Who Is Ahead in the Labor Queue? Institutions’ and Employers’ Perspective on Overeducation,Undereducation,and Horizontal Mismatches

Abstract

Using vignettes, this study compares employers’ assessments of matched and mismatched job applicants in England and the Netherlands. It contributes to the overeducation literature in several ways. First, matching is measured from the perspective of employers, who are better informed about job requirements than employees. Second, overeducated applicants are compared to matched applicants competing for the same job opening. This shift in focus toward applicant pools is necessary to properly test whether overeducation is rewarded during the hiring process, the central tenet of job competition theory. Third, vertical and horizontal mismatches are analyzed jointly: This more fine-grained differentiation refines sociological perspectives on credentialism and reveals the complex ways in which employers assign applicants to jobs. Results show that Dutch employers apply more rigid hiring floors and more strongly penalize horizontal mismatches: Compared to England, in the Netherlands, overeducation cannot compensate for the lack of occupation-specific training.

Keywords

employers overeducation horizontal mismatches credentialism vignettes institutions education systems

Overeducation refers to a misalignment between employees’ level of education and the level of education required to get or do their jobs (Halaby 1994; Sicherman 1991). The job competition model (Thurow 1975) provides a convincing explanation for the occurrence of overeducation in the labor market. During the hiring process, employers have only limited information to assess potential employees’ future productivity and look for the best available candidate at the lowest training costs. Employers rank candidates in an imaginary labor queue and offer the job to the top-ranked candidate. Education is a signal of trainability—that is, one’s willingness to learn new skills—and is one of the main criteria on which employers base their rankings. By hiring overeducated employees, employers try to save on training costs.

The job competition model is premised on the assumption that more education is always better in the eyes of employers; the content of education is largely ignored. This neglect is unfortunate as in some countries job placement depends on the content of qualifications as much as on their level. Recent contributions to the sociological literature on credentialism and occupational closure show that entry into occupations is often conditional on the acquisition of occupation-specific credentials, and social and legal barriers can put artificial restrictions on the supply of labor that is allowed to perform certain jobs (Bol and Weeden 2015; Brown 2001; Weeden 2002). From a closure perspective, a surplus of education in a field of study that does not match the occupation of destination is a horizontal mismatch and should not be rewarded (Solga and Konietzka 1999; Wolbers 2003). Furthermore, education systems vary in the extent to which their degree structure tracks workers into particular occupations. As a result, the prevalence and strength of occupational closure vary across countries (Bol and Weeden 2015). Where the pathways between education and the labor market are strongly institutionalized, employers demand specific educational credentials from prospective employees, and overeducation per se should be less of an asset. Building on this insight, I analyze vertical and horizontal mismatches jointly, and I compare employers’ evaluations of overeducated job applicants across two different national contexts.

Unlike most studies on overeducation, I focus on the demand side of the labor market. So far, only a handful of single-country studies have analyzed overeducation from the perspective of employers, and these studies hold the national context constant by design (e.g., Bills 1992; Shen and Kuhn 2012; Verhaest et al. 2016). Instead, I compare two very different national contexts, England¹ and the Netherlands, and restrict my focus to the information, communication, and technology (IT) sector. This design allows me to bring into sharper relief the role of the national context as I compare employers’ preferences across two labor market segments that are otherwise very similar in terms of business demographics, workforce characteristics, and economic climate (more information on the case selection is provided in the following).

I collected data from a sample of human resource (HR) professionals (i.e., recruiters, HR managers, and firm owners, henceforth called employers). Using vignettes of hypothetical job applicants, I analyzed how employers in the two countries assess applicants who are competing for the same job opening. Vignettes randomly varied on a number of characteristics, including level of education and field of study. Employers were asked to indicate the level of education necessary to get the same type of job in their organizations. I used this information to test (1) how employers evaluate horizontally and vertically mismatched applicants compared to fully matched applicants and (2) whether employers’ evaluations vary systematically across countries. Information about other aspects of human capital (e.g., grades, on-the-job training, and previous work experience) was also available, so I can rule out the possibility that employers may be using overeducation to compensate for skill heterogeneity.

With this study, I make a threefold contribution to the overeducation literature. First, I measure job requirements from the perspective of employers, who are arguably better informed about job requirements than are current employees. Second, I compare applicants who are competing for the same job opening. This shift in focus away from job incumbents and toward applicant pools is necessary to properly test the central tenet of the job competition model, which refers to the ranking of applicants during the hiring process. Third, I analyze vertical and horizontal mismatches jointly. This fine-grained classification shows how processes of credential closure influence employers’ hiring preferences.

Results reveal that employers place overeducated applicants ahead in labor queues—as the job competition model predicts—but only if applicants’ field of specialization matches the job’s occupational domain. Applicants with additional years of education but lacking occupation-specific training did not experience any advantage over applicants with fully matched qualifications, and in the Dutch context, they were even penalized. Therefore, focusing on only the vertical dimension of mismatches would be inappropriate in countries with a high degree of credential closure, like the Netherlands, where additional years of schooling cannot compensate for the lack of occupation-specific training.

It is important to understand how overeducation occurs in the labor market. Overeducation is an indication that one’s investment in education has not yielded the expected returns in the transition to the labor market and that a stock of human capital is going to waste or is not utilized efficiently. Even if provisional, overeducation is associated with low job satisfaction (Allen and van der Velden 2001) and can be as detrimental for future employment transitions as a spell of unemployment (Pedulla 2016). This scarring effect can trap labor market entrants in lower-level jobs with limited opportunities for career advancement or training, and it can have negative consequences on job quality in the long term. Moreover, employers’ preferences for overeducated workers may push individuals to overinvest in education as a defensive necessity to remain competitive in a knowledge-intensive society (Bills 2016). This can lead to processes of occupational displacement and further depress labor market opportunities for people with less education. From a theoretical point of view, this study responds to a recent call for research on how credentialist processes are sustained and embedded in larger institutional contexts (Bills and Brown 2011). It also relates to a recent proposal to analyze processes of occupational sorting in terms of both education levels and fields of study (DiPrete et al. forthcoming) as this granularity can improve our understanding of school-to-work transitions in international comparisons.

A Change of Perspective: From Overeducated Job Incumbents to Overeducated Job Applicants

Although queuing occurs at the pre-hire stage, empirical studies on overeducation are generally based on data of job incumbents. By pooling together individuals who have already been hired, these studies select on the dependent variable, and various forms of selection bias can affect the analysis (Fernandez and Weinberg 1997). Studies of job incumbents thus “fail to measure one of the defining dimensions of labor queues, i.e., the rank ordering of the set of people that hiring agents choose among” (Fernandez and Sosa 2005:1062). This rank ordering is key to the job competition model.

Nearly all work on overeducation focuses on employees, with a few exceptions. In a qualitative study of six organizations in the Chicago area, Bills (1992) urged hiring managers to make frequent and explicit comparisons between their most recently hired employee and other applicants who did not succeed in getting the job. Only 9 percent of employers declared they had inflexible hiring standards; many were willing to accept overeducated applicants or would even prefer them over matched applicants if given the choice. Similarly, Kulkarni, Lengnick-Hall, and Martinez (2015) found that employers from a wide range of industries in the United States were willing to adjust their requirements both upward and downward. These two studies reorient the overeducation debate from job incumbents to job candidates and draw attention to the role of employers. However, employers were asked to remember the exact composition of applicant pools, a retrospective method that is prone to recall bias. As Bills (1992:82) acknowledges, “one would optimally like data from appointing managers on candidates who were not selected,” that is, data on applicant pools.

Studies based on personnel records have an ideal research design to analyze applicant pools. For example, Fernandez and Sosa (2005) examined applications to customer service jobs at a large U.S. bank and found that overeducated applicants had higher—but not statistically significant—chances to be interviewed than did applicants with the required education level. Findings from this study, though, cannot be easily generalized to other settings. More recently, Shen and Kuhn (2012) had access to all applications submitted to the official Internet job board of a large Chinese urban area. Comparing the educational requirements mentioned in job advertisements with applicants’ education levels, they found that overeducation increased applicants’ likelihood to be shortlisted for a job. Interestingly, the overeducation premium was limited to graduates with a four-year university degree applying for jobs that required a three-year degree. A surplus of years of schooling did not bring any advantage—and in fact was even detrimental—to applicants with a three-year university degree applying for jobs that required a vocational degree from a technical school. A possible explanation for this finding is that occupationally focused training is not directly substitutable with more general college education.

The information recorded in personnel files is often very basic, and differences between applicants in the likelihood to be shortlisted or hired may be due to unobservables. Studies based on the correspondence testing method try to overcome this limitation. In the Belgian labor market, for example, Verhaest and colleagues (2016) sent matched applications in response to vacancies that mentioned a bachelor’s degree as an entry requirement. Applicants were equivalent in all respects except for their level of education: Half had a bachelor’s degree from a vocational college, and the other half had a master’s degree, which made them overqualified for the advertised jobs. On average, the probability of receiving a positive callback from employers was 11 percent higher for applicants with master’s degrees than for those with bachelor’s degrees; this difference was even higher for the probability of being invited to a job interview (19 percent). Employers did prefer overqualified applicants but not for vacancies with qualitative bottlenecks (i.e., difficult to fill vacancies due to a shortage of skills). This indicates that for Belgian employers, university education cannot compensate for the lack of more practical knowledge and skills.

Finally, a few studies on mismatches have used survey experiments. van Beek, Koopmans, and van Praag (1997) found that Dutch employers strongly penalized undereducated applicants but were indifferent toward overeducated ones. In a vignette study administered to personnel representatives of electronics manufacturing firms in Colorado, overeducated applicants were rated more positively with regard to motivation, productivity, promotion, training potential, commitment to the organization, and overall performance (Athey and Hautaluoma 1994).

Based on this brief literature review, we can make two observations. First, overeducated applicants are generally placed at the top of labor queues, as the job competition model predicts, but their relative advantage over adequately matched applicants varies across contexts. Second, overeducation is an imperfect substitute for the lack of more practical skills (Shen and Kuhn 2012; Verhaest et al. 2016), a result that invites us to consider horizontal and vertical mismatches jointly. I build on these two observations in the next section.

Theoretical Framework: Job–Education Matching in Context

Recently, Bills and Brown (2011), in a comment on the development of credentialism theory in the sociological literature, noted that the concept of overeducation has clear affinities with that of credential inflation (Brown 2001; Collins 1979), although these two literatures have developed largely independently. Both literatures originated in the United States during a period of dramatic educational expansion. Following the credential inflation argument, employers responded to an increasing supply of highly educated workers by raising their hiring standards. The upgrading of entry requirements did not reflect an upgrading of the occupational structure, and it gave rise to vertical mismatches: Employers preferred overeducated applicants for jobs that were previously accessible to high school graduates.

A different and more recent interpretation of educational credentialism stresses the role of specific educational credentials in processes of occupational allocation. The matching between qualifications and jobs—a process Bills and Brown (2011) call “credentialist hiring”—likely varies depending on various closure mechanisms that structure the acquisition of skills in the formal education system and the transfer of occupation-specific knowledge to the workplace. For example, legal as well as normative barriers set collectively by social partners, occupational groups, overseeing authorities, or professional associations can restrict entry into specific occupations. These barriers take the form of institutionalized regulations that bind employers—more or less forcefully depending on the labor market segment—to comply with formal qualification requirements for their recruitment and selection decisions (Bol and Weeden 2015; Weeden 2002).

I build my theoretical framework on this insight, relying on recent comparative research on overeducation and horizontal mismatches. For example, van de Werfhorst (2011) compares job–education mismatches across Dutch industries characterized by different degrees of occupational closure. He found a lower incidence of vertical mismatches in industries with high levels of occupational closure and a higher incidence in industries with high training intensity, where skills can be acquired on the job. This argument found support in some cross-national studies. Countries largely differ in the extent to which their education systems prepare students for labor market entry. Education systems may either offer occupationally focused training— based on strict cooperation between schools, employers, and social partners—or provide generalist training and a relatively weak system of standardized qualifications (Kerckhoff 2001; Müller and Gangl 2003). In the first group of countries, education provides students with occupation-specific skills; in the latter, such skills are acquired primarily on the job (Di Stasio and van de Werfhorst 2016). Indeed, Levels, van der Velden, and Di Stasio (2014) and Verhaest and van der Velden (2013) show that the likelihood of finding a matching job is higher in countries with an occupationally focused education system, where new hires can perform well from the outset and employers do not need to hire overeducated applicants to lower training costs. Verhaest, Sellami, and van der Velden (2015) reach a similar conclusion in relation to horizontal mismatches.

Following this line of reasoning, I test whether the likelihood that employers hire vertically or horizontally mismatched applicants varies across institutional settings. To bring the role of national institutions into sharp relief, I focus on England and the Netherlands, countries with very different education systems. The English education system is generalist: Educational credentials are only loosely coupled with occupational demands, the coordination of employment relations is weak, and the provision of training is largely decentralized to single employers. This education system does not provide institutionalized pathways to the workplace. In the Netherlands, in contrast, social partners are heavily involved in the coordination of occupational standards, and qualifications are highly standardized and generally well recognized by employers within occupational labor markets. Studies based on supply-side surveys have found that the correspondence between qualifications and occupational attainments is tighter in the Netherlands than in England (e.g., Iannelli and Raffe 2007; Müller 2005). Because mismatches result from informational asymmetries during the hiring process, I expect English employers to be more willing than Dutch employers to hire mismatched applicants of both types, vertical and horizontal. I also expect undereducated applicants will be penalized more severely in the Netherlands than in England.

I distinguish between three different stages of the hiring process: the initial screening of applications, the shortlist stage, and the interview stage. Following Thurow’s (1975) theoretical model, overeducated applicants should be placed ahead in the labor queue at each stage. This should occur to a lesser degree in the Netherlands, where the tight coordination of occupational labor markets leaves little room for mismatches and years of schooling should be of little value if they are not occupation specific.

Research Design

This study is a cross-national comparison of employers working in the IT sector and hiring IT professionals. This more focused research design has several advantages. First, IT professionals are one of the fastest growing occupations in recent years (Oesch and Rodrguez Mens 2011). They are employed in a variety of industries besides the core IT sector, including banking, finance, health care, and education. Thus, understanding how employers select IT professionals is important as such decisions affect a large number of job seekers.

Second, trainability is particularly important in environments with high training intensity, such as the IT sector, where core skills are for the most part learned at the workplace through intensive and frequent on-the-job training. In these settings, overeducated employees should be ranked ahead in labor queues due to their higher training potential.

Third, by choosing the IT sector as a case study, I gain traction to understand the influence of the national context on job matching processes. The many similarities shared by the two countries with regard to supply- and demand-side factors, summarized in Table 1, ease cross-national comparison. At least half of the IT workforce in both countries completed tertiary education. Furthermore, the economic crisis had a similar impact on employment figures in the two economies. This is important because employers’ perceptions of IT educational credentials may vary depending on the state of the labor market, as Van Noy and Jacobs (2012) show for the two very different labor markets of Seattle and Detroit. In this sector, investment per person employed, turnover per person employed, and percentage of high-growth enterprises are nearly identical in the two countries. Finally, Eurostat data on vacancy statistics reveal that job vacancy rates were very similar in the UK and the Netherlands at the time of the study.

Table 1.

A Comparison of the Information, Communication, and Technology (IT) Sector between the United Kingdom and the Netherlands.

	United Kingdom^a	Netherlands
Number of persons employed per enterprise	5.0	4.5
Enterprises as share of the total (percent)
10–49 employees	19.4	20.3
50–249 employees	17.0	17.8
More than 250 employees	28.5	32.4
Percent tertiary-educated employees	58.8	51.8
Investment per person employed	3.4	3.4
Turnover per person employed (thousand euro)	141.7	148.7
Wage-adjusted labor productivity^b	154.9	121.9
Average personnel cost (thousand euro)^c	50.9	58.6
Percent high-growth enterprises^d	19.1	18.2
Percent change in number of employees 2008–2010 (in FTE)	−2.6 (–3.6)	−1.8 (–2.9)
Percent change in number of employees 2010–2012 (in FTE)	3.7 (3.1)	4.2 (3.8)
Job vacancy rate 2011/2012^e	3.0/2.7	3.5/3.1

Source: Eurostat, Structural Business Statistics for NACE Rev. 2, J 62 (Computer programming, consultancy and related activities). Statistics refer to the year 2010, right before the start of data collection. Accessed March 3, 2016 (http://ec.europa.eu/eurostat/statistics-explained/index.php/Archive:Computer_programming_and_consultancy_statistics_-_NACE_Rev._2).

Data refer to the UK as a whole and not just England, as statistics broken down by region were not available.

Value added divided by personnel costs, which is then adjusted by the share of total employees in the total number of persons employed. This indicator is based on expenditure for labor input rather than a headcount of labor input, so it is relevant for comparison across countries with different incidence of part-time or self-employment.

Total remuneration, in cash or in kind, payable by an employer to an employee for work carried out. This is divided by the number of employees.

Data are not available for 2010; high growth for the year 2012 is shown instead. The share of high-growth enterprises is relative to the total number of enterprises with more than 10 employees (i.e., the same type of enterprises sampled in the study).

Job vacancy rates measure the proportion of total posts that are vacant; this is considered a key indicator of the business cycle because it reflects the unmet demand for labor.

Note: FTE = Full Time Equivalents.

In light of these data, the IT sector is a strong testing ground for my argument that cross-national variation in employers’ views is conditional on national institutions: Other possible intervening variables, like macroeconomic factors, business demographics, or the supply of tertiary graduates, are de facto held constant.

Of course, results cannot be immediately generalized to other countries or sectors of the economy. Nevertheless, it is important to bear in mind that I chose this sector strategically. The research design is a least likely case study, which provides strong analytic leverage to test the validity of an argument about the role of national institutions. The inferential logic of the least likely case study “is based on the ‘Sinatra inference’—if I can make it there I can make it anywhere” (Levy 2008:12). Self-learning and on-the-job informal training are paramount for skill acquisition in the IT sector. Work demands are continuously shifting, and qualifications adapt slowly to technological change, so workers are required to keep their skills up to date. Over the years, professional groups have tried, with only mixed success, to formalize their skill set and bodies of knowledge in educational credentials. Therefore, it is reasonable to assume that even in the Netherlands—where labor market coordination is strong and the education system is highly standardized and occupationally oriented—the correspondence between formal education and occupational attainment is generally less tight in IT fields than in fields like engineering or manufacturing (Adams and Demaiter 2008; DiPrete et al. forthcoming). Cross-national differences in the way employers match qualifications to jobs are thus likely to be on the conservative side. I come back to this issue and to the scope conditions of my argument in the conclusions.

Data and Methods

Data Collection

For the empirical analysis, I simulated a hiring contest with Dutch and English employers. I recruited employers at random from public records of organizations. Dutch organizations were randomly selected from the list of “computer programming, consultancy, and related activities” provided by the Chamber of Commerce. I merged this list with the member records of organizations affiliated with the IT trade association (Nederland ICT). In England, I randomly selected organizations from the member records of the IT trade association for small and medium-sized enterprises (UKITA) and the trade association for the technology industry (INTELLECT). I contacted the organizations by phone and invited a member of the HR department to participate in the study; if a specific HR department did not exist, I spoke with the person responsible for selection of personnel.

In total, 72 employers participated in the hiring simulation.² Compared to previous employer studies based on convenience samples (e.g., Bills 1992; Kulkarni et al. 2015), random sampling from official membership records casts a wider net in terms of reaching a more heterogeneous sample. Table A1 in the Appendix, available in the online version of the article, shows that participating employers are well distributed across firms of different size, and response patterns across firms do not show any evidence of nonresponse bias. I did not sample micro-firms (i.e., firms with fewer than 10 employees) because they tend to have less formalized recruitment procedures and rely on methods such as employee referrals. In these less bureaucratic organizations, job requirements are generally defined quite loosely, meaning that my operationalization of overeducation, undereducation, and horizontal mismatches (described below) is probably less relevant.

I invited employers to take part in a hiring simulation and fill in a brief survey about their recruitment and selection practices. The simulation was based on a factorial survey (Wallander 2009). Employers were asked to think about a hypothetical job opening at their organization and evaluate a series of hypothetical profiles of job applicants, described in vignettes. At the start of the hiring simulation, employers read the description of either a software engineer job or an IT business consultant job (whichever job they were more familiar with). In the International Standard Classification of Occupations (ISCO-08), software engineers and business consultants are classified as “software and application developers and analysts” (ISCO code 251) and belong to the occupational group of IT professionals. To make the hiring simulation more realistic, I did not randomly assign job type. With an average of nine years of personnel selection experience, I am confident that employers could provide realistic information about the qualification requirements of the jobs under study. Nearly all employers reported that the job description was highly similar, similar, or somewhat similar to the description used in their organizations, which guarantees that the measure of mismatches is externally valid.

Vignette Design

Each vignette contained information on eight variables, also called vignette dimensions.³ I used three of these variables—highest level of education, field of study, and study completion—to create the indicators of horizontal and vertical (mis)matches. A crucial decision was how to match levels of education to the corresponding national degree classification. In England, the advanced-level qualifications (widely known as A-levels) represent an important educational transition during secondary education: Students who complete A-levels are eligible to enter university, and access to some study programs depends on the grade obtained in specific subjects. A-levels in applied subjects are also available, including health and social care, applied business, engineering, and computing. Tertiary education is largely academic oriented and is vertically structured in two consecutive steps, an undergraduate program (or short-tertiary cycle) followed by a postgraduate program (or long-tertiary cycle). In English applicant vignettes, I distinguished between A-levels, undergraduate (bachelor’s degree), and postgraduate (master’s degree) tertiary education.

In the Netherlands, a tracking system sorts students into one of three different tracks very early (at the age of 12). Tracks differ in content, length, and the options they offer for further learning. Along with academic upper-secondary education, which is meant to prepare students for university, a number of highly standardized vocational qualifications are available at four different levels. Vocational education in the Netherlands has a strong occupational focus, and social partners play an important role in decisions regarding the quality of training provision and the standards of vocational qualifications. In the Dutch tertiary education system, universities of applied science (hogescholen) have a strong orientation toward the labor market and coexist with academic tertiary education. In the Dutch applicant vignettes, I distinguished between vocational upper-secondary education, undergraduate degrees (bachelor’s degree) from a university of applied science, and postgraduate degrees (master’s degree) from research-oriented universities. This different operationalization of tertiary education is appropriate for the Dutch context, where only a minority of graduates enter the labor market after obtaining an academic bachelor’s degree.

Vignette applicants specialized in one of three subjects: informatics, economics, or social sciences. They could have completed their studies on time, been delayed two years, or dropped out before obtaining a degree. Three additional variables referred to aspects of human capital rarely reported in employee surveys: the applicant’s grade point average (fair vs. very good), whether the applicant had been an intern for the employer in the past (yes vs. no), and work experience (two years vs. no work experience). Vignettes also mentioned the applicant’s gender and information about extracurricular activities (yes, the applicant was on the board of a student committee vs. no activities). Employers were told that all applicants were young labor market entrants, at the beginning of their careers.⁴

The Cartesian product of the eight vignette dimensions generated a total of 864 vignettes. I randomly assigned sets of 18 vignettes to employers, following earlier research with Dutch employers that showed 18 was a feasible number (van Beek et al. 1997). On the basis of the characteristics listed in the vignettes, employers were instructed to indicate the likelihood they would hire the job applicant on a scale ranging from 0 to 100. Because the job competition model hinges on the concept of trainability, employers also indicated the likelihood that the applicant would be easy to train.

Joint Measurement of Vertical and Horizontal Mismatches

Unlike most studies in this field, I used information provided by employers as a benchmark for job requirements. I classified applicants as undereducated or overeducated if there was a gap between the educational attainment reported in the vignette and the level of education that, according to the employer who rated the same vignette, was necessary to get a similar job in his or her organization. If the two levels coincided, the applicant was a vertical match. To take early drop-outs into account, I interpreted a matching level of education as a vertical match only if the study program was completed and the applicant obtained a degree. The minimum level of education required to get the job is also called a “reservation educational level,” indicating the level below which employers would not hire any school leaver (Verhaest and Omey 2006). This level may differ from the level of education required to do the job (i.e., to perform the job tasks reasonably well). The implications of this operationalization should be kept in mind while interpreting the results. Employers could specify the entry requirement in an open format or click the don’t know option in alternative.⁵

The measurement of horizontal mismatches was straightforward: Applicants with a background in informatics were classified as horizontally matched; those with a background in economics or social science were regarded as horizontally mismatched. By crossing information on vertical and horizontal matches, I distinguished between six groups of applicants: (1) undereducated and horizontally mismatched, (2) only undereducated, (3) only vertically matched, (4) both vertically and horizontally matched, (5) overeducated and horizontally mismatched, and (6) only overeducated. Figure 1 provides an example and illustrates how I constructed the indicator of mismatch.

Figure 1.

Construction of the type of (mis)match variable.

This operationalization improves on three alternative approaches that are commonly used in the overeducation literature to measure job requirements, namely, employee self-assessments, job analysis, and realized matches. Self-assessments are prone to bias: Employees may lack information on entry requirements, and their perceptions of job content, work autonomy, and even skill and qualification requirements may differ, sometimes to a large degree, from employers’ perceptions (Burchell et al. 1994; Green and James 2003). Compared to measurements based on job analysis, my approach does not assume a fixed level of required education for all the jobs that belong to the same occupational group. This level of detail is necessary as previous studies show that both job tasks (Autor and Handel 2013) and schooling requirements (Halaby 1994) also vary within occupations. Compared to the realized matches (or statistical) approach, I do not rely on an arbitrary cut-off point for comparison, such as the mean or median level of educational attainment in a given occupation, but on employers’ reported hiring floors. This way, entry requirements are more tightly linked to the organizational context.

Note that my indicator of vertical matching refers only to applicants’ level of education and not their skills. Individuals with the same education level can still vary in their literacy, numeracy, or vocational skills (see e.g., Allen and van der Velden 2001). If skills are not controlled for, what is taken at face value as genuine overeducation may in fact mask unobserved skill heterogeneity. The design of the vignettes addresses this concern: Employers observed not only job applicants’ schooling level but also their grade point average, previous work experience, and participation in internships (which is treated as a proxy for on-the-job training), and they were instructed at the beginning of the simulation that computer skills were comparable for all applicants. Thus, skill heterogeneity is ruled out by design.

After creating the measurement of mismatches, I compared the incidence of overeducation at three different stages of the hiring process: (1) applicants’ initial rating, (2) the shortlist, and (3) invitation to a job interview. The software automatically assigned the five applicants who received the highest ratings to the shortlist. Employers were then asked to rank the shortlisted applicants in order of preference, assigning the applicant they would most likely invite to a job interview to the first rank. This operationalization recognizes the multistaged nature of the hiring process, in which screening decisions are largely based on information reported in résumés, before employers or other members of the personnel department have met the applicant. Situational factors (e.g., gut feelings, applicants’ attire or demeanor) or interpersonal and presentation skills, which cannot easily be captured in a vignette study, are unlikely to influence employers’ assessments before the interview stage.

Results

Within-job Variation in Entry Requirements

Employers’ answers to the questionnaire reveal that entry requirements vary a great deal across organizations. Interestingly, the variation is more pronounced in England than in the Netherlands. In general, a bachelor’s degree was by far the most frequently mentioned level of required education, especially in the Netherlands, where 9 in 10 employers indicated that a degree from a vocational college was the minimum entry requirement for software engineer jobs. For business consultant jobs, 1 in 4 employers required a master’s degree; little more than 10 percent required a qualification from upper-secondary education. The latter was also sufficient for software engineer jobs for one-third of English employers. The larger variation in entry requirements in the English context is perhaps not surprising: For a nontrivial number of occupations, the official occupational classification of the UK (the Standard Occupational Classification) indicates there are no preset entry standards, or that entrants possess a variety of qualifications, or again that entry varies from employer to employer (Office of National Statistics 2010). This is consistent with the less coordinated nature of the English labor market.

Initial Ratings

The top panel of Table 2 reports the distribution of matched and mismatched applicants across all vignettes shown to employers. At the beginning of the hiring simulation, the distribution was virtually identical in the two countries, with both having 12 percent fully matched applicants. The other panels report the distribution of mismatches at the shortlist and interview stages, that is, after employers have screened the applicants. In the Netherlands, the proportion of full matches more than doubles from the initial pool to the shortlist. By contrast, it increases only slightly in England. In both countries, but particularly in the Netherlands, the proportion of undereducated and horizontally mismatched applicants is drastically reduced. We find a somewhat similar pattern when comparing the shortlist and interview stages. The proportion of applicants who are only vertically matched (i.e., their level of education meets the job requirements, but their field of study does not match the occupation) shrinks considerably in the Netherlands but not in England.

Table 2.

The Incidence of Job–Education Vertical and Horizontal (Mis)Matches (Percentages).

	England		Netherlands
	N	Percent	N	Percent
Initial applicant pool
Undereducated and horizontally mismatched	172	28.10	213	31.14
Only undereducated	80	13.07	103	15.06
Only vertically matched	138	22.55	144	21.05
Vertically and horizontally matched	75	12.25	83	12.13
Overeducated and horizontally mismatched	98	16.01	99	14.47
Only overeducated	49	8.01	42	6.14
Total	612	100.00	684	100.00
Shortlist (five top-rated applicants)
Undereducated and horizontally mismatched	26	15.48	14	7.37
Only undereducated	20	11.90	30	15.79
Only vertically matched	34	20.23	33	17.37
Vertically and horizontally matched	28	16.67	54	28.42
Overeducated and horizontally mismatched	30	17.86	23	12.10
Only overeducated	30	17.86	36	18.95
Total	168	100.00	190	100.00
Invitation to a job interview (first rank)
Undereducated and horizontally mismatched	3	8.82	1	2.63
Only undereducated	2	5.89	1	2.63
Only vertically matched	6	17.64	1	2.63
Vertically and horizontally matched	4	11.76	18	47.37
Overeducated and horizontally mismatched	7	20.59	2	5.26
Only overeducated	12	35.29	15	39.47
Total	34	100.00	38	100.00

Note: In England, some respondents indicated degree without specifying whether they meant a bachelor’s or a master’s qualification. In these cases, I assigned them the entry requirement of a bachelor’s degree. I assigned employers who did not mention any entry requirement to the lowest possible qualification, secondary education.

Figure 2 plots results from linear regression models, using the ratings as the dependent variable. Table A2 in the Appendix, available in the online version of the article, shows the full models. All models include employer fixed effects to restrict the focus to the rating of applicants within each applicant pool. The independent variables of interest are the mismatch dummies; full matches are the reference category. Additional human capital indicators (work experience, on-the-job training, and grades) and the other characteristics mentioned in the vignettes (gender, study delays, and extracurricular activities) are added as controls but not shown in the figure.

Figure 2.

Employers’ ratings of matched and mismatched applicants

The figure shows that Dutch employers applied fairly rigid hiring floors and screened out applicants who did not reach the minimum level of required education. English employers were somewhat more lenient. They barely differentiated between fully matched applicants and overeducated applicants with a horizontal mismatch (the difference is negligible and not statistically significant). This confirms that only in England can overeducation compensate for a lack of occupation-specific training. In the Netherlands, a horizontal mismatch cannot be overcome simply with more education: Overeducated but horizontally mismatched applicants are at a significant disadvantage compared to fully matched applicants. The mechanism assumed by job competition theory holds up only for applicants who are overeducated but horizontally matched. Somewhat surprisingly, a surplus of schooling in a relevant field of study does pay off in the Netherlands. One possible explanation is that employers expect overeducated applicants who are horizontally matched to possess more occupation-specific skills.

For a more straightforward reading of the regression coefficients, Table 3 shows the conditional mean ratings for matched and mismatched applicants after adjusting for all other covariates included in the model. In the Netherlands, the undereducation penalty was particularly severe if coupled with a horizontal mismatch: Compared to fully matched applicants, the ratings of undereducated applicants without a background in informatics decreased by nearly 80 percent, from 56 to 12. In comparison, the relative decrease in ratings in England was only half as large, from 40 to 24.

Table 3.

Predicted Probabilities for Matched and Mismatched Applicants, Given All Other Variables Included in the Vignette.

	England		Netherlands
	Margins	SE	Margins	SE
Ratings likelihood to hire the applicant
Undereducated and horizontally mismatched	23.70	(1.42)	12.31	(1.44)
Only undereducated	29.53	(2.05)	23.01	(2.02)
Only vertically matched	33.68	(1.53)	27.66	(1.73)
Vertically and horizontally matched	40.19	(2.07)	56.08	(2.27)
Overeducated and horizontally mismatched	37.75	(1.90)	29.21	(2.14)
Only overeducated	56.59	(2.60)	66.00	(3.19)
Ratings likelihood the applicant would be easy to train
Undereducated and horizontally mismatched	30.04	(1.40)	17.90	(1.79)
Only undereducated	33.72	(2.01)	31.25	(2.09)
Only vertically matched	38.64	(1.50)	33.80	(1.79)
Vertically and horizontally matched	46.66	(2.04)	61.58	(2.35)
Overeducated and horizontally mismatched	44.02	(1.87)	37.06	(2.21)
Only overeducated	59.01	(2.56)	69.67	(3.30)
Likelihood to be in the shortlist (five top-rated applicants)
Undereducated and horizontally mismatched	0.37	(0.17)	0.34	(0.15)
Only undereducated	0.55	(0.22)	0.74	(0.16)
Only vertically matched	0.61	(0.19)	0.73	(0.16)
Vertically and horizontally matched	0.76	(0.17)	0.93	(0.06)
Overeducated and horizontally mismatched	0.74	(0.16)	0.74	(0.17)
Only overeducated	0.90	(0.08)	0.98	(0.02)

Note: Predicted probabilities for matched and mismatched applicants. For the analysis of ratings, the statistics are computed from the models shown in Table A2 (without the linkage control) in the Appendix, available in the online version of the article. For the analysis of the shortlisting stage, the statistics are computed from the model shown in Table A3 in the Appendix.

Table A5 in the Appendix, available in the online version of the article, shows a pooled model with interaction effects between the mismatch dummies and a country dummy. Differences are statistically significant. The patterns of results are stable for both trainability and likelihood of being hired. As sensitivity checks, I ran random effects models with controls for firm size to ensure that cross-national differences are not driven by employers from different firms having different propensities to participate in the study (results are stable).

In the survey, employers also indicated their involvement in stable collaborations or partnerships with schools or universities. All Dutch respondents were involved in some kind of collaboration at the secondary or tertiary level. In contrast, 30 percent of English employers reported no collaboration of this sort. I gathered this information for each level of education mentioned in the vignettes; this allowed me to identify applicants who had a type of qualification that was reasonably familiar to the employer. Consistent with the strongly coordinated nature of the Dutch education system, the penalty for vertical mismatches in the Netherlands was somewhat reduced after introducing a control for linkages in the model. Note that informatics is an applied subject with a medium level of occupational specificity: the alignment between educational content and the labor demand is stronger than in academic subjects such as arts or social sciences, but it is weaker than in professional programs that grant a license to practice, like law or medicine. In applied sciences, coordination occurs through formal and informal mechanisms of the type captured by the linkage indicator; this coordination is more influential in the Dutch context.

The controls added in the models for other aspects of human capital, such as grades, work experience, and on-the-job training, matter only in England. In Britain, the competition for jobs is based on a ranking of candidates’ relative performance, signaled through grades and other measurable markers of achievement. In the Netherlands, job allocation largely depends on the match between job requirements and candidates’ credentials (for a more elaborated discussion, see Di Stasio and van de Werfhorst 2016). Perhaps surprisingly in light of the gender segregation in the IT sector, the coefficient for gender is not significant in any of the models (results available on request).

The Shortlist and Invitation to a Job Interview

I estimated applicants’ likelihood of being included in the shortlist with conditional logistic regressions (Figure 3). At the shortlist stage, employers used largely the same criteria as for the initial ratings, partly due to the study design, which automatically assigned the five top-rated applicants to the shortlist. Again, we see that overeducation does not compensate for a lack of occupation-specific training in the Netherlands. Dutch employers also applied very rigid hiring floors: The probability that undereducated applicants without occupation-specific training are shortlisted is negligible (Table 3).

Figure 3.

The shortlisting and ranking of matched and mismatched applicants.

The five shortlisted applicants were shown again to employers. Employers ranked them according to the likelihood they would be invited to a job interview (in England, the number of ranked vignettes at the interview stage is 168 instead of 170 because information from 2 vignettes is missing). Of the Dutch applicants who were invited to a job interview, nearly half were fully matched (18 out of 38, or 47 percent). In England, only 4 out of 34 applicants (12 percent) were fully matched. Figure 3 plots the coefficients obtained from rank-ordered logistic regressions, the appropriate model when the dependent variable refers to ranked alternatives. At the interview stage, country differences are largely in line with expectations. Overeducation alone in the absence of a horizontal match does not pay off and in the Dutch context is even penalized.

One may argue that this penalty is in fact a field of study effect rather than a negative effect of horizontal mismatch. Different fields of study are differently valued in the labor market, possibly because they vary in occupational focus, the degree of sorting by ability, or the size of the available supply of graduates. If so, cross-national differences may be due to the larger role of fields of study in processes of occupational allocation in the Netherlands. To rule out this alternative explanation, I split the analyses by field of study and compared horizontally matched applicants (i.e., applicants with a background in informatics) with economics and social science graduates separately. Although social science is often associated with limited employment opportunities, the field of economics, business, and management generally leads to high earnings and relatively good labor market returns. Figure A1 (available in the online version of the article) shows that economics is indeed a stronger field of study than social science (a background in economics is generally less penalized than a background in social science). Still, for both fields of study, employers’ assessments of mismatched applicants vary across countries in line with expectations. This suggests that cross-national differences are not simply due to stronger stratification by fields of study in the Dutch context.

The full models for the shortlist and the interview stages can be found in Tables A3 and A4 in the Appendix, available in the online version of the article. Again, a formal test of country differences is reported in Table A5 in the Appendix.

Discussion and Conclusions

Employers’ preferences affect employment opportunities on the supply side. Whether employers are willing to accept mismatched applicants or try to avoid them at all cost has important implications for the functioning of the labor market. My contribution to the overeducation literature is threefold.

First, I focused on the demand side of the labor market, which empirical research on overeducation has largely overlooked. This is unfortunate as the job competition model (Thurow 1975) gives central place to employers and the hiring process. To redress this neglect, I collected data from a sample of human resource professionals in England and the Netherlands and studied how they ranked applicants competing for a hypothetical job opening. The shift in focus from employees to applicant pools was a necessary step to properly test the job competition model.

Second, I analyzed vertical and horizontal (mis)matches jointly. This more fine-grained classification is meaningful: Overeducated applicants were indeed placed ahead in labor queues—as the job competition model predicts—but only if their field of specialization matched the occupational domain of the job. Applicants with additional years of education but lacking occupation-specific training did not experience any advantage over applicants with fully matched qualifications, and they were strongly penalized in the Dutch context. This important qualifier would go unnoticed in analyses that focus only on vertical mismatches.

Third, these findings draw our attention to the context of educational credentialing. Institutions structure the acquisition of skills during formal education and provide employers with a shared understanding of what is exchanged during the hiring process—whether certified and occupation-specific skills (Netherlands) or the promise of future potential and the willingness to acquire such skills directly on the job (England). In the Netherlands, where the education system is highly stratified and has a strong occupational focus, additional years of schooling cannot compensate for a lack of occupation-specific training. Furthermore, undereducated applicants encounter rigid barriers in the strongly coordinated Dutch labor market, and horizontally mismatched applicants are screened out of the applicant pool even if their level of education is appropriate for the job. In England, where the education system is less stratified and only loosely coupled with workplace training, employers are more willing to adjust their hiring standards. The adjustment is not only upward or downward but also lateral.

Admittedly, I focused on only one economic sector that was strategically chosen to bring into sharp relief the role of the national context. Although this choice precludes any generalization to other sectors or countries, a few considerations on the scope conditions of my general argument can still be made. For example, national institutions should play a larger role when employers are recruiting for lower-level bureaucratic positions (Brown 2001). To enter these positions, standardized degrees earned in the formal education system provide a guarantee of technical competence, especially in countries like the Netherlands, where curricula have a strong occupational orientation. By contrast, recruitment in professional and managerial labor markets is probably less dependent on national education systems. Moreover, the impact of national institutions on job–education matching likely varies across fields of study, depending on their degree of credential closure (DiPrete et al. forthcoming). Pathways into strongly closed occupations, such as law or medicine, where supply restrictions are regulated by professional groups, are likely to be comparable across countries. The national context should be more influential in the skilled trades, where employers have to comply with processes of labor market coordination involving unions and the state.

Finally, this study targeted only labor market entrants. This is because (over)education is certainly more important to employers when work experience is limited and alternative indicators of skills, trainability, or performance are not available. Whether initial mismatches can be corrected over the course of one’s career and whether the scarring effect of mismatched employment histories is more severe in some countries than in others are interesting avenues for future research.

One limitation of this study is that applicants were fictitious, and employers may have been affected by the artificiality of the hiring simulation. Although a recent validation study shows that respondents’ evaluations of profiles described in vignettes closely align with the evaluation of similar profiles in real life (Hainmueller, Hangartner, and Yamamoto 2015), future studies may try to replicate my findings in real employment contexts, for example, using matched pairs of résumés. Recently, two such studies in the United States found that applicants with a history of underemployment had a significantly lower callback rate than did applicants who were adequately employed (Nunley et al. 2016; Pedulla 2016). In both studies, however, the underemployed applicants in the matched pairs had a job below their level of education and in an irrelevant industry; it is unclear if the scarring effects of underemployment were due to overeducation, horizontal mismatch, or both. Résumé audits with a more complex design are necessary to answer this question.

Note, too, that in factorial surveys the supply of applicants is fixed. The random assignment of variables to vignettes and vignettes to respondents is necessary to isolate the actions of screeners. This is an important advantage of factorial surveys over observational studies because the effects of interest can be estimated without bias (e.g., Jasso 2006; Wallander 2009). In real labor markets, self-selection processes on the supply side determine the composition of applicant pools, and employers may adjust their hiring standards based on the available supply of candidates. A more realistic simulation would ideally model supply- and demand-side processes simultaneously.

To conclude, this study contributes to the further development of credentialism theory. Some scholars have warned that the irrational preferences of employers, keen on raising hiring standards regardless of the technical demands of workplaces, would inevitably lead to credential inflation (Berg 1971). At the very least, I showed that these concerns are misplaced in the strongly coordinated labor market of the Netherlands. National institutions set limits to the possibility that employers’ demands for educational credentials ratchet up in response to a supply of overeducated applicants. There is also an alternative interpretation to the fact that overeducation in an occupation-specific field was rewarded in both national contexts: Namely, credential inflation may increasingly be occurring within specific occupational domains, resulting in “occupation-specific credentialism.”⁶ In other words, employers do not simply look for more education but for more education of a specific kind, and they prefer overeducated but horizontally matched workers. Future research may look at this specific type of credentialism in more detail.

In light of these findings, Bills’s (2016:4) recent suggestion to shift measurements away from queues and toward hierarchies seems warranted: “Unlike queues, which imply a fairly simple vertical ranking, hierarchies can be significantly more complex systems, simultaneously containing both vertical and horizontal dimensions.” The fairly rigid hiring floors and ceilings applied by Dutch employers and their reluctance to hire overeducated applicants lacking occupation-specific training square remarkably well with this suggestion.

Footnotes

Acknowledgements

The author gratefully acknowledges funding support from the Dutch Organization for Scientific Research (NWO) through a VIDI grant assigned to Prof. Herman G. van de Werfhorst (grant number 452–07–002) and a grant from the Programme Council for Fundamental Research of the Netherlands Initiative for Education Research (grant number 411–10–920). The data collection took place when the author was affiliated to the University of Amsterdam, Amsterdam Institute for Social Science Research (AISSR). Earlier drafts of this article were presented during the Demosoc Seminar at Pompeu Fabra University (Barcelona, Spain), the 2014 Spring Meeting of Research Committee 28 of the International Sociological Association (Budapest, Hungary), the External Economics Seminar at Ghent University, and the departmental seminar of the Department of Sociology of the University of Bern. The author is grateful for the useful comments received from the participants.

Supplemental Material

The Appendices are available in the online version of the article.

Notes

Author Biography

Valentina Di Stasio, PhD, is a postdoctoral fellow in Sociology at Nuffield College, University of Oxford. She obtained her doctoral degree at the University of Amsterdam in 2014. This study is based on her doctoral research, which received the “Best Dissertation of the Year” award from the European Consortium for Sociological Research (ECSR). Her work is also published in Social Forces, European Sociological Review, and Annual Review of Sociology. Currently, Valentina is working on a cross-national field experiment on ethnic discrimination in the labor market.

References

Adams

Tracy L.

Demaiter

Erin I.

2008. “Skill, Education and Credentials in the New Economy: The Case of Information Technology Workers.” Work, Employment and Society 22(2):351–62.

Allen

Jim

van der Velden

Rolf

. 2001. “Educational Mismatches Versus Skill Mismatches: Effects on Wages, Job Satisfaction, and On-the-job Search.” Oxford Economic Papers 53(3):434–52.

Athey

Timothy R.

Hautaluoma

Jacob E.

1994. “Effects of Applicant Overeducation, Job Status, and Job Gender Stereotype on Employment Decisions.” The Journal of Social Psychology 134(4):439–52.

Autor

David H.

Handel

Michael J.

2013. “Putting Tasks to the Test: Human Capital, Job Tasks, and Wages.” Journal of Labor Economics 31(S1):S59–96.

Berg

Ivar E.

1971. Education and Jobs: The Great Training Robbery. Boston: Beacon.

Bills

David B.

1992. “The Mutability of Educational Credentials as Hiring Criteria: How Employers Evaluate Atypically Highly Credentialed Job Candidates.” Work and Occupations 19(1):79–95.

Bills

David B.

2016. “Congested Credentials: The Material and Positional Economies of Schooling.” Research in Social Stratification and Mobility 43:65–70.

Bills

David B.

Brown

David K.

2011. “New Directions in Educational Credentialism.” Research in Social Stratification and Mobility 29(1):1–4.

Bol

Thijs

Weeden

Kim A.

2015. “Occupational Closure and Wage Inequality in Germany and the United Kingdom.” European Sociological Review 31(3):354–69.

10.

Brown

David K.

2001. “The Social Sources of Educational Credentialism: Status Cultures, Labor Markets, and Organizations.” Sociology of Education 74(Extra Issue):19–34.

11.

Burchell

Brendan

Elliot

Jane

Rubery

Jill

Wilkinson.

Frank

1994. “Job Content and Skill: Managers’ and Employees’ Perspectives.” In Skill and Occupational Change, edited by Penn

Rose

Rubery

Oxford, UK: Oxford University Press.

12.

Collins

Randall

1979. The Credential Society: An Historical Sociology of Education and Stratification. New York: Academic Press.

13.

Di Stasio

Valentina

van de Werfhorst

Herman G.

2016. “Why Does Education Matter to Employers in Different Institutional Contexts? A Vignette Study in England and the Netherlands.” Social Forces 95(1):77–106.

14.

DiPrete

Thomas A.

Bol

Thijs

Ciocca

Christina

van de Werfhorst

Herman G.

Forthcoming. “School-to-work Linkages in the United States, Germany and France.” American Journal of Sociology.

15.

Fernandez

Roberto M.

Sosa

M. Lourdes

. 2005. “Gendering the Job: Networks and Recruitment at a Call Center.” American Journal of Sociology 111(3):859–904.

16.

Fernandez

Roberto M.

Weinberg

Nancy

. 1997. “Sifting and Sorting: Personal Contacts and Hiring in a Retail Bank.” American Sociological Review 62(6):883–902.

17.

Green

Francis

James

Donna

2003. “Assessing Skills and Autonomy: The Job Holder Versus the Line Manager.” Human Resource Management Journal 13(1):63–77.

18.

Halaby

Charles N.

1994. “Overeducation and Skill Mismatch.” Sociology of Education 67(1):47–59.

19.

Hainmueller

Jens

Hangartner

Dominik

Yamamoto

Teppei

2015. “Validating Vignette and Conjoint Survey Experiments against Real-world Behavior.” Proceedings of the National Academy of Sciences of the United States of America 112(8):2395–400.

20.

Iannelli

Cristina

Raffe

David

. 2007. “Vocational Upper-secondary Education and the Transition from School.” European Sociological Review 23(1):49–63.

21.

Jasso

Guillermina.

2006. “Factorial Survey Methods for Studying Beliefs and Judgments.” Sociological Methods & Research 34:334–423.

22.

Kerckhoff

Alan C.

2001. “Education and Social Stratification Processes in Comparative Perspective.” Sociology of Education 74:3–18.

23.

Kulkarni

Mukta

Lengnick-Hall

Mark L.

Martinez.

Patricia G.

2015. “Overqualification, Mismatched Qualification, and Hiring Decisions: Perceptions of Employers.” Personnel Review 44(4):529–49.

24.

Levels

Mark

van der Velden

Rolf

Di Stasio

Valentina

. 2014. “From School to Fitting Work: How Education-to- job Matching of European School Leavers Is Related to Educational System Characteristics.” Acta Sociologica 57(4):341–61.

25.

Levy

Jack S.

2008. “Case Studies: Types, Design, and Logics of Inference.” Conflict Management and Peace Science 25(1):1–18.

26.

Müller

Walter.

2005. “Education and Youth Integration into European Labour Markets.” International Journal of Comparative Sociology 46(5–6):461–85.

27.

Walter

Müller

Gangl

Markus

, ed. 2003. Transitions from Education to Work in Europe: The Integration of Youth into EU Labour Markets. Oxford, UK: Oxford University Press.

28.

Nunley

John M.

Pugh

Adam

Romero

Nicholas

Seals

R. Alan

. 2016. “The Effects of Unemployment and Underemployment on Employment Opportunities: Results from a Correspondence Audit of the Labor Market for College Graduates.” ILR Review, 1–28. doi:10.1177/0019793916654686

29.

Oesch

Daniel

Rodrguez-Mens

Jorge

. 2011. “Upgrading or Polarization? Occupational Change in Britain, Germany, Spain and Switzerland, 1990–2008.” Socio-Economic Review 9(3):503–31.

30.

Office for National Statistics. 2010. Standard Occupational Classification 2010. Volume 1—Structure and Descriptions of Unit Groups. Basingstoke: Palgrave Macmillan.

31.

Pedulla

David S.

2016. “Penalized or Protected? Gender and the Consequences of Nonstandard and Mismatched Employment Histories.” American Sociological Review 81(2):262–89.

32.

Raffe

David

Brannen

Karen

Croxford

Linda

Martin.

Chris

1999. “Comparing England, Scotland, Wales and Northern Ireland: The Case for ‘Home Internationals’ in Comparative Research.” Comparative Education 35(1):9–25.

33.

Sauer

Carsten

Auspurg

Katrin

Hinz

Thomas

Liebig.

Stefan

2011. “The Application of Factorial Surveys in General Population Samples: The Effects of Respondent Age and Education on Response Times and Response Consistency.” Survey Research Methods 5(3):89–102.

34.

Shen

Kailing

Kuhn

Peter

. 2012. “Do Chinese Employers Avoid Hiring Overqualified Workers? Evidence from an Internet Job Board.” IZA Discussion Paper No. 6848. Forschungsinstitut zur Zukunft der Arbeit Institute for the Study of Labor.

35.

Sicherman

Nachum.

1991. ““Overeducation” in the Labor Market.” Journal of Labor Economics 9(2):101–22.

36.

Solga

Heike

Konietzka.

Dirk

1999. “Occupational Matching and Social Stratification: Theoretical Insights and Empirical Observations Taken from a German-German Comparison.” European Sociological Review 15(1):25–47.

37.

Thurow

Lester C.

1975. Generating Inequality. New York: Basic Books.

38.

van Beek

Krijn W. H.

Koopmans

Carl C.

van Praag

Bernard M. S.

1997. “Shopping at the Labour Market: A Real Tale of Fiction.” European Economic Review 41(2):295–317.

39.

van de Werfhorst

Herman G.

2011. “Skills, Positional Good or Social Closure? The Role of Education across Structural-institutional Labour Market Settings.” Journal of Education and Work 24(5):521–48.

40.

Van Noy

Michelle

Jacobs

James

. 2012. “Employer Perceptions of Associate Degrees in Local Labor Markets: A Case Study of the Employment of Information Technology Technicians in Detroit and Seattle.” CCRC Working Paper No. 39. Accessed February 10, 2017 (http://ccrc.tc.columbia.edu/publications/employer-perceptions-associate-degrees.html).

41.

Verhaest

Dieter

Bogaert

Elene

Dereymaeker

Jeroen

Mestdagh

Laura

Baert

Stijn

. 2016. “Crowding Out in the Labour Market: Do Employers Lend a Hand?” IZA Discussion Paper No. 9654. Accessed February 10, 2017 (http://ftp.iza.org/dp9654.pdf).

42.

Verhaest

Dieter

Omey.

Eddy

2006. “Discriminating between Alternative Measures of Over-education.” Applied Economics 38(18):2113–20.

43.

Verhaest

Dieter

Sellami

Sana

van der Velden

Rolf

. 2015. “Differences in Horizontal and Vertical Mismatches across Countries and Fields of Study.” International Labour Review. Accessed February 10, 2017 (http://onlinelibrary.wiley.com/doi/10.1111/j.1564-913X.2015.00031.x/abstract).

44.

Verhaest

Dieter

van der Velden

Rolf

. 2013. “Cross- country Differences in Graduate Overeducation.” European Sociological Review 29(3):642–53.

45.

Wallander

Lisa.

2009. “25 Years of Factorial Surveys in Sociology: A Review.” Social Science Research 38(3):505–20.

46.

Weeden

Kim A.

2002. “Why Do Some Occupations Pay More than Others? Social Closure and Earnings Inequality in the United States.” American Journal of Sociology 108(1):55–101.

47.

Wolbers

Maarten H. J.

2003. “Job Mismatches and Their Labour-Market Effects among School-leavers in Europe.” European Sociological Review 19(3):249–66.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

3.51 MB