Under- or overexpansion of education? Trends in qualification mismatch in the United Kingdom and Germany, 1984-2017

Wiedner, Jonas Article — Accepted Manuscript (Postprint) Underor overexpansion of education? Trends in qualification mismatch in the United Kingdom and Germany, 1984-2017 Social Science Research Provided in Cooperation with: WZB Berlin Social Science Center Suggested Citation: Wiedner, Jonas (2024) : Underor overexpansion of education? Trends in qualification mismatch in the United Kingdom and Germany, 1984-2017, Social Science Research, ISSN 1096-0317, Elsevier, Amsterdam, Vol. 117, pp. 1-15, https://doi.org/10.1016/j.ssresearch.2023.102948 This Version is available at: https://hdl.handle.net/10419/313713 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. http://creativecommons.org/licenses/by-nc-nd/4.0/ 1 Underor overexpansion of education? Trends in qualification mismatch in the United Kingdom and Germany, 1984-2017 Jonas Wiedner WZB Berlin Social Science Research Center Abstract Prominent theoretical positions in sociology and labor economics disagree whether educational expansion has outstripped the demand for qualified labor (overexpansion), or whether economies face a skill shortage despite increases in education (underexpansion). Focusing on the United Kingdom and West Germany, two countries with dissimilar skill formation institutions, patterns of expansion, and labor markets, this paper asks to what degree expansion of education has been absorbed. I point out shortcomings of wage-centered analyses and develop an approach that focuses on trends in self-assessed overand underqualification. Using repeated surveys among workers and official labor market statistics, I estimate regression models that link the cohort-level expansion of education to the cohort-level prevalence of mismatch. Results suggest overexpansion in the United Kingdom, with overqualification increasing and underqualification decreasing over historical times and cohorts. West Germany, on the other hand, shows signs of underexpansion. While dominant theoretical accounts focus on the under-/overexpansion of tertiary education, my results show that mismatch-dynamics in both contexts are strongest for workers without university degrees. Keywords: educational expansion, mismatch, occupational change, comparative research This article was published by Elsevier in Social Science Research, Vol. 117 (2024), Art. 102948: https://doi.org/10.1016/j.ssresearch.2023.102948 © 2024 This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/ 2 1. Introduction Over the past century, completion of educational programs at all levels has increased dramatically. In 1953, just 19% of 17-year-old Britons attended full time schooling or training, but in the 2010s close to 90% did (House of Commons 2012). By 2018, half of 25to 34-year-olds had undergone tertiary programs (OECD 2019). In West Germany, 79% of 7th graders attended Volksschulen in 1952 and were thus bound to graduate by the age of 15. As a result, just 8% of a cohort enrolled in tertiary education in 1960. In 2010, by contrast, almost 40% did (Geißler 2014, p. 335 ff.). This dramatic expansion of education has been recognized by sociologists as one of the major forces shaping social change (Collins 1979; Baker 2014). Whether these growing numbers of higher-educated workers have been able to find adequate work is central to understanding stratification systems in modern societies. Yet, social science scholarship offers two contradictory answers to this question, each backed up by empirical literature. The first, mainly associated with sociological theories of social closure and credential inflation, argues that there are now too many higher qualified workers for labor markets to absorb in fitting work (Freeman 1976; Collins 1979; Wolf 2002). The second, rooted in variants of skill-biased technological change (SBTC) arguments in economics, argues that, given changes in skill demand, there are still too little (Goldin and Katz 2010; Acemoglu and Autor 2011; Carnevale and Rose 2014). I refer to these contrasting assessments as the diagnosis of overor underexpansion of education, respectively. An important implication of either overor underexpansion is that they increase the numbers of vertical education mismatches: in the case of underexpansion, workers whose education and training fall short of what is usual or nominally required in their occupation (underqualification), or, in the case of overexpansion, whose education and training are in excess of what is usual or required (overqualification) (Capsada-Munsech 2019). Trends in qualification mismatch are therefore indicative of the degree of labor market absorption of educational expansion. 1 1 I use the term “qualification“ to refer to certificates of graduation from education or training programs and to skill-certificates more broadly. I prefer the term “qualification” over “education” because it includes 3 Against this background and focusing on the United Kingdom and Germany, this article sets out to answer the following two questions: First, to what degree has expansion of education in these two societies been absorbed by labor markets? Is there evidence of overexpansion, with higher levels of overqualification, or do we see underexpansion, with increasing underqualification? Second, to which extent does this differ between the two countries? A focus on Germany and the United Kingdom, as, within the European context, very dissimilar cases, allows probing the extent of variation in different countries’ experience. The United Kingdom and Germany represent different varieties of capitalism with vastly different education systems (choicedriven progression in the United Kingdom, rigidly tracked in Germany), industry structures, labor market rules and patterns of expansion (strong expansion across all levels in the United Kingdom; a moderate expansion of tertiary education in Germany) (Müller and Gangl 2003; Hall and Soskice 2001). My analysis adds to an emerging literature that uses alternative measures to assess the labor market absorption of educational expansion directly. Evidence of underexpansion of education has traditionally been a byproduct of studies of increasing wage inequality, and therefore focused on wage-differences between education groups (Goldin and Katz 2010). Here, I develop a new approach which focuses on period and cohort trends in self-reported vertical mismatches between individuals’ education and their jobs, that is, on overand underqualification. I use time-series of mismatch-incidence based on repeated surveys (United Kingdom: Skills and Employment Survey), or a panel study (Germany: Socioeconomic Panel Study), respectively, as well as contextual information from official labor force surveys. My formal models use differences in education between cohorts in the same region and during the same time to estimate the relationship between a cohorts’ incidence of mismatch and its qualification structure. certified skills that have not been acquired in a formal educational setting, which have been central to training-policy debates particularly in the United Kingdom (Wolf 2002). My usage of the term “qualification” does, however, not include non-certified skills, abilities, and experience, such as general cognitive ability, physical strength, or work experience, which are sometimes subsumed under the term. 4 As an empirical contribution, I provide estimates of vertical mismatch-trends for the period 1986-2017 and for cohorts born between the 1920s and 1980s for the United Kingdom and Germany, and offer a novel, cohort-based perspective on the relationship between increasing educational levels and qualification mismatch (but cf. Horowitz 2018; and Pollmann-Schult and Mayer 2004). In contrast to most previous work (c.f., Shin and Bills 2021), my analyses include underqualification and systematically consider expansion at different qualification levels. I show that the relationship between educational expansion and qualification mismatch is highly contingent on institutional regimes. Educational expansion went hand in hand with increasing overand falling underqualification in the United Kingdom, but not in Germany. This finding supports recent arguments that global forces of technological and social change have very different effects in different institutional contexts (Oesch and Rodriguez Menes 2011; Fernández-Macías and Hurley 2017). Finally, my results document that changes in mismatch-rates with educational expansion are most pronounced for workers with qualifications below the university-level. In the next section, I review the two most important theories of the changing labor market balance of education and demand for it, skill-biased technological change and credential inflation, and examine the role of institutional context. I then introduce the data and the analytical approach I use. In section 4, I present descriptive and multivariate results before I summarize my findings and draw conclusions. 2. Theoretical background and previous research Perspective 1: technological change results in an underexpansion of education The dominant view on the changing balance between educational expansion and demand for it in economics is that of an increasing skill shortage (at least at stable prices, i.e. wages). While SBTC theory’s main concern is with increasing wage-inequality, the mechanism it assumes is that of an underexpansion of education. SBTC theory therefore has implications for 5 mismatch-trends. SBTC claims are rooted in the idea that technological innovations complement the productivity of workers with higher levels of education, but often substitute for less educated workers (Tinbergen 1956; Goldin and Katz 2010; Acemoglu and Autor 2012). Technological progress thus increases the relative demand for skilled vs. unskilled workers. Education's value on the labor market therefore depends on a “race between education and technology”: If the rate of technological innovation outpaces the expansion of education, the economy’s skill demand will outstrip supply, increasing highly educated relative to less educated wages (Goldin and Katz 2010). From an SBTC perspective, increasing wage inequality thus reflects a shortage of skilled labor, that is underexpansion of education (Carnevale and Rose 2014). A newer refinement of SBTC argues that the kinds of tasks that are being replaced by automation are not low-skill per se, but routine in nature. These, its proponents argue, predominate not in low, but in middle-skill occupations (Autor et al. 2003; Autor and Dorn 2013). The consequence of such routine-biased technological change (RBTC) would thus not be an upgrading, but a polarization of skill-demand. While SBTC’s and RBTC’s implications for educational expansion absorption coincide in the upper half of the education distribution, RBTC would expect not an underbut an overexpansion of middling certificates — assuming that educational upgrading is linear. Wage-based studies of underexpansion Work by economists on the United States documents dramatically increasing wage inequality between, but also within qualification groups, which is linked to the stagnation of college completion rates since the 1970s (Katz and Murphy 1992; Goldin and Katz 2010). Variants of the skill shortage thesis of SBTC are hence widely accepted among US economists (but cf. Card and DiNardo 2002; Acemoglu and Autor 2011). 2 In Germany, increases in wage inequality 2 It should be noted, however, that dissident labor market analysts have challenged the uncomplicated applicability of the polarization-thesis to the US labor market (Holzer and Lerman 2007, 2009; Holzer 2015). Drawing on disaggregated occupation-level wage and employment trends, they show that the supposed wide decline of middle-skill work is all but uniformly evident. Rather than expecting a secular decline of middle-skill work, they argue that institutional shortcomings in worker training are in part responsible for observed declines in some sub-segments of the skill-middle. As a consequence, and in 6 have been more moderate, which might reflect the substantial expansion of the education system since the 1950s. Nevertheless, studies conclude that SBTC, and hence a relative underexpansion of education, had a part in increasing wage inequality in Germany (Dustmann et al. 2009). In the United Kingdom, educational expansion was even more pronounced. At the same time, however, inequality between education groups – and wage inequality overall – have risen dramatically, suggesting that institutional changes may have played a dominant role there (Gosling and Lemieux 2004). Studies inferring changes in the ratio of demand of higher educated workers to its supply from wage data face criticism, however (Kalleberg 2011; Cappelli 2015). In order to provide evidence in favor of SBTC, they assume that observable wage changes are due to shifting demand or supply, and not to other factors. But much research demonstrates that factors such as changing minimum wages, de-unionization and other forms of rent-destruction and -creation have contributed significantly to increased wage inequality (for examples, see Kristal and Cohen 2017; Weeden and Grusky 2014; Fitzenberger et al. 2013; Gosling and Lemieux 2004). At least in simple wage analyses, these forces act as omitted variables and bias estimates of excess demand for education. The SBTC conclusion that wage inequality rose because of a shortage of highly educated labor may therefore be premature. Vice versa, this also puts important evidence in favor of the underexpansion thesis into question. Perspective 2: Credential inflation, social closure and relative education result in an overexpansion of education An important assumption of both variants of the technological change argument, SBTC and RBTC, is that skill and education are to a large extent synonymous. SBTC and RBTC are based on the human capital theory of education, which holds that education and qualifications are valued by employers and students because they create and certify relevant skills, which contrast to polarization theorists, they recommend policies to encourages skill-formation and better matching among middle-skill workers (Carnevale and Desrochers 2002). 7 then translate into high productivity and earnings. By contrast, many sociological approaches highlight that education’s role for students and society goes beyond technical know-how: formal education and productive skills are not the same thing. While these theories differ in the causal mechanisms they focus on, they agree that an increasing value attached to education for non-technical reasons has resulted in its expansion beyond labor markets’ absorption capacity. Prominent theories around the notion of credential inflation, for instance, argue that schooling is important to policy makers, parents, students and employers because it instils obedient work attitudes (Bowles and Gintis 1977), legitimizes social stratification (Bourdieu and Passeron 1970), equips students with cultural currency that gives access to closed occupational positions (Collins 1979), or is a means of self-realization and a symbol of family prestige (Baker 2014). Such ideological, political, or cultural reasons for expansion bear little connection to technological requirements. Expansion for these non-technical reasons is therefore likely to exceed employers’ demand for more highly qualified workers and result in overexpansion. Other social scientists point to micro-level processes which lead rational actors to acquire education beyond labor market demand and result in overexpansion. The central argument in such analyses is that education has a relative rather than an absolute value. According to the labor queue model, education is a signal of future productivity or trainability and employers rank applicants according to their level of qualification (Thurow 1975). The value of a given person’s qualification therefore depends on the levels of education of others in the queue. This mechanism is described as the relative or positional education hypothesis (Horowitz 2018). If education is positional, educational expansion translates into a heightened individual incentive to obtain yet more education in order to preserve one’s relative position (Bol 2015). Since education essentially becomes a race for the highest degree, overqualification, relative to jobs’ actual skill requirement, will be very prevalent as companies’ hiring standards escalate. To sum up: If forces other than technology-induced demand are behind increased qualification uptake, as argued by many sociological approaches, there will be overexpansion. In this 8 situation, holders of higher degrees will find it increasingly hard to find fitting work. Expansion will lead to higher rates of overqualification. Mismatch and the absorption of educational expansion As reviewed above, the evidence in favor of the underexpansion thesis has traditionally been a byproduct of studies of increasing wage inequality, and faces methodological concerns. In line with this criticism, studies using other approaches overwhelmingly find support for the overexpansion thesis. The principal alternative strategy focuses on mismatches between workers’ qualifications and the requirements of their jobs. Rates of (vertical) qualification mismatch can serve as an immediate measure of the degree of absorption of educational expansion by labor markets. 3 In contrast to inference based on relative wage-trends, mismatches offer a direct way to gauge the absorption of someone’s education on the labor market. The main advantage of mismatch-trend-based assessments of overand underexpansion is that unobserved processes (like changing minimum wages, de-unionization, or increasingly oligopsonic labor markets) do not directly affect the measurement of a qualification imbalance on the labor market in the way they do in relative-wage-based strategies. Even though mismatch-measurement may introduce ambiguities of its own (see Appendix B in the Online Supplement), substantive conclusions in a debate as central to policy and social science as the under-/overexpansion question should not rest on a single methodological approach. From a mismatch-perspective, rates of overqualification should decrease under SBTC, as employers seek to exploit the capacities of the relatively scarce highly qualified workers. At the same time, rates of underqualification should increase, as firms hire workers for positions above their nominal level of education to avoid interruptions of their business. Credential inflation, social closure and relative education approaches would expect the inverse pattern: rising overqualification and sinking underqualification over historical time, and in cohorts subject to 3 According to standard understandings in the literature, mismatch can be horizontal (between field of study/training and job) or vertical (between skill levels of training and skill level of employment) (Di Stasio 2017). While educational expansion has implications for the horizontal matching between workers and jobs (Elbers et al. 2021), underand overexpansion are most clearly expressed by rates of vertical mismatch that is by overand underqualification. 15 heart of the overexpansion debate. In line with my theoretical reasoning, I interpret a positive relationship as evidence of overexpansion. Independent variables My main independent variables are indicators of the contextual share of people of a given qualification level at a given point in time, in a given region, in a given extended cohort (defined as people born within +- 3 years from the cohort base year). I measure them using official labor market surveys. At this extended-cohort-region-period-level, I distinguish six categories (below secondary education, any non-minimal secondary education qualification, lower vocational qualifications, vocational qualifications including trade apprenticeships, lower tertiary qualifications, and tertiary qualifications) in the United Kingdom, and in Germany three categories (secondary or lower, vocational qualifications, and tertiary qualifications). This analytical choice reflects the consistently distinguishable levels of education in the different rounds of the labor force surveys I use. Control variables I adjust my estimates for individuals’ own highest qualification, which rules out that composition effects influence results. Based on the available individual-level survey data, I distinguish six (United Kingdom) and five (Germany) groups, according to a respondent’s highest qualification: • no qualifications, • (non-minimal) secondary qualifications (these two are combined in Germany), • two categories of vocational qualifications o in the United Kingdom: ▪ lower vocational qualifications from short programs, i.e. level 1 or 2 in the NVQ classification, and ▪ more advanced vocational qualifications (NVQ level 3), e.g. apprenticeships or SCOTEC/SCOTBEC qualifications; o in Germany: ▪ workers with vocational training, and ▪ workers with higher vocational training, i.e. Meister and Techniker, • lower tertiary certificates 16 o United Kingdom: NVQ level 4, e.g. university certificates or nursing degrees; o Germany: universities of applied sciences (Fachhochschule), and • university graduates. I limit the comparison to cohort variation by including survey year-region fixed-effects into my model. Such fixed effects rid my estimates of any period and region variation. This strategy therefore holds changes on the demand side constant (if they do not differ across cohorts) and allows me to focus interpretation on supply-side changes, i.e., the role of educational expansion. It also controls for supply-side confounders such as immigration. I further discuss the assumptions of this approach in the Online Appendix C. In addition to these fixed-effects, I adjust for a range of other variables that might confound estimates of the effects of educational expansion. At the individual level, I control for respondents’ migration background (Germany), or their ethnic group (United Kingdom), respectively, and an interaction of gender with their partnership status. Gender-specific results can be found in Section D of the Online Supplement. At the period-region-cohort level, I adjust for the size of workers’ extended cohort (as above defined as people born within +- 3 years from the base year), relative to the entire working age population 30 to 65 in that year in that region. This is to account for the possibility that members of relatively larger cohorts might face increased competition in accessing matching jobs. In Section F of the Online Supplement, I report specification curves for the main results, which show the full range of estimates obtained for all plausible specifications. Analytical strategy In order to estimate the relationship between educational expansion, measured as the extended cohort-specific share of qualification 𝑄𝑙, and the probability to be mismatched, I use the following random-effects linear probability model: 𝑝(𝑀𝑀𝑖,𝑟,𝑐,𝑡 =1)=∑𝛼𝑙𝑄𝑟,𝑐,𝑡,𝑙 𝑙=𝐿 𝑙=1 + ∑(𝛽𝑘𝑋𝑘,𝑖) 𝑘=𝐾 𝑘=1 +δ𝑟×𝑡 + 𝑢𝑐. (Eq. 1) 17 This model estimates the probability that an individual 𝑖, member of cohort 𝑐, surveyed in region 𝑟 at historical time 𝑡 will be mismatched. In this model, δ𝑟×𝑡 represents the period-region fixed-effects, 𝑢𝑐 is a cohort-level random effect to account for the clustering of observations and measures within cohorts, and the term ∑(𝛽𝑘𝑋𝑘,𝑖) 𝑘=𝐾 𝑘=1 represents 𝐾 control variables including individual education. I estimate Eq. 1 separately for underand overqualification as the mismatch outcome. When estimating the probability of overqualification, underqualified workers form part of the reference category of non-overqualified workers, and vice versa. Workers who cannot indicate a mismatch in the survey questionnaire (university graduates, for instance, cannot be formally underqualified and people without qualifications cannot be overqualified) are excluded from the respective models. The quantity of interest in both kinds of models are the 𝛼𝑙, the partial relationship between the cohort-specific share of qualification 𝑄𝑙 in 𝑟, at 𝑡 (𝑙 indexes the qualification levels) and the corresponding underor overqualification probability. The Online Supplement C discusses the principal conditions under which a causal interpretation can be assigned to this coefficient. There are two more specific threats to identifying the causal effects of educational expansion in my framework: endogenous migration in response to labor market conditions and sample selection bias arising from selective unemployment. I describe my strategy to tackle these issues, and the resulting estimates in Appendix G of the Online Supplement. 4. Results Educational expansion, occupational upgrading, and mismatch trends Before I turn to the question of how rates of vertical mismatch have evolved, Figure 1 gives a descriptive impression of the supply of qualifications, the change of the occupational structure, and the resulting demand for qualifications on British and German labor markets. Panel A of Figure 1 takes a birth-cohort perspective and shows the highest qualification reported by 30year-olds in large-scale labor force surveys. It documents significant educational expansion in 18 the United Kingdom, and, to a lesser degree, in Germany for cohorts born between the mid1950s and the mid-1980s. While roughly a fifth of a cohort born in the 1950s has undergone tertiary education in both societies, the figure for Germany stood at 25% in the 1983 cohort and at more than 35% in the same birth-cohort in the United Kingdom. A second difference between the two countries is that German expansion was limited to the top, i.e., to tertiary education, with the share of those without any vocational qualifications not showing any clear trend. British education, on the other hand, expanded at all levels, thereby dramatically decreasing the share of those without formal vocational education from over 60% of a cohort born in the 1950s to around 30% for cohorts born in the 1980s. This figure is, however, still more than twice as high as the share of 30-year-old Germans without vocational qualifications. Panel B of Figure 1 changes perspectives and shows the evaluation of the occupational structure of the German and the British Labor markets over historical time. It uses the first digit of the ISCO88 classification to define occupational groups. In both countries, higher skilled occupations (managers, professionals, technicians) expanded, while lower-skilled ones (craft workers, machine operators, elementary occupations) contracted between the 1980s and the 2010s. However, there are differences, when it comes to the overall frequency of occupations: by the end of the 2010s, for instance, I estimate the share of workers in the manager category to be approximately 3.3 times higher in the United Kingdom than in Germany, (20% vs. 6%), while the opposite is true for technicians (15% vs 25%). There is thus clear evidence of occupational upgrading and no evidence of polarization—at least at the relatively coarse level of occupational groups—in both countries, but the descriptive analysis also documents persistent differences in the composition of the two labor markets. Panel C moves to the core of my analysis and shows the qualifications workers reported were needed for their current job in a given survey-year. It presents evidence that the occupational upgrading apparent in Panel B translated into a significantly increasing demand for (higher) qualifications on both British and German labor markets. This is true even though my subjective measure of mismatch likely underestimates the extend of actual demand shifts, as qualification requirements of unfilled vacancies are not counted. 19 In Germany, the share of jobs requiring no formal qualifications at all has halved between 1984 and 2018, from more than 44% to 22%, while tertiary-education-requirements expanded from 11% of jobs to 26%. This shows that, against RBTC-arguments, demand upgrading was linear: not only did demand for highly qualified workers grow, but the relative demand for not formally qualified workers shrank. Interestingly, this is different from the form of educational expansion. Panel A documents that it was limited to tertiary degrees. This hints at a first source of underexpansion in the German case: the collapsing demand for untrained workers. In the United Kingdom, no vocational education beyond secondary school diplomas was required to get into half of jobs in 1986, in 2017, this was only true for a third. A tertiary-education requirement rose from a quarter of jobs in the 1980s to just under half in 2017. In contrast to Germany, these two trends imply that the demand for middling post-secondary, non-tertiary vocational qualifications declined markedly, from 37% to 18% in 2017. In line with Oesch and Picitto (2019), and in contrast to the trends at the level of coarse occupational groups in Panel B, the United Kingdom therefore shows some signs of polarized upgrading. 20 Figure 1: Educational expansion, occupational change, and upgrading of jobs’ qualification requirements in the United Kingdom and in Germany Source: Own calculation based on GMZ 1976-2013, UKLFS 1979-2017, GSOEP 1984-2017, and UKSESS 1986-2017. The first round of the UKSESS was not carried out in all regions. 1986 figures are hence not strictly comparable to later periods. 21 Combining demandand supply-side perspective, I now move to an examination of mismatch rates. I first examine what descriptive trends in mismatch prevalence can tell us about the validity of the two universalist theoretical perspectives, before I turn to a formal assessment of the relationship between expansion and mismatch at the cohort level. Figure 2 shows the overall and cohort-specific trend in self-reported mismatch rates over survey-years in the two countries and thereby offers a first descriptive assessment, limiting the analysis to workers between the ages of 30 and 60. It documents that mismatch trends in the United Kingdom and in Germany have been strikingly different. While there are signs of overexpansion in the United Kingdom, Germany’s labor market seems to be moving into a mild qualification shortage. Overall, overqualification rates rose in the United Kingdom but declined in Germany. The opposite is true for underqualification. A second noteworthy observation is that there are only small differences between cohorts (lines in gray) in Germany, whereas British cohorts face vastly different situations, even at the same point in time. Respective younger cohorts faced the highest risk to be overqualified, and the lowest risk to be underqualified during almost all survey years. The rising rate of overqualification and the declining rate of underqualification thus appear to be mainly driven by cohort replacement. Trends in Germany, on the other hand, appear to take place mainly at the period level, with mismatch rates of different cohorts being almost indistinguishable. If anything, younger cohorts appear to show lower self-assessed overqualification rates from the late 2000s onward. Note that country differences, thus, evolve in opposing directions to any possible bias introduced by the slightly differing item wording in the United Kingdom and in Germany. 22 Figure 2 suggests that overand underexpansion are highly context-dependent phenomena. However, it does not account for the fact that with educational expansion, a higher share of workers becomes eligible for overqualification in the first place, and likewise is no longer at risk of underqualification. Moreover, at this point, it is unclear whether mismatch trends are predominantly driven by the demand or the supply side. I thus now turn to models described in Eq. 1 that link overand underqualification incidence to cohort-level qualification shares — the main indicator of educational expansion. Figure 2 Trends in mismatch rates in the United Kingdom and in Germany 23 Evidence from regression models Table 1 provides a global assessment of the link between educational expansion over birthcohorts and cohort level-over-/underqualification incidence. It answers the core question at stake in the overexpansion debate whether we observe more or less over-/underqualification in cohorts where higher levels of education are more widespread. Using a conceptually different approach, it confirms the significant difference in the relationship between expansion and mismatch in the two countries suggested by Figure 1: overexpansion in the United Kingdom, but an overall balance in Germany. Table 1 suggests that British expansion at both the lower and the upper end of the qualification hierarchy was only partially absorbed by labor markets. Even net of own qualifications, and net of period-region differences in demand, overqualification is higher in region-cohorts, where the share of upper tertiary and secondary educated workers is higher. Expansion also severely limited underqualification: this type of mismatch is dramatically lower in region-cohorts where upper tertiary and secondary education is more prevalent. Expansion of middling qualification levels, however, is not associated with changed mismatch rates. I do not find any evidence of a comparable overexpansion in the German data. On the contrary, column 3 shows that in Germany overqualification decreased in cohort regions, where tertiary education increased. The effect sizes I find are large: I estimate that a 10-percentage point increase in the number of graduates has historically been associated with a 2.6-percentage point increase in overqualified and a 5-percentage point reduction in underqualified workers in the United Kingdom. In Germany, overqualification reduced at more than half the rate at which tertiary education expanded. 24 United Kingdom Germany Overqualification Underqualification Overqualification Underqualification (1) (2) (3) (4) Expansion of… …secondary 0.192* -0.195* (1.98) (-2.29) …lower voc. -0.264 0.151 (-0.96) (0.61) …vocational 0.0293 0.0953 0.173 0.0636 (0.14) (0.54) (1.14) (0.61) …lower tert. -0.175 0.382 (-0.58) (1.40) …tertiary 0.259* -0.495*** -0.535*** 0.0879 education (2.41) (-4.89) (-3.55) (0.85) Individual controls Yes Yes Yes Yes Cohort controls Yes Yes Yes Yes N 17878 16560 21489 17971 NYears 7 7 29 29 NRegions 12 12 10 10 NCohorts 61 61 56 56 Variance components VarIntercept 0.0000668 2.22e-20 1.51e-15 3.18e-17 VarResidual 0.227 0.170 0.153 0.0610 t statistics in parentheses. Cohort random effects. Individual controls: Interaction gender and partnership status, and ethnicity. Cohort controls: relative size of cohort. Results controlled for period-region fixed-effects. * p < 0.05, ** p < 0.01, *** p < 0.001 Table 1 Relationship between educational expansion and qualification mismatch, RE-LPM results Table 1 reports results, when all qualification shares are entered simultaneously, omitting the “no qualification” category. Results for using other references categories can be found in Section E of the Online Supplement. These additional estimates underline the robustness of my results. As I discuss in more detail in the Online Supplement C, no causal interpretation can be applied to the German results. A valid, but more cautious interpretation of columns 3 and 4 in Table 1 is that tertiary expansion in Germany was not enough to stabilize rates of overqualification, which were falling for other reasons. Even without a causal interpretation, this pattern rules out the overexpansion thesis for the German case. How did the apparent overexpansion of British secondary and tertiary education play out across different levels of education? And what kinds of workers found it easier to move out of overqualification in Germany? Figures 3 and 4 break down the association between expansion 31 A third contribution of this study is to take a holistic approach by studying mismatch and expansion across levels of education. This perspective affords new insights. I find that tertiary expansion in the United Kingdom was not associated with overqualification of university graduates, but of graduates of lower tertiary institutions and vocational programs – suggesting a labor queue model in which expansion at the top fuels overqualification among workers with middling qualifications, as “their” jobs are increasingly filled with graduates. My analysis of the German data suggests that any possible underexpansion is not to be found among the most highly qualified workers, university graduates. Supplementary analyses even yielded some evidence that graduates’ access to executive positions has become somewhat harder with tertiary expansion. Instead, underexpansion seems to be located among workers with (advanced) vocational qualifications. If this interpretation is correct, educational policy in Germany should focus on shoring up the vocational sector, rather than expanding university access further. References Acemoglu, Daron, and David Autor. 2011. “Skills, Tasks and Technologies: Implications for Employment and Earnings.” Pp. 1043–1171 in Handbook of Labor Economics, edited by O. Ashenfelter, R. Layard, and D. Card. Amsterdam; New York: Elsevier. ———. 2012. “What Does Human Capital Do? A Review of Goldin and Katz’s The Race between Education and Technology.” Journal of Economic Literature 50(2): 426–63. Allen, Jim, and Rolf van der Velden. 2001. “Educational Mismatches versus Skill Mismatches: Effects on Wages, Job Satisfaction, and on-the-Job Search.” Oxford Economic Papers 53(3): 434–52. Autor, David H, and David Dorn. 2013. “The Growth of Low-Skill Service Jobs and the Polarization of the US Labor Market.” American Economic Review 103(5): 1553–97. Autor, David H., Frank Levy, and Richard J. Murnane. 2003. “The Skill Content of Recent Technological Change: An Empirical Exploration.” The Quarterly Journal of Economics 118(4): 1279–1333. Baker, David. 2014. The Schooled Society: The Educational Transformation of Global Culture. Stanford University Press. Bol, T. 2015. “Has Education Become More Positional? Educational Expansion and Labour Market Outcomes, 1985-2007.” Acta Sociologica 58(2): 105–20. Bol, Thijs, and Kim A. Weeden. 2015. “Occupational Closure and Wage Inequality in Germany and the United Kingdom.” European Sociological Review 31(3): 354–69. Bourdieu, Pierre, and Jean-Claude Passeron. 1970. La Reproduction: Éléments pour une théorie du système d’enseignement. Minuit. Bowles, Samuel, and Herbert Gintis. 1977. Schooling In Capitalist America: Educational Reform And The Contradictions Of Economic Life. New edition. New edition. New York: Basic Books. Burris, Val. 1983. “The Social and Political Consequences of Overeducation.” American Sociological Review 48(4): 454. 32 Busemeyer, Marius R. 2014. “The Politics of Education and Training Reform.” Pp. 58–122 in Skills and Inequality: Partisan Politics and the Political Economy of Education Reforms in Western Welfare States. Cambridge: Cambridge University Press. Cappelli, Peter H. 2015. “Skill Gaps, Skill Shortages, and Skill Mismatches Evidence and Arguments for the United States.” ILR Review 68(2): 251–90. Capsada-Munsech, Queralt. 2015. “The Role of Social Origin and Field of Study on Graduates’ Overeducation: The Case of Italy.” Higher Education 69(5): 779–807. ———. 2019. “Measuring Overeducation: Incidence, Correlation and Overlaps Across Indicators and Countries.” Social Indicators Research 145(1): 279–301. Card, David, and John E. DiNardo. 2002. “Skill-Biased Technological Change and Rising Wage Inequality: Some Problems and Puzzles.” Journal of Labor Economics 20(4): 733–83. Carnevale, Anthony P, and Donna M Desrochers. 2002. “The Missing Middle:Aligning Education and the Knowledge Economy.” 25(1). Carnevale, Anthony P, and Stephen J Rose. 2014. “The Undereducated American.” Georgetown University Center on Education and the Workforce. Collins, Randall. 1979. The Credential Society : An Historical Sociology of Education and Stratification. New York : Academic Press. Croce, Giuseppe, and Emanuela Ghignoni. 2012. “Demand and Supply of Skilled Labour and Overeducation in Europe: A Country-Level Analysis.” Comparative Economic Studies 54(2): 413–39. Davia, Maria A., Seamus McGuinness, and Philip J. O’Connell. 2017. “Determinants of Regional Differences in Rates of Overeducation in Europe.” Social Science Research 63: 67–80. Delaney, Judith, Seamus McGuinness, Konstantinos Pouliakas, and Paul Redmond. 2020. “Educational Expansion and Overeducation of Young Graduates: A Comparative Analysis of 30 European Countries.” Oxford Review of Education 46(1): 10–29. Di Pietro, Giorgio. 2002. “Technological Change, Labor Markets, and ‘Low-Skill, Low-Technology Traps.’” Technological Forecasting and Social Change 69(9): 885–95. Di Stasio, Valentina. 2017. “Who Is Ahead in the Labor Queue? Institutions’ and Employers’ Perspective on Overeducation, Undereducation, and Horizontal Mismatches.” Sociology of Education 90(2): 109–26. Di Stasio, Valentina, Thijs Bol, and Herman G. Van de Werfhorst. 2016. “What Makes Education Positional? Institutions, Overeducation and the Competition for Jobs.” Research in Social Stratification and Mobility 43: 53–63. Dustmann, Christian, Johannes Ludsteck, and Uta Schönberg. 2009. “Revisiting the German Wage Structure.” Quarterly Journal of Economics 124(2): 843–81. Elbers, Benjamin, Thijs Bol, and Thomas A DiPrete. 2021. “Training Regimes and Skill Formation in France and Germany An Analysis of Change Between 1970 and 2010.” Social Forces 99(3): 1113–45. Esping-Andersen, Gosta. 1999. Social Foundations of Postindustrial Economies. Oxford University Press. Felstead, Gallie, Green, and Inanc. 2014. “Skills and Employment Surveys Series Dataset, 1986, 1992, 1997, 2001, 2006 and 2012.” Colchester, Essex: UK Data Archive. Fernández-Macías, Enrique, and John Hurley. 2017. “Routine-Biased Technical Change and Job Polarization in Europe.” Socio-Economic Review 15(3): 563–85. Fitzenberger, Bernd, Karsten Kohn, and Alexander C. Lembcke. 2013. “Union Density and Varieties of Coverage: The Anatomy of Union Wage Effects in Germany.” ILR Review 66(1): 169–97. Freeman, Richard. 1976. The Overeducated American. Academic Press. Furey, Jane. 2021. “Relative to Whom? Comment on ‘Relative Education and the Advantage of a College Degree.’” American Sociological Review 86(5): 1000–1010. Geißler, Rainer. 2014. Die Sozialstruktur Deutschlands. Wiesbaden: Springer Fachmedien Wiesbaden. Glendinning, Richard, Viv Young, Alexandra Bogdan, and GfK UK Social Research. 2018. Skills and Employment Survey 2017. Technical Report. 33 Goebel, Jan, Markus M. Grabka, Stefan Liebig, Martin Kroh, David Richter, Carsten Schröder, and Jürgen Schupp. 2019. “The German Socio-Economic Panel (SOEP).” Jahrbücher für Nationalökonomie und Statistik 239(2): 345–60. Goldin, Claudia, and Lawrence F. Katz. 2010. The Race between Education and Technology. 2/28/10 edition. 2/28/10 edition. Cambridge, Mass.: Belknap Press. Gosling, Amanda, and Thomas Lemieux. 2004. “Labor Market Reforms and Changes in Wage Inequality in the United Kingdom and the United States.” in Seeking a Premier Economy: The Economic Effects of British Economic Reforms, 1980-2000, edited by D. Card, R. Blundell, and R. B. Freeman. University of Chicago Press. Hall, Peter A., and David Soskice. 2001. “Introduction to Varieties of Capitalism.” Pp. 1–68 in Varieties of Capitalism. The Institutional Foundations of Comparative Advantage. Oxford, New York: Oxford University Press. Heisig, Jan Paul. 2018. “Measuring the Signaling Value of Educational Degrees: Secondary Education Systems and the Internal Homogeneity of Educational Groups.” Large-scale Assessments in Education 6(1): 9. Holzer, Harry. 2015. “Job Market Polarization and U.S. Worker Skills: A Tale of Two Middles.” Economic Studies at Brookings. Holzer, Harry J, and Robert I Lerman. 2007. “America’s Forgotten Middle-Skill Jobs: Education and Training Requirements in the Next Decade and Beyond.” Workforce Alliance. ———. 2009. “The Future of Middle-Skill Jobs.” Horowitz, Jonathan. 2018. “Relative Education and the Advantage of a College Degree.” American Sociological Review 83(4): 771–801. House of Commons. 2012. Education: Historical Statistics. London. Kalleberg, Arne L. 2011. Good Jobs, Bad Jobs : The Rise of Polarized and Precarious Employment Systems in the United States, 1970s-2000s. New York: Russell Sage Foundation. Katz, L. F., and K. M. Murphy. 1992. “Changes in Relative Wages, 1963-1987: Supply and Demand Factors.” The Quarterly Journal of Economics 107(1): 35–78. Korpi, Tomas, and Michael Tåhlin. 2009. “Educational Mismatch, Wages, and Wage Growth: Overeducation in Sweden, 1974–2000.” Labour Economics 16(2): 183–93. Kristal, Tali, and Yinon Cohen. 2017. “The Causes of Rising Wage Inequality: The Race between Institutions and Technology.” Socio-Economic Review 15(1): 187–212. Mayer, Karl Ulrich, Walter Müller, and Reinhard Pollak. 2007. “Germany: Institutional Change and Inequalities of Access in Higher Education.” in Stratification in Higher Education: A Comparative Study, edited by Y. Shavit, R. Arum, and A. Gamoran. Stanford, California: Stanford University Press. McGuinness, Seamus, Adele Bergin, and Adele Whelan. 2018. “Overeducation in Europe: Trends, Convergence, and Drivers.” Oxford Economic Papers 70(4): 994–1015. Müller, Walter, and Markus Gangl. 2003. Transitions from Education to Work in Europe: The Integration of Youth Into Eu Labour Markets. Oxford ; New York: Oxford Univ Pr. Nolan, Brian, Wiemer Salverda, Daniele Checchi, Ive Marx, Abigail McKnight, István György Tóth, and Herman G. van de Werfhorst (eds.). 2014. Changing Inequalities and Societal Impacts in Rich Countries: Thirty Countries’ Experiences. Oxford, New York: Oxford University Press. OECD. 2019. Education at a Glance 2019: OECD Indicators. OECD. Oesch, D., and J. Rodriguez Menes. 2011. “Upgrading or Polarization? Occupational Change in Britain, Germany, Spain and Switzerland, 1990-2008.” Socio-Economic Review 9(3): 503–31. Oesch, Daniel, and Giorgio Piccitto. 2019. “The Polarization Myth: Occupational Upgrading in Germany, Spain, Sweden, and the UK, 1992–2015.” Work and Occupations 46(4): 441–69. Pollmann-Schult, Matthias, and Karl Ulrich Mayer. 2004. “Returns to Skills: Vocational Training in Germany 1935-2000.” Yale Journal of Sociology 4: 24. Powell, Justin J. W., and Heike Solga. 2011. “Why Are Higher Education Participation Rates in Germany so Low? Institutional Barriers to Higher Education Expansion.” Journal of Education and Work 24(1–2): 49–68. 34 Rossen, Anja, Christina Boll, and André Wolf. 2019. “Patterns of Overeducation in Europe: The Role of Field of Study.” IZA Journal of Labor Policy 9(1). Shin, Dong-Hoon, and David Bills. 2021. “Trends in Educational and Skill Mismatch in the United States.” Social Sciences 10(10): 395. Thurow, Lester C. 1975. Generating Inequality: Mechanisms of Distribution in the U.S. Economy. Basic Books. Tinbergen, Jan. 1956. “On the Theory of Income Distribution.” Weltwirtschaftliches Archiv 77: 22. Vaisey, Stephen. 2006. “Education and Its Discontents: Overqualification in America, 1972– 2002.” Social Forces 85(2): 835–64. Verhaest, Dieter, and Rolf Van der Velden. 2013. “Cross-Country Differences in Graduate Overeducation.” European Sociological Review 29(3): 642–53. Wagner, Gert G., Joachim R. Frick, and Jürgen Schupp. 2007. “The German Socio-Economic Panel Study (SOEP)-Evolution, Scope and Enhancements.” Weeden, Kim A., and David B. Grusky. 2014. “Inequality and Market Failure.” American Behavioral Scientist 58(3): 473–91. Wiedner, Jonas. 2021. “Political and Social Consequences of Qualification Mismatches: A Bounding Approach to Status Inconsistency.” Social Forces (soab120). Wolf, Alison. 2002. Does Education Matter? Myths About Education and Economic Growth. Penguin Books Limited. Wren, Anne (ed.). 2013. The Political Economy of the Service Transition. Oxford, UK: Oxford University Press.