.
A first aim of this paper is to investigate whether students from similar study fields in technical and vocational education do have a different transition path from school to labour market. In addition we would like to quantify the impact of a number of student, school and regional characteristics on the unemployment duration after leaving school.
In the empirical analysis data from existing administrative databases were used. The student database from the Flemish Department of Education contains student and school characteristics. We were able to pair this database with the database of the Flemish Employment Agency (VDAB). By (anonymously) linking both databases we were able to reconstruct individual paths from secondary education to higher education and/or entry onto the labour market. The source of all data, tables and empirical analyses presented in this paper is the combined database of the Flemish Department of Education and the Flemish Employment Agency. Section two comprises an overview of recent empirical literature regarding the transition from school to work. In section three a brief description of the data is presented and in section four the results of the empirical analyses will be reported. Section five concludes this paper.
2 Literature
A number of studies investigated various determinants of the transition speed from education to work. Most studies find that school-leavers with a less favoured socio-economic background have longer unemployment spells and that study field and regional characteristics have a significant impact on the probability (and speed) of finding a first significant job. Evidence on the impact of gender is mixed.
Caspi, Moffitt, Entner Wright, & Silva (1998) find that individual characteristics and family background indicators at (early) childhood age have a significant impact on the prediction of future unemployment probability and unemployment duration after leaving school. Low intelligence, delinquent behaviour or other behavioural problems, the lack of a school certificate, poor reading skills, living in a single-parent family with a lack of school involvement, family problems and being male are significant predictors for an increased probability of unemployment and a longer duration of unemployment.
Bratberg & Nilsen (2000) estimate a model to determine the impact of human capital and family and individual background on the duration of the first unemployment spell of Norwegian school leavers. They find that females, married school leavers, apprentices and those with higher levels of education have shorter search periods while non-Scandinavians and school leavers with children have a longer first unemployment spell.
Eide & Showalter (2001) find that grade retention has negative impact on post-high school labour market performance, more specifically on earnings.
Lassibille, Navarro Gómez, Aguilar Ramos, & de la O Sánchez (2001) investigate the impact of human capital on the duration of unemployment and find that males and people with higher education have shorter unemployment spells. The socio-economic background of the family does not have a significant impact on the duration of unemployment.
A study by Andrews, Bradley, & Stott (2002) demonstrates that male and female school leavers with higher grades are more likely to rapidly find a matching job, while disadvantaged (i.e. school leavers with poor home background, involvement in criminal activities, etc.) males and females are less likely to obtain a matching job after leaving school.
Riphahn (2002) estimates a multinomial logit model in order to explain transitions after secondary education. She finds that being non-native (or having parents with foreign nationality) significantly increases the probability of becoming unemployed after leaving school. For females, this effect is even more pronounced. In addition, higher levels of schooling, belonging to a two parents household and not having siblings under the age of 16 lead to lower probabilities of unemployment. Finally, regional differences and the degree of urbanisation also have a significant impact on the transition rate from school to work.
A proportional hazard model that was estimated by Nielsen, Rosholm, Smith, & Husted (2003) indicated that higher educated students with high income parents are likely to find a job faster than others. They also report that second generation immigrants and females have longer durations of their first unemployment spells than ethnic Danes and males.
Cañada-Vicinay (2005) uses a parametric Weibull duration model to investigate the effect of family background on the transition from school into the labour market and to permanent employment and finds that school-leavers with vocational training and higher educated parents have a shorter duration of their unemployment period while females need more time to obtain permanent employment.
In a study by Semeijn, Boone, van der Velden, & van Witteloostuijn (2005) logistic regressions were estimated in order to quantify the impact of personality on the probability of having a job within 3 months after graduating from university. They find that gender, grade point average and final thesis result do not have a significant impact on the probability of finding a job, while personality characteristics and study field do have a significant impact.
Fernández (2006) estimates a hazard model and concludes that male school-leavers have a higher probability of finding a job and that socio-economic status of the parents is also important : the probability of leaving unemployment is higher for students whose father or mother is employed.
Brekke (2007) uses a Cox proportional hazard regression to study the impact of ethnic background on the speed of transition to the first regular job. She finds that graduates belonging to an ethnic minority group take longer to find their first job than native Norwegians.
The results of a proportional hazard model that was estimated by Salas-Velasco (2007) indicate that males and graduates with higher-educated fathers have shorter transitions from university to the labour market. In addition, job search effort, study field and regional differences also have a significant impact on the hazard rate.
Wolbers (2007) estimates a Cox regression and finds evidence for regional differences between 11 European countries in the transition speed from education to work. In addition the results show that higher educated and female school leavers have higher entry speeds into their first significant job.
3 Description of the data
In order to estimate similarities or differences between technical and vocation education in the transition from school to work, a number of similar study fields was selected. In order to obtain meaningful results, we only selected study fields that were offered in a sufficient number of schools and with a sufficient number of students. This resulted in 10 study fields, of which the entire population of 9991 students enrolled in the 6th year (i.e. the last year of secondary education) was included in this research. The majority was enrolled in vocational education (6311 pupils) while technical education accounted for 3680 students (more details can be found in Table 1).
Table 1. Selected study fields (last year secondary school, school year 2003-2004)
Technical education / Vocational educationGeneric name of selected study field / Schools (N.) / Students (N.) / Schools (N.) / Students (N.)
Car engineering (CARS) / 35 / 353 / 64 / 652
Construction (CONS) / 13 / 112 / 32 / 298
Office (OFFI) / 116 / 1373 / 158 / 2475
Woodwork (WOOD) / 43 / 376 / 123 / 1123
Fashion (FASH) / 17 / 124 / 57 / 244
Agriculture (AGRI) / 12 / 71 / 12 / 87
Horticulture (HORT) / 20 / 204 / 20 / 243
Mechanics (MECH) / 78 / 681 / 64 / 458
Bakery and pastry (BAKE) / 4 / 41 / 15 / 187
Hotel and catering (HOCA) / 23 / 345 / 27 / 544
Total number of students / 3680 / 6311
Source : Combined database Department of Education and Employment Agency
The databases of the Department of Education (database secondary education and database tertiary education) on the one hand and the database of the employment agency (VDAB) were linked using the NRN (National Registration Number, a unique identifier associated with every citizen living in Belgium). However, a small fraction (105 students, approximately 1% of the population included in this study) of students does not have a NRN (e.g. refugees). These students are entitled to go to school and thus can be tracked in all education databases, however due to the lack of NRN they cannot be paired to the VDAB-data.
The Department of Education data include information on students characteristics (e.g. gender, nationality, year of birth, grade retention, problematic non-attendance, grant, study results), detailed individual enrolment information in secondary education (SE) and higher education (HE) for four school years (starting from school year 2003-2004 and ending in school year 2006-2007), as well as a number of school characteristics (e.g. number of students, type of educational network, number of study fields, city). The VDAB-data include monthly information on the individual employment or unemployment status from July 2004 to April 2007 (unemployed, professional training during unemployment period, individual professional training during unemployment period or not-unemployed).
4 Empirical analysis
4.1 Profile of students and study fields
We will first describe some major differences between the characteristics of students in vocational and technical education. Vocational education has more non-Belgian students (6.9% in vocational versus 2.4% in technical education), more students with problematic non-attendance (approximately 1% in vocational versus 0.3% in technical education) and a higher share of students with grade retention (57% in vocational versus 45% in technical education). Moreover, the grade retention is more severe in vocational education.
These differences persist when we compare similar individual study fields in technical and vocation education. Regarding the different study fields, we can distinguish typical male study fields (car engineering, woodwork, construction and mechanics) from a typical female study field such as fashion. The other study fields (office, agriculture, horticulture, bakery & pastry and hotel & catering) have a more balanced gender distribution. Further we observe larger grade retention in car engineering and mechanics.
4.2 Career after last year of secondary education
After the school year 2003-2004, a variety of options emerges for the pupils : from continuing (secondary or higher) education to a transition to the labour market. Figure 1 depicts the possibilities in the school year 2004-2005. From the 9991 students we were originally following, 7248 (73%) are still studying during the next school year. The majority (55% of the original population) of them can be found in secondary education (SE), mainly in a 7th year (specialisation year). A smaller share (18%) is enrolled in higher education (HE), predominantly in colleges. About 27% of the original population does not study anymore in school year 2004-2005. The majority of them registers with the employment agency (VDAB) and can be considered unemployed for at least one month after leaving school. About 5% (510 students) of the original population is not enrolled in any educational institution nor registered with VDAB during the school year 2004-2005. Since school leavers have to register with VDAB in order to obtain unemployment benefit when they remain unemployed, there is a strong incentive to register. Therefore, it can be assumed that the vast majority of these 510 school leavers has found a job immediately after leaving school (other possibilities may be emigration abroad, admission in a psychiatric institution, internment, etc.).