demographic changes, labor markets and their consequences on post-school-transitions in west germany...

19
Research in Social Stratification and Mobility 32 (2013) 65–83 Available online at www.sciencedirect.com Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005 Corinna Kleinert a,, Marita Jacob b a Institute for Employment Research (Institut für Arbeitsmarkt- und Berufsforschung, IAB), Regensburger Str. 104, D-90478 Nuremberg, Germany b Research Institute for Sociology and Social Psychology, University of Cologne, Greinstr. 2, D-50939 Koeln, Germany Received 14 February 2012; received in revised form 31 December 2012; accepted 28 January 2013 Available online 5 February 2013 Abstract The German system of post-school education and training is well-known for its high degree of standardization and stratification. Whereas transitions from this system to the labor market have been examined widely, much less is known about the transition patterns of school-leavers to post-school education. In particular, there is a shortage of research applying a comprehensive view of school-leavers from different tracks and investigating changing overall transition patterns over time. In our paper we fill this research gap by investigating school-leavers’ transitions to post-school education in West Germany over the last 30 years. Particular attention is given to the influence of contextual conditions, such as demographic fluctuations, economic cycles, and changes in the social composition of school-leavers. The empirical analyses are based on the retrospective life-course survey ALWA, which contains detailed information on the education and work trajectories of 10,000 adults in Germany. Our results show that school-leavers with low and medium attainment need more time to enter training when unemployment is high or when the school-leaver cohort is large. Overall, access to post-school education has become increasingly difficult for these groups, over and above the market shortages and the changes in group composition. By contrast, the transitions of school-leavers from higher secondary schools are unaffected by market fluctuations and long-term structural change. In the period looked at here, their chances remained almost stable. Accordingly, we conclude that the transition rates of school-leavers with different educational attainment have diverged in (West) Germany over the last 30 years. © 2013 International Sociological Association Research Committee 28 on Social Stratification and Mobility. Published by Elsevier Ltd. All rights reserved. Keywords: School-to-work transition; Post-school education; Long-term change 1. Introduction In Germany, as in many other countries, the transition from school to work has long lasting consequences for future employment careers. Due to the high degree of standardization and stratification in the German post- school educational system, analyses of school-to-work transition often divide the process into two stages: first, Corresponding author. E-mail addresses: [email protected] (C. Kleinert), [email protected] (M. Jacob). the transition from general school to vocational or higher education, and second, the transition to employment after completing a course of vocational or higher education. In this paper we concentrate on the first stage, i.e. the first transition after leaving general secondary school. On the German labor market, a standardized, acknowledged vocational certificate or a tertiary degree is regarded as a basic requirement for attaining a skilled posi- tion (Müller & Gangl, 2003; Shavit & Müller, 1998). Accordingly, the first hurdle on the way from school to post-school education is often more decisive in terms of future employment and career prospects than the 0276-5624/$ see front matter © 2013 International Sociological Association Research Committee 28 on Social Stratification and Mobility. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.rssm.2013.01.004

Upload: marita

Post on 12-Dec-2016

228 views

Category:

Documents


3 download

TRANSCRIPT

Page 1: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

A

WpsgicdlOtmwt©L

K

1

ffsst

m

0

Research in Social Stratification and Mobility 32 (2013) 65–83

Available online at www.sciencedirect.com

Demographic changes, labor markets and their consequences onpost-school-transitions in West Germany 1975–2005

Corinna Kleinert a,∗, Marita Jacob b

a Institute for Employment Research (Institut für Arbeitsmarkt- und Berufsforschung, IAB), Regensburger Str. 104, D-90478 Nuremberg, Germanyb Research Institute for Sociology and Social Psychology, University of Cologne, Greinstr. 2, D-50939 Koeln, Germany

Received 14 February 2012; received in revised form 31 December 2012; accepted 28 January 2013Available online 5 February 2013

bstract

The German system of post-school education and training is well-known for its high degree of standardization and stratification.hereas transitions from this system to the labor market have been examined widely, much less is known about the transition

atterns of school-leavers to post-school education. In particular, there is a shortage of research applying a comprehensive view ofchool-leavers from different tracks and investigating changing overall transition patterns over time. In our paper we fill this researchap by investigating school-leavers’ transitions to post-school education in West Germany over the last 30 years. Particular attentions given to the influence of contextual conditions, such as demographic fluctuations, economic cycles, and changes in the socialomposition of school-leavers. The empirical analyses are based on the retrospective life-course survey ALWA, which containsetailed information on the education and work trajectories of 10,000 adults in Germany. Our results show that school-leavers withow and medium attainment need more time to enter training when unemployment is high or when the school-leaver cohort is large.verall, access to post-school education has become increasingly difficult for these groups, over and above the market shortages and

he changes in group composition. By contrast, the transitions of school-leavers from higher secondary schools are unaffected byarket fluctuations and long-term structural change. In the period looked at here, their chances remained almost stable. Accordingly,

e conclude that the transition rates of school-leavers with different educational attainment have diverged in (West) Germany over

he last 30 years. 2013 International Sociological Association Research Committee 28 on Social Stratification and Mobility. Published by Elseviertd. All rights reserved.

rm cha

eywords: School-to-work transition; Post-school education; Long-te

. Introduction

In Germany, as in many other countries, the transitionrom school to work has long lasting consequences foruture employment careers. Due to the high degree of

tandardization and stratification in the German post-chool educational system, analyses of school-to-workransition often divide the process into two stages: first,

∗ Corresponding author.E-mail addresses: [email protected] (C. Kleinert),

[email protected] (M. Jacob).

276-5624/$ – see front matter © 2013 International Sociological Association Research Committee

http://dx.doi.org/10.1016/j.rssm.2013.01.004

nge

the transition from general school to vocational or highereducation, and second, the transition to employment aftercompleting a course of vocational or higher education.

In this paper we concentrate on the first stage, i.e. thefirst transition after leaving general secondary school. Onthe German labor market, a standardized, acknowledgedvocational certificate or a tertiary degree is regardedas a basic requirement for attaining a skilled posi-

tion (Müller & Gangl, 2003; Shavit & Müller, 1998).Accordingly, the first hurdle on the way from school topost-school education is often more decisive in termsof future employment and career prospects than the

28 on Social Stratification and Mobility. Published by Elsevier Ltd. All rights reserved.

Page 2: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

ial Strat

66 C. Kleinert, M. Jacob / Research in Soc

second one. This is aggravated by the fact that the high-est share of post-school education is provided by firms(Autorengruppe Bildungsberichterstattung, 2008, p. 89).Since more than half of all former apprentices are hiredby their training firms once they have completed train-ing (Autorengruppe Bildungsberichterstattung, 2008, p.180; Bellmann & Hartung, 2010), the second hurdle oflabor market entry in Germany is often fairly smooth forthese (Gangl, 2003; Pollmann-Schult & Mayer, 2004).Besides firm-based training, the post-school educationsystem also offers school-based vocational education infull-time vocational schools. Furthermore, when look-ing at the transition patterns displayed by a wholeschool-leaver cohort, one has to take into account alsoenrollment as another alternative for the continuation ofeducation after leaving school.

Although they differ in their formal prerequisitesfor access, these sections of the post-school systemare strongly interdependent, notably with regard tostructural contextual conditions. For example, if in aphase of economic recession there are not enoughapprenticeships available in firms, school-leavers mayswitch to school-based vocational education. Compe-tition for apprenticeships may become even keener ifsome school-leavers with the qualifications for tertiaryeducation do not enroll there but try to embark on voca-tional training instead because in a downturn it appears tobe the less risky alternative. Surprisingly, there are onlyfew empirical studies that comprehensively examine thetransition patterns of all school-leavers into all forms ofpost-school education over an extended period of time.In our paper we apply such a comprehensive perspec-tive and examine changes in post-school transitions.InGermany, students are separated very early in their lives(at about the age of 10) into three different secondary-school tracks, lower, medium, and higher secondaryschools. The criteria are academic performance, theirteachers’ recommendations, and their parents’ prefer-ences (e.g. KMK, 2011).1 These hierarchically stratifiedtracks award different school-leaving certificates: lowersecondary (Hauptschulabschluss), medium secondary(Mittlere Reife), and higher secondary diploma ([Fach-]Hochschulreife or Abitur). Permeability between tracks

is low, at least in the upward direction (Jacob& Tieben, 2009). The level of educational attain-ment has a major effect on pathways to post-school

1 There is some variation between federal states in the timing of thetransition from primary to secondary school, the number and types ofsecondary school tracks, and the availability of comprehensive schools(for details, see Section 3.1).

ification and Mobility 32 (2013) 65–83

education because some sections of it are open to partic-ular levels only and the three school-leaving certificatesoffer easily decodable signals for employers. Before1970, in former West Germany, the majority of a birthcohort attended lower secondary schools and afterwardseither entered the labor market or embarked on firm-based training. Only a small fraction of adolescents wentto higher secondary schools and studied at universities.From 1970 onwards, this picture changed. In the courseof educational expansion, medium secondary and highersecondary schools have grown in number, and nowadaysonly a minority of mostly socially disadvantaged stu-dents attend lower secondary schools (BIBB, 2009, p.130). Accordingly, our first descriptive research questionis whether and how the transition chances of school-leavers from different tracks have changed over time.

This change in educational attainment has beenaccompanied by changes in the social composition ofschool-leavers. In the course of educational expansion,school-leavers with higher attainment have becomemore heterogeneous in terms of social background,whereas students with low attainment have becomeincreasingly homogeneous (Solga & Wagner, 2007).Nowadays they often stem from disadvantaged familiesor immigrant backgrounds. At the same time, majorstructural changes have taken place in the labor mar-ket, including skill-biased technological change andsectoral change. This has resulted in more challengingrequirements for future workers and changing demandsfor occupations and qualifications. In the period inquestion, these long-term changes were overlaid bycyclically increasing unemployment rates and a varyingdegree of demographic pressure caused by the sizeof school-leavers’ cohorts. Taking these long-termdevelopments and short-term fluctuations together, thesecond question we ask is how structural developmentsand cyclical changes affected the transition patterns ofschool-leavers with differences in attainment.

In this paper we provide a long-term investigationof the transitions of different educational groups fromschool to post-school education. For this purpose weadopt a dual perspective for our analysis. On the onehand, we ask whether the chances for entering voca-tional training, vocational school or tertiary education

have changed between school-leavers with different edu-cational qualifications over the past 30 years in (West)Germany.2 On the other hand, we investigate changes

2 We disregard transitions in East Germany. Before 1990, the edu-cational system and access to vocational education and enrolment inthe GDR were not comparable to their counterparts in West Germany.

Page 3: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

ial Strat

otcFtgcdFscgi

snitflaeq

2

suelsgsto

oeMcohv

Ftatc

i

C. Kleinert, M. Jacob / Research in Soc

ver time within each school track to point up accesso post-school education for young people with identi-al school-leaving certificates at different points in time.rom these two perspectives, we pay particular attention

o both, the effects of compositional changes and demo-raphic and economic fluctuations. Our aim is to directlyompare transition chances between school-leavers fromifferent tracks despite the different options they have.urthermore, we aim to examine transition patterns ofchool leavers with the same educational level underhanging macro conditions over time. To this end, weroup the three sectors of post-school education togethern one single category.

A response to these questions calls for comprehen-ive longitudinal data enabling us to compare a largeumber of birth cohorts. The study ‘Working and Learn-ng in a Changing World’ conducted in 2007/2008 byhe Institute of Employment Research (IAB, Germany)ulfills these criteria.3 It contains detailed retrospectiveife-course data on the education and employment ofbout 10,000 persons born between 1956 and 1988. Inmpirical terms, we set about answering our researchuestions by using event history modeling.

. Previous research

Although there have been many empirical studies onchool-to-work transitions in Germany, their scope issually restricted in one way or another. Various studiesxamine transitions from post-school education to theabor market instead of looking at the paths taken bychool-leavers. Others only take into account particularroups of school-leavers; others focus on transitions topecific sectors of post-school education. Furthermore,he majority of the studies conducted so far only analyzene point in time.

A much smaller line of research is concerned withverall changes in transitions from school to post-schoolducation over time (Hillmert, 2002; Pollmann-Schult &ayer, 2004; Schneider & Tieben, 2011). These studies

ontain important empirical observations with respect to

ur research questions. The share of school-leavers fromigher secondary schools who choose non-academicocational training has increased, whereas the par-

or those who left school after 1990, the structurally different situa-ion of East German post-school education would have to be taken intoccount, but the sample size in our data is too small to provide consis-ent estimates broken down in terms of school-leaving certificates andohorts.3 In Germany, the study is known as ALWA (‘Arbeiten und Lernen

m Wandel’).

ification and Mobility 32 (2013) 65–83 67

ticipation of lower-qualified school graduates hasdecreased, at least until the mid 1990s (Pollmann-Schult& Mayer, 2004). At the same time, the transition periodfrom school to post-school education has lengthened(Hillmert, 2002). Transition patterns have becomelonger, more complex, and more heterogeneous (Jacob& Weiss, 2010; Jacob, 2004). A considerable share ofpresent day school-leavers enters post-school educationafter detours, e.g. vocational preparatory programsor acquisition of higher school-leaving certificates(Autorengruppe Bildungsberichterstattung, 2008, p.96f.). In particular, school-leavers with lower certifi-cates and grades (Solga, 2002), young people fromsocially disadvantaged families, and males with animmigration background (Aybek, 2011) encounterincreasing problems in entering post-school education.

Most of these results do not reveal causes of thechanges they describe, though. Thorough examinationof contextual effects is particularly conspicuous by itsabsence. In the discussion about the capacities of the dualsystem, causes are attributed mainly to market shortagescreated by economic cycles and demographic pressure.Yet there is little empirical evidence to bolster thisassumption. While some transition studies take a ratherrough approach by statistically controlling for labor mar-ket situation in addition to individual characteristics(e.g. Bender & Dietrich, 2001), only very few of themsystematically examine economic, demographic, andstructural changes jointly. In an aggregated regional anal-ysis, Heineck, Kleinert, and Vosseler (2011) show thatthe share of all school-leavers embarking on firm-basedapprenticeships in a given region displays a negative cor-relation with the regional unemployment rate and the sizeof the school-leaving cohort. Schweri and Müller (2007)and Muehlemann and Wolter (2011) yield similar resultsfor the Swiss training market. Hillmert (2001) investi-gates the effects of unemployment rate and cohort sizeon transitions to firm-based training in West Germanyfor the period 1976–1995. He shows that the transitionrate tends to decrease as unemployment increases, but hefinds no significant overall effect of cohort size. Youngpeople leaving higher secondary schools will more fre-quently enter vocational training if the employmentsituation is poor and cohort sizes are large. Graduatesfrom lower secondary schools are negatively affected byboth factors, i.e. fewer of them find training positions.

In sum, previous research leaves several questionsunanswered. First, there are far more studies on tran-

sitions from post-school education to work (e.g. onemployment entry, occupational status, income andunemployment risks) than on transitions from schoolto post-school education. Second, historical differences
Page 4: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

ial Strat

lege, school-leavers need to have successfully completedhigher secondary school.

68 C. Kleinert, M. Jacob / Research in Soc

in overall transition opportunities between and withingroups of school-leavers from the various tracks haveyet to be systematically analyzed. Third, while economicfluctuations do figure in some studies, connections haverarely been made with changes in access to post-schooleducation between and within groups. Also, the indi-cators used (e.g. Germany-wide unemployment rates)often either disregard regional differences or analyses ofmacro-factors are limited to aggregate data and do notinvestigate individual transitions. Accordingly, there isan absence of analyses of the interplay between structuralconditions and individual characteristics for individualtransitions. One main reason for this defect is the dataavailable up to now. The relevant datasets provided onlylimited scope for estimates on changes in individual tran-sitions and the effects of macro-factors on transitionpatterns over a number of decades.

3. From school to post-school education inGermany

3.1. The German educational system and itsinterdependencies

The German school system tracks students intodifferent types of school relatively early in theirlives. The most common ones are lower secondaryschools (Hauptschule), which operate with reducedcurricula and a slower pace of instruction, mediumsecondary schools (Realschule) with medium-level edu-cation, and higher secondary schools (Gymnasium)which take 2–3 years longer than the other twotypes and offer advanced academic instruction prepar-ing students for higher education.4 Besides differenttypes of school, the German school system awardshierarchically differentiated school-leaving certificates:

lower (Hauptschulabschluss), medium (Realschulab-schluss or Mittlere Reife), and upper secondary diploma(Hochschulreife or Abitur).5

4 There are several other tracks not present in all federal states,such as comprehensive schools (Gesamtschulen) which offer all lev-els of educational attainment under one roof, or schools combininglower and medium secondary level education (Mittelschulen). Varia-tion between federal states in terms of tracking age and school typeshas been increasing in the last years, but in the period we are lookingat here (1975–2005) such variation was considerably less evident thanit is today.

5 Despite the similarity in labeling, most of these diplomas can beachieved at more than one type of school, but only students who havesuccessfully got through a higher secondary school will be awardedthe Abitur, the entry ticket for university education. School-leaverswith medium secondary certificates can also continue their studies at

ification and Mobility 32 (2013) 65–83

After finishing school, adolescents can directly enterthe labor market.6 Most of them, however, decide tocontinue education because a vocational or higher edu-cation diploma is regarded as a minimum prerequisitein the German labor market. This diploma is ofteneven more important than the school attainment certifi-cate for labor market entry and the occupational andemployment career (Shavit & Müller, 1998; Solga &Konietzka, 1999). In addition, the option of embark-ing on a career in unskilled labor after leaving schoolis increasingly bound up with risks of long-term unem-ployment, precarious forms of employment, and lowwages (Giesecke & Heisig, 2011; Nickell & Bell,1995).

Roughly speaking, post-school education in Germanydivides into three sectors: dual vocational training, train-ing at full-time vocational schools, and higher education(Franz & Soskice, 1995; Soskice, 1994).7 The dual sys-tem of vocational training, which takes place parallelin firms and vocational schools, is the largest sector(BIBB, 2009, p. 95). In principle, it is open to all schoolleavers, but since training firms select from the numberof applicants, access depends on employers’ decisions.Hence, apprenticeships are offered and entered viamarket mechanisms (Schweri & Müller, 2007). In com-parison, admission to vocational schools is restricted byformal criteria, and most programs require a mediumgeneral school certificate. School-based training repre-sents only a small segment of post-school educationand is mainly geared to health-care, social and assistantoccupations (Autorengruppe Bildungsberichterstattung,2008, p. 96ff.). To be eligible to study at institutions ofhigher education such as a university or a technical col-

higher secondary vocational schools (Fachoberschulen). Successfulcompletion of these studies confers eligibility for higher educationat technical colleges or universities of applied science, as they callthemselves (Fachhochschulen). These institutions provide practicallyoriented programs such as engineering, business administration, orsocial work.

6 In our terminology we do not consider vocational training in thedual system and school-based vocational training as part of secondaryeducation. However, other authors use the term “upper secondaryschool” for both, general and vocational education after school grade9 or 10 (e.g. Schneider & Tieben, 2011).

7 During the past few decades, these sectors have increas-ingly been supplemented by vocational preparation measures,shorter programs that prepare students for regular training butdo not award acknowledged vocational certificates (AutorengruppeBildungsberichterstattung, 2008, p. 96f). Taken together, these pro-grams are often categorized as “transition system”.

Page 5: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

ial Stratification and Mobility 32 (2013) 65–83 69

ceAnattmmApepstvrswtspdctwt

3

cocvttatct

v52

obte

Quantitative Qualitative

Supply Size of school-

leave rs’ cohort

Composition of school-

leavers’ cohort (between and

within educational groups)

Demand Number of places and

positions in firms, schools,

and universities

Requirements for places and

positions in firms, schools,

and universities

C. Kleinert, M. Jacob / Research in Soc

Although these three sectors of post-school edu-ation have different structural characteristics andntry requirements, they are highly interdependent.t the micro-level, entering one sector voluntarily oron-voluntarily normally involves forgoing the otherlternatives, at least for the moment. For example, ifhere are more applicants for dual training than there areraining places offered by firms, unsuccessful candidatesay switch to vocational schools—and this behavioray then result in a scarcity of places in this sector.nother plausible strategy might be to wait for a vacantlace while improving one’s qualifications, e.g. in gen-ral upper secondary education.8 At the macro-level,olitical priorities for one sector may change the wholeystem. If, for example, policy-makers prompt firmso offer more dual training, the number of students atocational schools and universities may decrease as aesult. Reforms in tertiary education, such as offeringhorter programs, may attract non-traditional studentsho would otherwise have refrained from studying. This

rend may increase the opportunities for less qualifiedchool-leavers to enter vocational training. These exam-les are a neat illustration of the fact that in order toescribe and analyze the transitions of school-leaversomprehensively and meaningfully interpret changes ofransition patterns over time, any appropriate analysisill have to include all segments of post-school educa-

ion simultaneously.9

.2. Matching processes in post-school education

Entering post-school education is a two-sided pro-ess. On the one hand, school-leavers will go in searchf vacant places and positions, on the other hand edu-ation and training opportunities are offered by firms,ocational schools, and universities. Looking at transi-ions to post-school education from a macro-perspective,his results in a matching process of aggregate demandnd supply. Two dimensions of matching can be dis-

inguished (cf. Jacob, 2004, p. 26f.). On the one hand, aertain number of persons looking for post-school educa-ion (the supply side) have to be quantitatively allocated

8 The share of apprenticeship applicants who had left school in pre-ious years increased considerably from 36 percent in 1996 to nearly5 percent in 2007. Afterwards, it decreased again to 41 percent in012 (Statistik der Bundesagentur für Arbeit, own analyses).9 Alternatively we could describe patterns in one type of trainingnly—as several existing studies have done—but then we would note able to indicate and/or explain overall changes in transition pat-erns over time by comparing school-leavers with different levels ofducation.

Fig. 1. Factors influencing matching processes in post-school educa-tion.

to a certain number of open positions (the demand side).On the other, matching also takes place in qualitativeterms between the qualifications of the persons look-ing for post-school education and the skill requirementsof firms, schools, and universities offering places andpositions (Fig. 1).

The quantitative supply depends on the size of aschool-leaver cohort. The more adolescents leave schoolin a certain year the keener is the competition for train-ing positions in firms and places at vocational schoolsand universities. Whereas state-financed institutions usu-ally adapt the number of places they provide to expectedchanges in demography (or at least attempt to do so),employers do not. Accordingly, increased competitionwill be particularly pronounced in the sector of postschool education, which relies on firm-based training.

Besides adaptations to the demographic situation, thequantitative demand of training positions in vocationalschools and universities is determined by political objec-tives. State-financed institutions can only react slowly toshort-term market shortages, because financial means forthe necessary infrastructure cannot be provided in time.By contrast, firms regulate their number of apprentice-ship positions by business considerations (Troltsch &Walden, 2010). At the macro-level this means that theeconomic situation has a considerable impact on the tran-sition chances of a cohort. A quantitative shortage of postschool education opportunities due to economic fluctu-ations can be expected to have negative consequencesfor the transition chances of school-leavers, not onlyin connection with dual training but also in the othersectors (see Section 3.1). For both demographically andeconomically conditioned scenarios of increased com-petition, the same mechanism can be expected. If weassume that low-qualified school leavers are always lastin the queue for training positions and school places,this group will be more severely affected by keen com-petition than medium- or higher-educated school-leavers

(Thurow, 1979).

In qualitative terms, the supply side is determinedby the composition of a school-leaving cohort. First

Page 6: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

ial Stratification and Mobility 32 (2013) 65–83

Sources: Statis tik der Bundesagentur für A rbeit, Arbeitslos igkeit

im Zeitverlauf, Table 2.1.1; Fede ral Statis tical Office

(Destatis) , Fachserie 11 .1 (owncalcu latio ns)

0.0

2.0

4.0

6.0

8.0

10.0

12.0

0

200

400

600

800

1000

1200

in %

in 1

, 00

0

Unemploy ment rate

School lea vers

Fig. 2. Trends in demographic pressure and economic cycle, WestGermany 1975–2005.

place in formal access to school-based training and ter-tiary education in the period we are looking at.11 As forthe provision of training positions in firms that depends

10 Konietzka (1999) shows in his analyses of school-to-work transi-tions in Germany that the share of men and particularly women whoattended post-school education has increased over time, but only untilthe birth cohorts 1954–1956. These cohorts had entered post-schooleducation before the time span under view here. In the cohorts bornaround 1960 this trend came to halt.11 Regarding actual enrolment rates however, a remarkable expan-

sion took place (see also the remarks on changes in educational

70 C. Kleinert, M. Jacob / Research in Soc

and foremost, the distribution of educational attainmentin a certain cohort will influence the matching pro-cess. Since low-qualified school leavers are last in thequeue of opportunity seekers, their chances of enter-ing post-school education will decrease (or increase)with group size. This mechanism is particularly pro-nounced in Germany because the hierarchically orderedtypes of school-leaving certificate provide strong andreadily decodable signals for training firms, vocationalschools, and universities (Allmendinger, 1989). In addi-tion, changing educational attainment affects the socialcomposition of students within the different schooltracks because fluctuations of social groups betweenschool tracks are not random processes.

The qualitative demand is reflected in the means ofaccess to the different types of post-school education, i.e.in the qualifications that employers need and the entryrequirements stipulated by state-based educational insti-tutions. Vocational schools and universities manage theirinflow by formal requirements in terms of credentials andgrades. By contrast, employers can freely decide whichapplicants they want to hire for vocational training andthey are not obliged to disclose their decision criteria.Here, the type of school-leaving certificate and the finalgrades are often the most important criteria. So if the sup-ply of applicants and their educational composition staysconstant, upward changes in the qualification require-ments made by firms may result in decreasing transitionchances for low-qualified school-leavers because thereare no school-based alternatives for this group. Viceversa, downward changes in firm requirements mayresult in an increasing number of highly qualified school-leavers applying for universities and vocational schools.Besides education, ascriptive criteria such as social back-ground, ethnicity, or sex may play a role in employers’decisions and influence individual transition chances.This point is of particular relevance if there are changesin the social composition of students within single schooltracks, because a real or perceived accumulation of socialproblems within a single educational group may lead totransition disadvantages. In this case, the level of educa-tion is an indicator not only for formal qualifications butalso for the likelihood of social problems (Solga, 2002).

3.3. Long-term trends in supply and demand in WestGermany

How did the four factors that determine transitions

develop in West Germany in the period under review?As regards quantitative supply, the size of school-leavercohorts, we can identify a relatively long-term cyclicaldevelopment (Fig. 2). The number of school-leavers

Sources: Statistik der Bundesagentur für Arbeit, Arbeitslosigkeit imZeitverlauf, Table 2.1.1; Federal Statistical Office (Destatis), Fachserie11.1 (own calculations).

increased until the early 1980s, when the “babyboomers” born in the 1960s finished school. Subse-quently (up to the mid 1990s) it decreased significantlydue to the decline in birth rates triggered by the use ofthe contraceptive pill. Afterwards, school-leaver cohortsstarted increased again, albeit slowly. Besides, notonly the size of a school-leaver cohort matters but thesize of the cohort with ambitions to enter post-schooleducation.10

In terms of the quantitative demand, the number ofplaces in state-provided post-school education, tradi-tional tertiary education had been supplemented sincethe 1970s by the introduction of universities of appliedscience, polytechnics, and dual university programscombining in-firm training with academic education.Vocational schools expanded greatly during the 1970s;since then the increase in the number of students has beenslight. In general, however, no dramatic changes took

attainment of secondary school leavers below). The share of first-year students among the population in the same age increased fromless than 20 percent in 1980 to 37 percent in 2005 (AutorengruppeBildungsberichterstattung, 2008, p. 292).

Page 7: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

C. Kleinert, M. Jacob / Research in Social Strat

Source: Federal Statistical Offi ce (Destatis), Fachserie 11.1 (own calculations)

1913 11 8 8 8 8 8

49

4340

33 30 31 28 27

21

2932

3635 37 38 40

11 14 1623 27 25 25 25

1970 1975 1980 1985 1990 1995 2000 2005

Higher

secondary

Medium

secondary

Lower

secondary

No

certificate

Fig. 3. School-leavers’ educational attainment, West Germany1970–2005 (%).Source: Federal Statistical Office (Destatis), Fachserie 11.1 (own cal-c

orGwa

lltptacscbsiL

spmtSb&iw2oade

and universities have not greatly changed over time,changes in access to post-school education mainly cameabout in connection with the market-based matching

12 These changes are reflected in the dual training sector: the numberof training firms in manufacturing declined, whereas training pos-itions in clerical and service occupations increased (Miller Idriss,2002). Lower educated school leavers are increasingly concentratedin a narrow scope of training occupations (Uhly & Erbe, 2007).

ulations).

n the economic cycle, we look at the unemploymentate as an indicator for economic developments. For Westermany, this rate displays pronounced cyclical changesith peaks in the mid 1980s, the late 1990s, and 2005,

nd an increase after each recession (Fig. 2).Turning to the qualitative aspects of supply, Germany

ike many other developed countries went through aengthy period of educational expansion that started inhe 1960s and resulted in a significant change in the com-osition of school-leaver cohorts. Fig. 3 shows, however,hat in West Germany this trend more or less came to

halt in 1990. In later cohorts there was hardly anyhange in educational attainment. Educational expan-ion was accompanied by within-group changes in socialomposition. While students at higher secondary schoolsecame more socially heterogeneous, students at lowerecondary schools became more homogeneous, notablyn terms of disadvantaged social backgrounds (Breen,uijkx, Müller, & Pollak, 2009; Solga & Wagner, 2007).

Finally, changes in the qualitative aspects of postchool education (and labor market entry) were mostrobably affected by vast structural changes in laborarkets. The most important of these was skill-biased

echnological change (for surveys, see Acemoglu, 2002;anders & ter-Weel, 2000). Routine jobs that coulde taken over by machines disappeared (Autor, Levy,

Murnane, 2003), and jobs with transferable tasksn the production process were switched to countriesith lower wage levels (Grossmann & Rossi-Hansberg,008). These developments took place partly withinccupational fields (for Germany, see Spitz-Oener 2006)

nd partly they are reflected in sectoral change. In recentecades, we have observed continued growth and differ-ntiation in the service sector, whereas many occupations

ification and Mobility 32 (2013) 65–83 71

in manufacturing have gradually disappeared.12 Calcu-lation of the employment trend in different occupationalfields in West Germany shows up these developments(Fig. 4). The dwindling of jobs is to be found mainlyin occupations in the production sector, for example inmining, metal production and other production occu-pations, whereas the majority of thriving occupationsare related to the service sector, most pronounced in ITand computing, but also visible in health, social andeducation occupations. Most occupational fields withlow skill requirements are stagnating or in decline, forexample cleaners, hosts and hairdressers or jobs in trans-portation and logistics, while the jobs on the increasetypically call for high, often academic qualifications.In contrast to the stagnation in educational attainment,occupational changes were more pronounced from 1990to 2005 than from 1975 to 1989.13 In recent years, educa-tional expansion thus seems to have evolved more slowlythan structural labor market change in West Germany.

3.4. Changes in transition opportunities betweenand within educational groups

Taking into account our theoretical considerationsand the developments in demography, education, econ-omy and the labor market in West Germany, we can nowderive some hypotheses on trends in connection withtransition chances for school-leavers from the differentschool tracks. We start with the assumption that the tran-sition opportunities for low-qualified school leavers havedecreased over time, whereas those for highly qualifiedschool-leavers have remained constant or even increased.There is of course nothing surprising about this trend; itis a tendency that has been delineated in some detailby previous research. Our interest here is to explicateseveral reasons for this development that have not beenexamined systematically yet.

First, cyclical shortages in the training marketmay be part of the explanation for this trend. Sinceformal requirements for entering vocational schools

13 In order to compare the relative development of occupationalgroups in both time periods, the trends are divided by the overallemployment trend in the two periods.

Page 8: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

72 C. Kleinert, M. Jacob / Research in Social Stratification and Mobility 32 (2013) 65–83

% change i n 1989, b ase d on nu mbe rs of 1975( =0%), and % ch ange i n 2005, b ase d on nu mbe rs of 1990 (=0%),

corrected for changei n tota l empl oym ent.

Source: Sa mple of Integrated La bor M arket Biographies (SIAB), own calculations

8%

-28%

-26 %

-22 %

0%

7%

2%

-19 %

17%

7%

3%

-11 %

13%

97%

5%

16%

20%

70%

65%

63%

-12 %

-16 %

-69 %

-38 %

-37 %

-14 %

-24 %

-5%

-41 %

7%

3%

13%

-6%

30%

104%

12%

13%

23%

49%

46%

76%

-6%

Agriculture

Mining

Other produ ction occ .

Metal produ ction

Metal enginee ring and install ation

Elec tricity

Food

Con struction and wood

Tec hnicians and science occ .

Salespersons

Trade, banking and accou nting

Transpo rtation and logistics

Organisation and administration

IT and compu ting

Clerical occup ations

Order and sec urity

Media, hu manities and artists

Doctors and ph armacists

Other health occ upations

Social and education occ .

Cleaners, ho sts, hairdressers

1975-198 9

1990-200 5

Fig. 4. Changes in the number of employees in different occupational groups, 1975–1989 and 1990–2005. % change in 1989, based on numbers of, correc

calcula

1975(=0%), and % change in 2005, based on numbers of 1990 (=0%)Source: Sample of Integrated Labor Market Biographies (SIAB), own

process in the sector of firm-based training. As set outabove, in a market with increased competition thosewith the worst prerequisites have the greatest difficultiesin finding a training position, as they are in an inferiorposition over and against the more highly qualifiedapplicants. Additionally, they are formally excluded

from most of the school-based and all of the academictracks of post-school education. Thus we assume thatschool-leavers with low-level attainment certificates aregenerally more affected by demographic pressure and

ted for change in total employment.tions.

economic cycles than those with high-level certificates(H1a). Furthermore, we expect that the decline intransition opportunities for low-qualified school leaversfrom 1975 to 2005 can partly be explained by increasedcompetition between school-leavers due to cyclicalmacro developments (H1b).

Second, we assume that changes in the social com-position of students in the different tracks are partlyresponsible for the negative tendencies identifiable inconnection with low-qualified school-leavers entering

Page 9: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

ial Strat

paatsestf

cbdtrlth1tettksifivAotvvtaitldc

4

tGiR2tpG

higher education are all assembled in one category.If no such outcome is identifiable in the data, processtime is right-censored.16 In order to exclude outliers,

14 One should bear in mind that ALWA is retrospective data. First, thislimits the information that could be elicited in the survey. For example,there are no questions on the respondents’ educational decisions afterleaving school, so we are not able to identify risk sets of persons whointended to enter certain sectors of post-school education but did not doso. Second, episodes that took place a long time before, that were briefor unimportant, and that are negatively associated are prone to recallerror and thus may be underreported. However, we assume that bothleaving school and the first ‘real’ course of post-school education oneactually has entered are episodes that are too salient to be forgotten.15 As stated above, our paper aims at identifying global trends in

school-to-training transitions and not at examining detailed develop-ments in the different sectors of vocational education. Substantial andmethodological considerations prompted us not to define the threetypes of post-school education as competing events. First, we are inter-ested in comparing the transition rates of different educational groupsindependently of the kind of post-school education they embark on.Second, those models would estimate the transition rate of lower-and medium-educated school-leavers into higher education, whichin reality is non-existent. Third, the models require independence ofalternatives (IIA), whereas our theoretical assumption is that the sectorsare highly interrelated (see Section 3.1). Additionally, we estimated amultinomial logistic regression model distinguishing different typesof post-school education as sensitivity test. Results support these con-siderations: Low educated school leavers have only little chances toenter school-based training and virtually none to enter higher educa-tion. Medium educated school leavers have only little chances to enterhigher education. Higher educated school leavers have lower chancesthan medium educated ones to enter this type of training. This counter-intuitive result is explained by the fact that the risk set of persons who

C. Kleinert, M. Jacob / Research in Soc

ost-secondary education. Low-educated school-leaversre highly dependent on dual training, where social char-cteristics may play a role in the matching process. Soheir increased homogeneity may have caused their tran-ition opportunities to decline. By contrast, we do notxpect any changes for students with higher secondarychool degrees, because they can choose from a mul-itude of school-based and academic alternatives thatunction independently of market mechanisms (H2).

While these two explanations can be verified empiri-ally, there may be other causes for the growing divideetween educational groups that we cannot observeirectly. Queuing theory suggests that the decline inhe group size of low-qualified school leavers hasesulted in the displacement of low-educated school-eavers by more highly educated competitors in theransition to post-school education. This mechanism,owever, can only have been effective up to the mid980s, because afterwards educational expansion cameo a standstill. Since then, the macro data suggest thatmployers’ skill requirements have risen more quicklyhan school-leavers’ qualifications. The diverging timerends in educational expansion and structural labor mar-et change point to a qualitative imbalance betweenupply and demand, i.e. the qualifications of an increas-ng share of applicants may have satisfied neither therms’ requirements nor the formal prerequisites ofocational schools and institutions of higher education.gain, this may have resulted in decreasing transitionpportunities for low-educated school-leavers. Finally,he decline in group size probably reduced the signalingalue of a low-level school certificate beyond its actualalue. From this perspective, the perceived social charac-eristics of Hauptschule students—irrespective of theirctual composition—result in stigmatization, thus reduc-ng the likelihood of training opportunities. In sum,hese mechanisms suggest declining opportunities forow-educated school-leavers even when controlling foremography, the economic situation, and their socialomposition (H3).

. Data and methods

For our analysis of school to post-school educa-ion transitions we use data from the ALWA study (theerman acronym for ‘Working and Learning in a Chang-

ng World’) conducted by the Institute of Employmentesearch (IAB, Germany) in 2007/2008 (Antoni et al.,

010; Kleinert, Matthes, & Jacob, 2008). This represen-ative retrospective survey involved more than 10,000ersons born between 1956 and 1988 and living inermany. The data provide detailed information on the

ification and Mobility 32 (2013) 65–83 73

respondents’ life courses, e.g. schooling and educationalattainment, type of post-school education, if any (i.e.vocational preparation courses, firm-based and school-based training, and enrollment), military or civil service,episodes of employment, phases of unemployment andinactivity, and information about partners and children.14

From the raw ALWA life-course data we compiled ananalysis dataset depicting the transition from school topost-school education. Process time starts with the endof the first uninterrupted career of general schooling,i.e. between episodes of school attendance a maximumbreak of four months is allowed. The advantage of thisdefinition is that school-leavers really are at risk of start-ing post-school education, while this is not the case withconsecutive episodes of school attendance.

Process time ends with entry into the first post-school education episode. We only distinguish whetherpost-school education has been entered or not, i.e. tran-sitions into dual training, school-based training, and

15

intended to enter dual training is not defined correctly. Accordingly,tests show that the IIA assumption is violated.16 Theoretically, transitions to work might be an alternative to tran-

sitions to post-school education. However, transitions from school to

Page 10: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

ial Strat

74 C. Kleinert, M. Jacob / Research in Soc

i.e. rare cases of transitions happening very late, weobserve every person for only eight years after leavingschool.

Preparatory training programs such as one-year voca-tional schools were not counted as post-school educationbecause they do not lead to a recognized occupationalcertificate. These episodes are classified as search time,just like any other kind of activity taking place beforeinitial entry into regular post-school education. Onlyperiods of military and civil service during the transi-tion phase were deducted from process time, becausethey are institutionally predetermined and only malesare affected by them. Since our data do not allow fordistinction between periods of actual search and vol-untary non-entry periods, we included two variablescontrolling for the main activities conceivably produc-ing ‘lock-in effects’, i.e. phases where persons usuallydo not continue their search because they are insti-tutionally involved in full-time programs. These twotime-dependent covariates measure whether preparatoryprograms were completed in the transition phase orwhether individuals returned to school and achievedhigher qualifications there.17

Finally, several groups of persons were excludedfrom our analysis data: individuals who had not fin-ished school at the time of the interview (left-censoredcases), persons who had graduated from school abroador in Eastern Germany, and persons who completed theirschool education before 1975 or after 2005.18 After theseadjustments, we are left with the transitions of 6834persons and 62,976 person-months.

The central covariate in the models is the highestschool-leaving certificate achieved when leaving

general school and thus entering process time. Fourgroups are distinguished: school-leavers with no cer-tificate (dropouts), with lower secondary certificates

work are rare in our data: only 8 percent of the school leavers in our sam-ple enter a regular job before starting post-school education and halfof this group switches later on to post-school education. This patternsuggests that many of the observed employment spells have only transi-tional character. In a further sensitivity test, a multinomial competingrisk model of transitions to education and work our central resultsremain unchanged. Additionally, the model shows that transitions toemployment became more seldom over time.17 Several vocational preparatory programs offer the opportunity to

attain general schooling certificates. In this case, both dummy variableschange to 1 when such programs were finished successfully.18 This was done to avoid biased estimates. Before 1975, no transi-

tions to higher education are observed due to the birth cohort range ofthe ALWA population. For the young cohorts leaving school after 2005,the ALWA sample has selectivity problems with regard to educationalgroups.

ification and Mobility 32 (2013) 65–83

(Hauptschulabschluss), with medium secondary certifi-cates (Mittlere Reife), and with higher secondary cer-tificates (Abitur/[Fach-]Hochschulreife). An additionalvariable measures the self-rated school performance inGerman and mathematics in order to control within-track attainment. It contains ten values, 0 indicates anaverage rating, +2 a very good one and -2 a very badone.

Second, we use indicators for cyclical macro-trends.To measure the economic situation we use unemploy-ment rates for the whole dependent workforce in orderto avoid endogeneity problems.19 To indicate demo-graphic pressure, we use the share of 15- to 17-year-oldsin the population.20 Both variables are measured annu-ally for the federal state the respondent was living inwhen he/she left school. We have no direct measure forstructural labor market change.21 To test indirectly forlong-term developments, three aggregated cohorts weredistinguished: those who left school between 1975 and1984, between 1985 and 1994, and between 1995 and2005.

Finally, several variables accounting for the socialcomposition of the educational groups are included inthe models: parents’ education, migration background,growing up in an incomplete family, as well as respon-dent’s sex, age, and number of children. For parents’education, we decided to use the highest vocationalor resp. higher education degree, regarding the factthat vocational and higher education diploma are moreimportant in the German labor market than schoolcertificates and thus have a stronger impact on inter-generational transmission of stratification. Here, wedistinguish three levels of education: unskilled, voca-tional and higher education. To control for migrationbackground, we use a variable that indicates the pre-dominant usage of a foreign (non-German) language at

home at the age of 15. Family problems are measuredwith a variable distinguishing whether young peoplehave predominantly grown up with both parents or not.

19 Unemployment rates were obtained from the Federal Employ-ment Agency (Bundesagentur für Arbeit). We use unemployment ratesinstead of more direct indicators such as GDP because the former usu-ally lags behind GDP and is a far better measure for bottlenecks in thevocational training market.20 The data stem from the Federal Statistical Office (Destatis). We

preferred to use an indicator for birth rates rather than school-leavercohorts, because the latter may be endogenous. If the number of peersleaving school in the same year is high, students may decide to stay ata general school and postpone access to post-school education.21 The sectoral, qualificational, and occupational structure of employ-

ees never reflects the pure demand for skills, it is also dependent onthe supply of workers and the qualifications they have.

Page 11: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

ial Strat

TaTA

sbfWlppa&stmtigtiap

lwiro&st(haeT2

cfcoHm

oIdd

C. Kleinert, M. Jacob / Research in Soc

he number of children is monthly time-dependent, i.e. change in the number of children indicates a new birth.he distribution of all variables is shown in Table A1 inppendix.We investigate the transition from school to post-

chool education by using event history modelingecause it enables us to estimate the effects of dif-erent covariates on the transition rate simultaneously.

e decided to use a model for discrete time with aogistic link function (Rabe-Hesketh & Skrondal, 2008,. 331ff.),22 because it is particularly well suited torocesses that are not only measured discretely butre discrete in nature (Allison, 2010, p. 416; Singer

Willett, 2003, p. 426). Both the beginning (leavingchool) and the end (entering post-school education) ofhe process analyzed here usually take place in certainonths of the year, which results in a large number of

ies. We assume that the transition rate is relatively highn the first months after leaving school and decreases pro-ressively every year. Accordingly, we have collapsedhe transition durations measured on a monthly basisnto five dummy variables, four for the first four yearsnd one for the fifth to eighth year of the transitioneriod.

In general, unobserved heterogeneity can pose a prob-em in models with binary dependent variables, evenhen omitted variables are uncorrelated with variables

n the model. In this case, odds ratios are biased with theesidual variance and thus cannot be interpreted directlyr compared throughout subgroups or models (Auspurg

Hinz, 2011; Mood, 2010). Both types of compari-on, however, are central for our analyses. One solutiono this problem is to estimate average marginal effectsAME) because they are hardly biased by unobservedeterogeneity (Mood, 2010). χ2 tests of the respectiveverage marginal effects are used to examine whetherffects significantly differ between groups (cohorts in

able 1, educational groups in Table 2) (Auspurg & Hinz,011).23

22 Since dummy variables are included in order to control for dis-retely measured time, this model makes no assumption regarding theunctional form of the survival function. The model used here can beharacterized as semi-parametric because it is assumed that the effectsf the covariates are linear and additive on a logistic scale (Rabe-esketh & Skrondal, 2008: 342f.). It is estimated by conventionalaximum-likelihood estimation for binary dependent variables.

23 This solution requires that the groups compared do not affect eachther in their transition decisions. We think this assumption is plausible.nteraction models are no alternative because in models with binaryependent variables their results may be biased by different marginalistributions in the subgroups (Ai & Norton, 2003; Buis, 2011).

ification and Mobility 32 (2013) 65–83 75

A new decomposition method developed by Karlson,Holm, and Breen (2010) for binary non-linear models isused to test whether effects significantly differ in nestedmodels.24 Here, the total effect of a covariate of interestis decomposed into direct and indirect effects. The totaleffect is obtained from a reduced model without pos-sibly intervening variables. The direct effect is obtainedfrom the full model and can be interpreted as the share ofthe total effect not affected by intervening variables. Viceversa, the indirect effect is the share of the total effect thatis confounded by these variables. With the KHB methodit is also possible to test whether confounding is signifi-cant. We expect that transition chances of low-educatedschool leavers have systematically declined over time,which in technical terms corresponds to a negative cohorteffect. We assume further that this effect can be explainedin part by cyclical macro factors (H1b) and by a changein their social composition (H2). With the KHB method,we test whether such confounding of the cohort effect issupported by the data.

5. Results

Table 1 shows the results for the multivariate modelon the determinants of school-to-post-school educationtransitions separately for three school-leaving cohortsin order to compare the transition rates between dif-ferent groups of school-leavers in different periods oftime.25 Table 2 contains results of the same multivariatemodel, this time separately for three groups of school-leavers26 in order to compare the developments withinthe three educational groups over time. Aside from thecoefficients, the tables also show the results of statisticaltests of effect differences between groups (χ2 tests). Inthe following, the central results of both estimations arediscussed jointly.

We first turn to the findings on economic and demo-graphic fluctuations. As expected, the level of theregional unemployment rate has a negative effect on

the transition rate in Table 1. This is statistically sig-nificant only from the mid 1980s onwards. This resultis related to the development of overall unemployment

24 The KHB method is implemented as user-written program in Stata(see Kohler, Karlson, & Holm, 2011).25 Average marginal effects (AME) are shown in both tables. On

the average of all observations, they measure how many percentagepoints the probability of a transition will change when the regres-sor increases in one measurement unit. Negative values indicate aprobability decrease, positive values a probability increase.26 We did not estimate a model for school dropouts for reasons of

small sample size.

Page 12: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

76

C.

Kleinert,

M.

Jacob /

Research

in Social

Stratification and

Mobility

32 (2013)

65–83Table 1Transitions from school to post-school education by cohorts.

Effects Effect differences (χ2 tests)

Cohort 1 1975–1984 Cohort 2 1985–1994 Cohort 3 1995–2005 Cohort 2–1 Cohort 3–2 Cohort 3–1

Educational attainment (ref. upper secondary)No certificate −0.072*** (−9.66) −0.090*** (−10.23) −0.096*** (−9.30) 2.434 0.194 3.540+

Lower secondary −0.012* (−2.08) −0.017+ (−1.70) −0.053*** (−6.62) 0.187 8.452** 17.898***

Medium secondary −0.007 (−1.40) 0.010 (1.27) −0.024** (−3.20) 3.253+ 9.546** 3.679+

Macro factorsUnemployment ratea −0.001 (−0.95) −0.004** (−2.78) −0.002* (−2.04) 3.776+ 0.913 1.143Share of 15–17-year-oldsa −0.016** (−2.89) −0.009+ (−1.69) −0.023 (−1.53) 0.689 0.719 0.197

Social compositionAgeb −0.004+ (−1.72) −0.001 (−0.37) 0.000 (0.12) 0.405 0.125 1.155Female −0.027*** (−6.73) −0.007 (−1.00) −0.003 (−0.54) 7.285** 0.159 12.003***

Children −0.041** (−2.67) −0.070*** (−3.45) −0.085*** (−5.27) 1.320 0.348 3.994*

Parental education: unskilledc −0.000 (−0.03) −0.030* (−2.03) −0.004 (−0.40) 3.038+ 1.962 0.087Parental education: vocationalc −0.003 (−0.61) −0.001 (−0.17) 0.013* (2.08) 0.047 2.019 3.799+

Non-German language of origin −0.024* (−2.32) 0.019 (1.04) −0.028*** (−3.91) 4.230* 5.720* 0.084Incomplete family −0.001 (−0.12) −0.006 (−0.54) −0.023* (−2.57) 0.140 1.568 3.655+

ControlsSchool performance 0.016*** (5.47) 0.012* (2.45) 0.020*** (4.79) 0.686 1.677 0.475Additional certificate reached 0.272*** (6.54) 0.171** (2.72) 0.043 (1.03) 1.795 2.841 14.892Voc. prep. program finished 0.163*** (14.47) 0.219*** (10.43) 0.164*** (8.96) 5.438* 3.930 0.000

Duration (ref. 1–4 months)5–16 months −0.370*** (−38.43) −0.348*** (−28.39) −0.270*** (−24.79) 1.882 22.656*** 46.851***

17–28 months −0.423*** (−40.86) −0.397*** (−30.12) −0.315*** (−25.33) 2.392 20.572*** 44.674***

29–40 months −0.438*** (−42.12) −0.407*** (−30.47) −0.328*** (−26.15) 3.481+ 18.604*** 46.048***

41–52 months −0.443*** (−42.67) −0.419*** (−32.15) −0.333*** (−26.53) 2.049 22.725*** 45.740***

53–96 months −0.448*** (−43.74) −0.423*** (−33.13) −0.337*** (−27.27) 2.288 23.642*** 47.998***

AIC 12,498 7389 8594BIC 12,673 7549 8756N person-months 31,599 14,630 16,747N persons 3176 1789 1869

Source: ALWA 2007/2008, own calculations.Logistic regression models, average marginal effects (AME), robust standard errors, z-statistics in brackets.

a Cohort means-centered.b Educational groups means-centered.c Ref. academic.+ p < 0.10.* p < 0.05.

** p < 0.01.*** p < 0.001.

Page 13: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

C.

Kleinert,

M.

Jacob /

Research

in Social

Stratification and

Mobility

32 (2013)

65–83

77Table 2Transitions from school to post-school education by educational groups.

Effects Effect differences (χ2 tests)

Lower secondary Medium secondary Higher secondary Lower-medium Medium-higher Lower-higher

Cohort (ref. 1975–1984)1985–1994 −0.027+ (−1.94) 0.001 (0.15) −0.021 (−1.58) 2.732+ 1.800 0.1051995–2005 −0.059*** (−4.04) −0.027* (−2.37) −0.019 (−1.25) 3.011+ 0.152 3.447+

Macro factorsUnemployment ratea −0.002* (−2.12) −0.003*** (−3.34) −0.000 (−0.19) 0.143 4.201* 2.105Share of 15–17-year-oldsa −0.021** (−2.82) −0.014* (−2.55) −0.008 (−1.11) 0.590 0.457 1.632

Social compositionAgeb −0.003 (−1.05) −0.003 (−1.22) 0.002 (0.61) 0.000 1.323 1.214Female −0.038*** (−6.89) −0.028*** (−5.63) 0.014* (2.44) 1.528 30.480*** 41.978***

Children −0.044** (−2.99) −0.069*** (−3.49) −0.044* (−2.09) 0.977 0.774 0.001Parental education: unskilledc 0.002 (0.14) −0.005 (−0.47) 0.030 (1.60) 0.171 2.658 1.608Parental education: vocationalc 0.011 (0.96) 0.002 (0.28) 0.002 (0.26) 0.435 0.002 0.511Non-German language of origin −0.030*** (−4.05) −0.024** (−2.68) 0.007 (0.45) 0.285 3.121+ 4.923*

Incomplete family −0.008 (−1.07) −0.009 (−1.24) −0.009 (−0.78) 0.018 0.000 0.008Controls

School performance 0.019*** (4.43) 0.012** (3.02) 0.023*** (5.72) 1.302 3.765+ 0.565Additional certificate reached 0.099** (2.83) 0.098 (1.64) – 0.000 – –Voc. prep. program finished 0.192*** (17.85) 0.180*** (16.32) 0.148** (3.04) 0.619 0.413 0.787

Duration (ref. 1–4 months)5–16 months −0.418*** (−35.35) −0.396*** (−35.80) −0.285*** (−29.53) 1.859 57.024*** 75.870***

17–28 months −0.488*** (−39.41) −0.447*** (−38.61) −0.318*** (−31.01) 5.706* 69.746*** 111.496***

29–40 months −0.498*** (−39.72) −0.455*** (−38.79) −0.344*** (−33.10) 6.380* 50.133*** 89.736***

41–52 months −0.503*** (−40.21) −0.461*** (−39.18) −0.353*** (−34.43) 6.064* 47.543*** 85.769***

53–96 months −0.505*** (−40.62) −0.465*** (−40.09) −0.364*** (−37.96) 5.677* 44.453*** 80.347***

AIC 5213 9861 12,408BIC 5366 10,023 12,558N person-months 15,577 24,275 19,298N persons 1371 2574 2786

Source: ALWA 2007/2008, own calculations.Logistic regression models, average marginal effects (AME), robust standard errors, z-statistics in brackets.

a Cohort means-centered.b Educational groups means-centered.c Ref. academic.+ p < 0.10.* p < 0.05.

** p < 0.01.*** p < 0.001.

Page 14: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

ial Strat

78 C. Kleinert, M. Jacob / Research in Soc

in Germany. Before the mid 1980s, the unemploymentrate was probably too low to have any measurable impacton transition rates (see Fig. 2). Our second indicator forthe market situation at school-leaving time, the regionalcohort size, also has a negative effect on young people’schances of entering post-school education. This is partic-ularly true for school-leavers from the oldest cohort, the“baby boomers” born in the mid 1960s that left school inthe first time period. Looking at the χ2-test we see thatfluctuations over time are not significant in either of themacro-variables, and their negative signs do not changein the three cohorts.

In accordance with our first hypothesis on differencesbetween educational groups, the effects shown in Table 2indicate that only the two lower-educated groups are dis-advantaged by cyclical processes in the economy and bydemographic pressure. The χ2-tests reveal that only inone case there is a significant difference of effects, that isthe effect of unemployment between medium-educatedand high-educated school leavers.

Table 2 shows that school-leavers with low andmedium educational attainment have decreasing chancesto enter post school education: compared to the referencegroup of 1975–1984 those leaving school in later cohortsare less likely to embark on training. In Table 3, thiscohort effect is decomposed to find out whether it is con-founded by macro factors (H1b) or by school-leavers’social characteristics (H2). Here we test statisticallywhether the declining transition chances of low- andmedium-educated school-leavers can be partly explainedby changing impact of cyclical macro factors and socialcomposition.

In the first decomposition model shown in Table 3,a reduced model that does not control for the impactof economy and demography is compared with the fullmodel that was shown in Table 2. Contrary to our expec-tations, in the reduced model the (total) cohort effect issmaller than in the full model, and in the second cohort(1985–1994) it is not significant anymore. Accordingly,the indirect effect of cohort that is mediated by cycli-cal macro factors has a positive sign, and for the secondcohort it is significant at the 10 per cent-level. This meansthat the impact of the cyclical macro factors resultedin decreasing competition over time, whereas theoreti-cally we had expected increasing competition. In otherwords, the decline in access to post-school educationfor low- and medium-educated school-leavers cannotbe explained by growing competition due to unemploy-

ment and demographic pressure. On the contrary, if theunemployment rate and cohort size would have beenas high in later times as in 1975–1984, the transitionchances of low- and medium-educated school leavers

ification and Mobility 32 (2013) 65–83

would have deteriorated even further than they reallyhave. Hypothesis 1b is thus not supported by the empir-ical data.

In the second decomposition model shown in Table 3,a model including school leavers’ social characteristicsonly is compared with the full model shown in Table 2.The results indicate that for the two groups with declin-ing transition rates, low and medium educated schoolleavers, the differences between direct and total effectsare small and the indirect effects are not significant. How-ever, in this model the indirect effects show the expectedsign: in both cohorts they are negative. In substantialterms this means that a small part of the decline in accessto post-school education for low- and medium-educatedindividuals may be explained by growing social selectiv-ity over time, but statistically this effect is insignificant.Hypothesis 2 thus has to be rejected as well.

Finally, testing our third hypothesis we answerthe question whether differences in transition chancesbetween and within educational groups remain whentaking into account economic and demographic fluc-tuation and controlling for social composition. Goingback to Table 1, the estimation results show that transi-tion chances between groups actually diverge over time.Individuals without a school-leaving certificate had a sig-nificantly lower transition rate than the reference groupof graduates from higher secondary schools in all threecohorts. Accordingly, in the mid 1970s already, suc-cessful completion of secondary school was a basicprerequisite for entering post-school education. Up tothe mid 1990s, young people with lower secondarycertificates had slightly lower chances of entering post-school education than their peers with higher secondaryattainment. In the youngest school-leaving cohort, theirtransition rate dropped considerably. The two χ2-valuesfor the third cohort compared to the first and the secondcohort show that this difference is highly significant. Asimilar, but less pronounced, trend appears for youngpeople with medium educational attainment. In the firstand second cohort they do not differ from school leaverswith higher secondary certificates. Later, their chancesof entering post-school education are lower as well. Instatistical terms, though, the youngest school-leavingcohort only differs significantly from the previous one.In sum, the chances of access to post-school educationdiverge remarkably over time. Notably in the youngestcohort, those with lower and medium secondary schooleducation had considerably lower transition rates than

those with higher secondary education.

Looking at the effects within educational groups,Table 2 shows that the chances of low educated schoolleavers only dropped slightly from the mid 1980s to the

Page 15: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

C. Kleinert, M. Jacob / Research in Social Stratification and Mobility 32 (2013) 65–83 79

Table 3Decomposition of cohort effect in direct and indirect effects via macro factors and social composition.

Lower secondary Medium secondary Higher secondary

AME z AME z AME z

Indirect effects via cyclical macro factors1985–1994

Total (reduced model) −0.009 −1.16 0.010 1.63 −0.010 −1.34Direct (full model) −0.027+ −1.94 0.001 0.15 −0.021 −1.58Indirect (difference) 0.018+ 1.68 0.009 1.11 0.011 1.01Confounding in % −202.1 85.5 −115.0

1995–2005Total (reduced model) −0.034*** −5.65 −0.016** −2.99 −0.006 −0.88Direct (full model) −0.059*** −4.04 −0.027* −2.37 −0.019 −1.25Indirect (difference) 0.025 1.62 0.011 0.99 0.013 0.96Confounding in % −71.2 −65.8 −203.2

Indirect effects via social composition1985–1994

Total (reduced model) −0.034* −2.44 0.002 0.22 −0.021 −1.62Direct (full model) −0.027+ −1.94 0.001 0.15 −0.021 −1.58Indirect (difference) −0.007 −0.51 0.001 0.09 −0.001 −0.19Confounding in % 21.3 35.7 2.7

1995–2005Total (reduced model) −0.066*** −4.46 −0.033** −2.92 −0.021 −1.37Direct (full model) −0.059*** −4.04 −0.027* −2.37 −0.019 −1.25Indirect (difference) −0.007 −0.45 −0.006 −0.64 −0.002 −0.60Confounding in % 10.1 18.0 9.3

Source: ALWA 2007/2008, own calculations.KHB method for comparing the estimated coefficients between two nested non-linear probability models, using average marginal effects; controlledfor all the other variables in cohort model (see Table 2).

mcssccao

cttyssldwctwe

id 1990s in comparison to the reference group, theohort who left school from 1975 to 1984, while tran-ition rates of the youngest cohort are considerably andignificantly lower. Similar findings apply for adoles-ents with medium attainment. Here again, the youngestohort has significantly lower chances than the oldest. Soccess to post-school education became more difficultver time for both groups.

In sum, controlling for social characteristics andyclical macro-factors, we observe a trend of divergingransition rates of school-leavers with different educa-ional attainment in West Germany over the last 30ears that cannot be explained by increasingly negativeocial selection or by cyclical shortages. Chances forchool-leavers with lower secondary education (and to aesser degree of those with medium secondary education)ecreased over time. In our theoretical considerationse attributed this long-term development to qualitative

hanges on the demand side, i.e. upgrading of occupa-ions and changes in the economic sectors. By contrast,e observe no change over cohorts for the highest-

ducated group of school leavers. For them, access to

post-school education has remained constant during thelast 30 years. Accordingly, our third hypothesis is sup-ported by the data.

6. Summary and conclusion

In this article, we examined how the transition fromschool to post-school education has changed in Germanyfrom 1975 to 2005 by comparing school-leavers withdifferent educational attainment. Our analysis of allschool-leavers and all types of post-school educationenables us to depict changes in overall transition pat-terns over time. This comprehensive approach is justifiedby our assumption that transitions into the differentsectors of post-school education are not independentof each other. We analyzed two different aspects ofchange by first comparing transitions into post-school

education between school leavers from different tracksand then examining how access to post-school educa-tion has developed within these groups over time. Toidentify trends, we have taken into account structural
Page 16: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

ial Strat

80 C. Kleinert, M. Jacob / Research in Soc

parameters as well as school-leavers’ social composi-tion.

Comparing transitions between educational groupswe show that the transition chances diverged consider-ably over the last 30 years. Particularly in recent years,graduates with low- and medium-level attainment hadgreater difficulty entering post-school education thanthose with higher secondary-school attainment whereasstudents with higher schooling have not been affectedby the long-term structural changes in the labor marketover the three decades analyzed here. Cyclical macrofactors such as unemployment rate and relative birthcohort size actually affect the transition chances of stu-dents with low and medium educational attainment only.By contrast, the transitions of graduates from highersecondary schools do not vary with market fluctua-tions. Hence, those with low educational attainment aremostly affected by short-term bottleneck-situations afterschool due to their weak position in competition forfirm-based training and restricted access to schools anduniversities. For this group of school leavers, a long gapbetween school and post-school education might evenbe considered as additional negative signal by employ-ers, which in turn may deteriorate their chances evenmore.

Regarding changes within educational groups, ourresults show that the difficulties of those graduatingfrom lower secondary schools have increased remark-ably over time. From the mid 1990s onwards, we alsoobserve prolonged transitions in the group of school-leavers with medium education. This trend remainssignificant even when the economic and demographicsituation and group composition are controlled for. Con-trary to our expectations, these factors cannot explainthe declining transition opportunities of lower edu-cated school-leavers. Competition due to quantitativeshortages in the training market caused by high unem-ployment and demographic pressure even decreasedin the analyzed period of time and social composi-tion of low- and medium educated school leavers didnot contribute significantly to their declining transitionchances. The latter result is remarkable, because in theGerman public debate increasing socially problematicselectivity of students from lower secondary schools isassumed to be the main factor responsible for their tran-sition problems after having left school. In contrast, ourresults show that their transition chances to post-schooleducation decreased over and above social composi-

tion.

There are some limitations to our study. First, notall forms of institutional and structural change havebeen taken into account. During the last few decades,

ification and Mobility 32 (2013) 65–83

several reforms have been initiated in the post-schooleducation system, such as the expansion of prepara-tory training programs (Skrobanek, Reissig, & Müller,2011), reforms and changes within the dual sys-tem (Hoeckel & Schwartz, 2010), and expansion anddifferentiation in higher education (Reimer & Jacob,2011). Unlike the structural factors we have analyzed,these reforms are endogenous to the observed devel-opment of transitions over time, so we decided not toinclude any measures for these changes in our analy-ses. Their respective importance will need to be clarifiedin further analyses. In particular the introduction andexpansion of the preparatory training programs mightexplain partly the prolonged transitions in the youngestcohort leaving school after 1995. Nevertheless, beingmostly preparatory and/or providing general school cer-tificates, these programs cannot substitute for vocationaltraining in firms or vocational schools that we have beeninterested in our analyses.

Second, our article is limited to a description ofoverall trends in the transition from school to post-school education, while the preferences and decisionsof school-leavers and the recruitment strategies offirms providing apprenticeships could only be addressedmarginally. Since we observe diminishing chances andlonger transition periods for school graduates with low-level certificates, our results suggest that the value oflow and medium educational attainment has changed.By contrast, we find constant (but not improved) chancesfor graduates from higher secondary schools. Further-more, by looking at durations only, we cannot distinguishbetween ‘problematic’ unsuccessful search periods andvoluntary gaps between leaving school and enteringpost-school education. So part of the phenomenon ofprolonged transition phases we observe might be dueto an increase in voluntary delayed decisions aboutwhether, when and how to continue one’s educa-tion.

Summing up our results, major changes regarding thetransition between general schooling and post-schooleducation have occurred in the last 30 years in WestGermany. Lower educated school-leavers are particu-larly affected by long-term structural trends—they mightbe considered as ‘losers’ of the educational expan-sion and the profound changes in the labor market.We conclude from our results that transition problemsof these groups might even continue in times withless competition caused by demographic pressure and

unemployment—at least, unless the conditions and cir-cumstances in the German school system, in access topost school education, and in the labor market do notchange fundamentally.
Page 17: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

C. Kleinert, M. Jacob / Research in Social Stratification and Mobility 32 (2013) 65–83 81

Appendix.

Table A1Variables in the models (means and standard deviations).

Variable Person-months Persons (last observed month)

Mean SD Mean SD

Entering post-school education 0.10 0.30 0.95 0.22School leaver cohort

1975–1984 0.50 0.50 0.46 0.501985–1994 0.23 0.42 0.26 0.441995–2005 0.27 0.44 0.27 0.45

Educational attainmentDrop out with no degree 0.06 0.24 0.02 0.12Lower secondary certificate 0.25 0.43 0.20 0.40Medium secondary certificate 0.39 0.49 0.38 0.48Higher secondary certificate 0.31 0.46 0.41 0.49

Regional unemployment rate (%) 7.41 3.20 7.53 3.09Regional share of 15–17-year-olds (%) 4.21 0.89 4.14 0.89Age (in years) 17.79 1.72 18.00 1.74Female 0.57 0.50 0.50 0.50Children 0.07 0.25 0.02 0.13Highest parental education

Unskilled 0.11 0.31 0.07 0.26Vocational 0.74 0.44 0.74 0.44Academic 0.15 0.35 0.19 0.39

Non-German language of origin 0.10 0.30 0.06 0.23Incomplete family 0.11 0.32 0.09 0.28School performance (scale from −2 to +2) 0.29 0.67 0.43 0.67Additional certificate reached 0.02 0.14 0.01 0.12Vocational preparation program finished 0.16 0.37 0.11 0.31Transition duration

1–4 months 0.21 0.40 0.72 0.455–16 months 0.29 0.45 0.17 0.3717–28 months 0.13 0.34 0.04 0.2029–40 months 0.09 0.28 0.02 0.1341–52 months 0.07 0.26 0.01 0.0953–96 months 0.21 0.41 0.05 0.22

N 62976 6834

S

R

A

A

A

A

A

A

ource: ALWA 2007/2008, own calculations

eferences

cemoglu, D. (2002). Technological change, inequality, and the labourmarket. Journal of Economic Literature, XL(1), 7–72.

i, C., & Norton, E. C. (2003). Interaction terms in logit and probitmodels. Economics Letters, 80(1), 123–129.

llison, P. D. (2010). Survival analysis. In G. R. Hancock, & R. O.Mueller (Eds.), The reviewer’s guide to quantitative methods inthe social sciences (pp. 413–425). New York: Routledge.

llmendinger, J. (1989). Career mobility dynamics. Berlin: EditionSigma.

ntoni, M., Drasch, K., Kleinert, C., Matthes, B., Ruland, M., &Trahms, A. (2010). Working and learning in a changing world. Part

I: Overview of the study. FDZ Methodenreport 05/2010. Nuern-berg: Institute for Employment Research.

uspurg, K., & Hinz, T. (2011). Gruppenvergleiche bei Regres-sionen mit binären abhängigen Variablen—Probleme und

Fehleinschätzungen am Beispiel von Bildungschancen im Kohort-enverlauf. Zeitschrift für Soziologie, 40, 62–73.

Autor, D. H., Levy, F., & Murnane, R. J. (2003). The skill content ofrecent technological change: An empirical exploration. QuarterlyJournal of Economics, 118, 1279–1333.

Autorengruppe Bildungsberichterstattung. (2008). Bildung inDeutschland 2008. Bielefeld: W. Bertelsmann Verlag.

Aybek, C. (2011). Varying hurdles for low-skilled youth on the wayto the labour market. In M. Wingens, M. Windzio, H. de Valk,& C. Aybek (Eds.), A life-course perspective on migration andintegration (pp. 55–74). Heidelberg: Springer.

Bellmann, L., & Hartung, S. (2010). Übernahmemöglichkeiten imAusbildungsbetrieb. Eine Analyse mit dem IAB-Betriebspanel.

Sozialer Fortschritt, 59(6–7), 160–167.

Bender, S., & Dietrich, H. (2001). Unterschiedliche Startbedingungenhaben langfristige Folgen. Der Einmündungsverlauf der Geburt-skohorten 1964 und 1971 in Ausbildung und Beschäftigung—

Page 18: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

ial Strat

82 C. Kleinert, M. Jacob / Research in Soc

Befunde aus einem IAB-Projekt. IAB-Werkstattbericht 11/2001.Nuernberg: Institute for Employment Research.

BIBB. (2009). Datenreport zum Berufsbildungsbericht 2009. Infor-mationen und Analysen zur Entwicklung der beruflichen Bildung.Bonn: Bundesinstitut für Berufsbildung.

Breen, R., Luijkx, R., Müller, W., & Pollak, R. (2009). Non-persistent inequality in educational attainment: Evidence fromeight European countries. American Journal of Sociology, 114,1475–1521.

Buis, M. (2011). The consequences of unobserved heterogeneity ina sequential logit model. Research in Social Stratification andMobility, 29(3), 247–262.

Franz, W., & Soskice, D. (1995). The German apprenticeship system.In F. Buttler, W. Franz, R. Schettkat, & D. Soskice (Eds.), Institu-tional frameworks and labour market performance (pp. 185–206).London: Routledge.

Gangl, M. (2003). Returns to education in context: Individual educationand transition outcomes in European labor markets. In W. Muller,& M. Gangl (Eds.), Transitions from education to work in Europe.The integration of youth into EU labor markets (pp. 156–185).Oxford: Oxford University Press.

Giesecke, J., & Heisig, J. P. (2011). Destabilization and destandard-ization: For whom? The development of West Germanjob mobility since 1984. Schmollers Jahrbuch, 131(2),301–314.

Grossmann, G. M., & Rossi-Hansberg, E. (2008). Trading tasks: Asimple theory of offshoring. American Economic Review, 98(5),1978–1997.

Heineck, G., Kleinert, C., & Vosseler, A. (2011). RegionaleTypisierung: Was Ausbildungsmärkte vergleichbar macht. IAB-Kurzbericht 13/2011. Nürnberg: Institut für Arbeitsmarkt- undBerufsforschung.

Hillmert, S. (2001, April). Cohorts and competition. Transitionsfrom school to work in the context of economic and demo-graphic change. Paper prepared for the Conference MarketExpansion, Welfare State Retrenchment and their Impact onSocial Stratification organised by International SociologicalAssociation—Research Committee on Social Stratification (RC28)Mannheim.

Hillmert, S. (2002). Labour market integration and institutions: AnAnglo-German comparison. Work, Employment & Society, 19(4),675–701.

Hoeckel, K., & Schwartz, R. (2010). Learning for jobs: OECD reviewsof vocational education and training—Germany. OECD.

Jacob, M. (2004). Mehrfachausbildungen in Deutschland. Karriere,Collage, Kompensation? Wiesbaden: VS Verlag für Sozialwis-senschaften.

Jacob, M., & Tieben, N. (2009). Social selectivity of track mobilityin secondary schools. A comparison of intra-secondary transi-tions in Germany and the Netherlands. European Societies, 11(5),747–773.

Jacob, M., & Weiss, F. (2010). From higher education to work patternsof labor market entry in Germany and the US. Higher Education,60, 529–542.

Karlson, K. B., Holm, A., & Breen, R. (2010). Comparing regressioncoefficients between models using logit and probit: A new method(CSER Working Paper No. 0003). Aarhus: Centre for StrategicEducational Research, Aarhus University.

Kleinert, C., Matthes, B., & Jacob, M. (2008). Die Befragung ,Arbeitenund Lernen im Wandel‘. Theoretischer Hintergrund und Konzep-tion. IAB-Forschungsbericht 05/2008. Nuernberg: Institute forEmployment Research.

ification and Mobility 32 (2013) 65–83

KMK. (2011). The Education System in the Federal Republic ofGermany 2010/2011. A description of the responsibilities, struc-tures and developments in education policy for the exchangeof information in Europe. http://www.kmk.org/fileadmin/doc/Dokumentation/Bildungswesen en pdfs/secondary.pdf (accessed13.02.12).

Kohler, U., Karlson, K. B., & Holm, A. (2011). Comparing coefficientsof nested nonlinear probability models. Stata Journal, 11(3),420–438.

Konietzka, D. (1999). Ausbildung und Beruf. Die Geburtsjahrgänge1919–1961 auf dem Weg von der Schule in das Erwerbsleben.Opladen, Wiesbaden: Westdeutscher Verlag.

Mood, C. (2010). Logistic regression: Why we cannot do what we thinkwe can do, and what we can do about it. European SociologicalReview, 26, 67–82.

Muehlemann, S., & Wolter, C. (2011). Firm sponsored training andpoaching externalities in regional labor markets. Regional Scienceand Urban Economics, 41, 560–570.

Müller, W., & Gangl, M. (Eds.). (2003). Transitions from edu-cation to work in Europe. The integration of youth intoEU labor markets. (pp. 251–276). Oxford: Oxford UniversityPress.

Miller Idriss, C. (2002). Challenge and change in the German voca-tional system since 1990. Oxford Review of Education, 28(4),473–490.

Nickell, S., & Bell, B. (1995). The collapse in demand for the unskilledand unemployment across the OECD. Oxford Review of EconomicPolicy, 11(1), 40–62.

Pollmann-Schult, M., & Mayer, K. U. (2004). Return to skills: Voca-tional training in Germany 1935–2000. Yale Journal of Sociology,4, 73–99.

Rabe-Hesketh, S., & Skrondal, A. (2008). Multilevel and longitudi-nal modeling using Stata (2nd ed.). College Station, TX: StataPress.

Reimer, D., & Jacob, M. (2011). Differentiation in higher education andits consequences for social inequality. Higher Education, 61(3),223–227.

Sanders, M., & ter-Weel, B. (2000). Skill-biased technological change:Theoretical concepts, empirical problems and a survey of the evi-dence (MERIT. Discussion Paper 2000-012).

Schneider, S., & Tieben, N. (2011). A healthy sorting machine?Social inequality in the transition to upper secondaryeducation in Germany. Oxford Review of Education, 37,139–166.

Schweri, J., & Müller, B. (2007). Why has the share of training firmsdeclined in Switzerland? Zeitschrift für Arbeitsmarktforschung,40(2/3), 149–167.

Shavit, Y., & Müller, W. (Eds.). (1998). From school to work. Acomparative study of educational qualifications and occupationaldestinations. Oxford: Clarendon Press.

Singer, J. D., & Willett, J. B. (2003). Applied longitudinal data anal-ysis: Modeling change and event occurrence. New York: OxfordUniversity Press.

Skrobanek, J., Reissig, B., & Müller, M. (2011). Successful placementor displacement in the transition from school to vocational train-ing: The case of lower secondary school pupils. Journal of YouthStudies, 14(7), 811–836.

Solga, H. (2002). Stigmatization by negative selection: Explain-

ing less-educated persons’ decreasing employment opportunities.European Sociological Review, 18(2), 159–178.

Solga, H., & Konietzka, D. (1999). Occupational matching and socialstratification. Theoretical insights and empirical observations

Page 19: Demographic changes, labor markets and their consequences on post-school-transitions in West Germany 1975–2005

ial Strat

S

Sforschung, 43, 107–124.

C. Kleinert, M. Jacob / Research in Soc

taken from a German–German comparison. European SociologicalReview, 15(1), 25–47.

olga, H., & Wagner, S. (2007). Die Zurückgelassenen—diesoziale Verarmung der Lernumwelt von Hauptschülerinnen undHauptschülern. In R. Becker, & W. Lauterbach (Eds.), Bildungals Privileg (pp. 187–216). Wiesbaden: VS Verlag für Sozialwis-senschaften.

oskice, D. (1994). Reconciling markets and institutions: The Germanapprentice system. In L. M. Lynch (Ed.), Training and the pri-vate sector (pp. 25–60). Chicago and London: The University ofChicago Press.

ification and Mobility 32 (2013) 65–83 83

Thurow, L. C. (1979). A job competition model. In M. J. Piore(Ed.), Unemployment and inflation (pp. 17–32). New York:M. E. Sharpe.

Troltsch, K., & Walden, G. (2010). Beschäftigungsentwicklung undDynamik des betrieblichen Ausbildungsangebotes. Eine Analysefür den Zeitraum 1999 bis 2008. Zeitschrift für Arbeitsmarkt-

Uhly, A., & Erbe, J. (2007). Auszubildende mit Hauptschulabschluss:vom Normalfall zur Randgruppe? Berufsbildung in Wissenschaftund Praxis, 4/2007, 17–20.