Sage Journals: Discover world-class research

Abstract

Evidence on gender inequality in the labor market is extensive. However, little is known about the potential role of overeducation and horizontal mismatch in explaining women’s labor-market disadvantages. We draw on recent data from the Eurograduate pilot survey to investigate the role of overeducation, field-of-study mismatch and field-specific overskilling for gender gaps in labor income in the European landscape. We found considerable variations in the extent of both gender earning gaps (GEGs) and wage gaps (GWGs) across countries. However, our decomposition analyses show that neither overeducation nor horizontal mismatch contribute to explaining these gender gaps. The lack of mediation seems related to either the absence of gender differences in overeducation and horizontal mismatch, or to the nonexistence of income penalties associated with the mismatch.

Keywords

Eurograduate pilot survey gender earnings gap gender wage gap higher education graduates overeducation education mismatch horizontal mismatch

Introduction

Gender inequality in labor income is a core topic in social stratification research because it calls into question the meritocratic principles of modern societies (Grusky, 2014). Women’s disadvantage has persisted over the last 30 years in most Western countries (Bar-Haim et al., 2018). While occupational segregation plays an increasingly important role in explaining the gender gap in labor income, the role of gender disparities in educational attainment has declined across cohorts due to the educational expansion.

As a result of the new millennium higher education expansion (e.g. Marginson, 2016), more than half of the population between the age of 25 through 34 years completes higher education in nine out of the 35 OECD countries (OECD, 2018). Women have even overtaken men in educational attainment in most high-income countries (Sauer and Van Kerm, 2021). And yet a considerable number of single-country studies (e.g. Francesconi and Parey, 2018; Leuze and Strauß, 2009; Goldan, 2021; Passaretta and Triventi, 2021) and fewer comparative studies (e.g. Garcia-Aracil, 2007; Triventi, 2013) consistently found gender gaps in labor income even among higher education graduates. This persisting gap is remarkable, especially if we consider that the highly educated are expected to have meritocratic opportunities in the labor market.

Explaining gender gaps in income, the existing literature points toward the importance of individual- (e.g. Achatz et al., 2005; Weichselbaumer and Winter-Ebmer, 2005), occupational- (e.g. Blau and Kahn, 2017), and country-level (e.g. Budig et al., 2010) determinants. Working hours stand out as one of the major determinants of gender gaps in earnings. Although the importance of working time varies substantially across countries, women’s lower work intensity represents an important driver of women’s disadvantage even among the highly qualified. As a result, the gender gap in monthly earnings (GEG, hereafter) is generally higher than the gender gap in hourly wages (GWG, hereafter). This holds true especially in countries where women’s labor market participation is lower (e.g. Triventi 2013).

Our paper considers both the GEG and the GWG as two complementary aspects of gender inequality in labor income. On the one hand, monthly earnings represent the actual economic resources made available from labor. On the other hand, hourly wages are closer measures of individuals’ productivity and earning potential. On the other hand, monthly earnings represent the actual economic resources made available from labor.

To date, only a few studies investigate the role of overeducation and horizontal mismatch for gender gaps in labor income (e.g. Addison et al., 2020). While there is evidence on the effect of education–job mismatches for gender inequality in labor income in the Anglo-American contexts (e.g. Addison et al., 2020; McGuinness and Sloane, 2011), evidence is scant in Europe (e.g. Allen and van der Velden, 2001; Christl and Köppl–Turyna, 2020). Moreover, to the best of our knowledge, European evidence is based on single-case studies that draw on a limited set of countries and load on different data, methodologies, and overall study designs.

Against this backdrop, we address two sets of research questions:

(a) Do GEGs and GWGs exist among recent cohorts of higher education graduates in Europe? Do GEGs and GWGs vary substantially across European countries?

(b) To what extent do overeducation and horizontal mismatch contribute to explaining GEGs and GWGs? Does the importance of overeducation and of horizontal mismatch differ across countries?

To address these research questions, we draw on brand-new comparative data—the Eurograduate pilot survey—that collected information for eight European countries: Austria, Czech Republic, Greece, Germany,¹ Croatia, Malta, Lithuania, and Norway (Meng et al., 2020; Muehleck et al., 2020).² We estimate and decompose gender gaps in labor income separately in each country. By mapping cross-country differences in those countries, our potential contribution to the literature is twofold.

First, we draw on a recent and unique data set for higher education graduates. Our internationally comparative perspective adds to existing evidence that mostly consists of single-country studies (see Table 1). Furthermore, the data offers the unique opportunity to describe and compare gender inequality in labor income across countries that have received little attention so far (mostly due to the unavailability of data for these countries). Second, we focus on the potential role of overeducation and horizontal mismatch in explaining gender gaps in labor income among higher education graduates within each country.

Table 1.

Literature review—GEG and GWG among higher education graduates in European countries.

Study	Country/countries	Population	Data	Dependent variable(s)	Methods	Raw gap	Unexplained/residual gap	Determinants
Single-country studies
Goldan (2021)	Germany	Doctoral graduates in year 2014	DZHW “PhD Panel” 2014, graduates surveyed in 2015	GEG: (logarithm of) gross monthly earnings in € 5 years after graduation (including the monthly share of annual bonuses)	OLS	30.4%	10.6%	Human capital (with a special focus on scientific productivity), employment, working hours, socio-demographics
Francesconi and Parey (2018) ^a	Germany	Higher education graduates in years 1989, 1993, 1997, 2002, 2005, 2009, full-time workers only	DZHW “graduate surveys,” graduates from the corresponding 6 graduation cohorts surveyed 12–18 months after graduation	GEG: (logarithm of) monthly earnings, measured in 2001 Euros	OLS, decomposition^b	ALL	ALL	Personal characteristics, family background, human capital (university, grades, field of study)
						19.7%	8.4%
						1989: 24.5%	1989: 10.0%
						1993: 15.1%	1993: 9.5%
						1997: 18.6%	1997: 8.3%
						2001: 19.5%	2001: 7.8%
						2005: 23.5%	2005: 9.1%
						2009: 16.4%	2009: 5.0%
Leuze and Strauß (2014)	Germany	Higher education graduates in year 1997	DZHW “graduate survey,” graduates surveyed 12–18 months and 5 years after graduation	GWG: (logarithm of) gross hourly wages in € 1 and 5 years after graduation	Random effects panel regression	21.2/16.5%	6.2%	Human capital, female-dominated subjects, female-dominated occupations, employment, socio-demographics
Leuze and Strauß (2009)	Germany	Higher education graduates in year 1997, full-time workers only	DZHW “graduate survey,” graduates surveyed 12–18 months and 5 years after graduation	GEG: (logarithm of) gross monthly income 5 years after graduation (without variable salary components)	OLS, (Oaxaca-Blinder) decomposition	26.7%	10.8%	Human capital, female-dominated subjects, occupational segregation, employment, socio-demographics
Reimer and Schröder (2006)	Germany	Higher education graduates (ISCED 5A programmes), social sciences only, between winter semester 1994/1995 and summer semester 2000	“Mannheim University Social Sciences Graduate Survey,” all graduates surveyed in 2001/2002	GWG: (Logarithm of) hourly wage (in “Deutsche Mark”) at the beginning of the first job	OLS, (Blinder-Oaxaca) decomposition	4.7%	7.9%	Human capital, job search, economic sector, job characteristics, job attitudes, socio-demographics
Passaretta and Triventi (2021)	Italy	Doctoral graduates in years 2004 and 2008	“ISTAT Survey on Italian PhD Graduates, graduates from 2004 surveyed in 2009, graduates from 2008 surveyed in 2014	GEG: (Logarithm of) monthly earnings (adjusted for purchasing power parity)	OLS and quantile regression, (Kitagawa Oaxaca-Blinder) decomposition	16.0% (for GWG)	- - -	Human capital (with a special focus on scientific productivity), occupation, working hours, reconciliation
Passaretta and Triventi (2021)	Italy	Doctoral graduates in years 2004 and 2008		GWG: (logarithm of) hourly wages		16.0% (for GWG)	- - -
Piazzalunga (2018)	Italy	Higher education graduates (ISCED 5A programmes) in year 2007	“ISTAT Survey on University graduates”, graduates from 2007 surveyed in 2011	GWG: (logarithm of) hourly wages	OLS, Oaxaca-Blinder and quantile) decomposition	5.6%	5.5%	Socio-demographics, social background, human capital (field of study), employment
Amilon and Persson (2013)	Sweden	Doctoral graduates (1970 between 2004)	Swedish register data 2004	GEG: (logarithm of) annual earnings in 2004	OLS	- - -	14.5%	Experience, family, academic versus non-academic sector
Chevalier (2007)	United Kingdom	Higher education graduates, only full-time employed	UK graduate survey, graduates from 1995 surveyed in 1998	GEG: (logarithm of) annual gross earnings (category mid-points)	OLS, (Oaxaca-Blinder) decomposition	12.6%	- - -	Field of study, occupational characteristics, character traits, motivation, expectations
Two-country studies
Machin and Puhani (2003)	United Kingdom (UK) and Germany (DE)	University graduates	Labor force surveys in the UK and (western) Germany (“Mikrozensus”) in 1996	GWG: (logarithm of) gross wages (UK) and net wages (DE)	OLS, (Blinder-Oaxaca) decomposition	UK: 20.8% DE: 28.0%	UK: - - -	Field of study, age, occupational characteristics
Machin and Puhani (2003)	United Kingdom (UK) and Germany (DE)	University graduates		GWG: (logarithm of) gross wages (UK) and net wages (DE)	OLS, (Blinder-Oaxaca) decomposition	UK: 20.8% DE: 28.0%	DE: - - -	Field of study, age, occupational characteristics
Three-country studies
Figueiredo et al. (2015) ^a	Italy (IT), Portugal (PT) and Spain (ES)	Higher education graduates (ISCED 5A programmes) in year 2000	“Research into employment and professional FLEXibility (REFLEX),” graduates from 2000 surveyed in 2005/2006 + European Labour Force Survey (EULFS)	GWG. (logarithm of) hourly wages	OLS, (Blinder-Oaxaca) decomposition	ALL: 14.1%	ALL: - - -	Employment history (first job, unemployment), current employment, sociodemographics, education and skill mismatches
						IT: 13.6%	IT: - - -
						PT: 23.8%	PT: - - -
						ES: 11.0%	ES: - - -
Comparative studies
Triventi (2013)	11 European countries	Higher education graduates (ISCED 5A programmes) in year 2000	“Research into Employment and professional FLEXibility (REFLEX),” graduates from 2000 surveyed in 2005/2006	GEG: (Natural logarithm of) gross monthly earnings (adjusted for purchasing power parity)	OLS, (Blinder-Oaxaca) decomposition	ALL: 23%	ALL	Employment (26.3%), working hours (19.7%), human capital (9.8%), reconciliation (6.4%)
						AT: 30.0%	AT: 11.4%
						BE: - - -	BE: 03.6%
						CZ: - - -	CZ: 12.5%
						DE: 32.0%	DE: 09.9%
						ES: - - -	ES: 05.7%
						FI: - - -	FI: 02.2%
						FR: - - -	FR: 08.0%
						IT: - - -	IT: 07.2%
						NL: - - -	NL: 06.7%
						NO: - - -	NO: 02.0%
						UK:	UK: 04.3%
Garcia-Aracil (2007)	11 European countries	Higher education graduates in year 1995	“Careers after higher Education – a European research survey (CHEERS),” graduates from 1995 surveyed in 1999	Labor force participation rate, unemployment rate	OLS, (Oaxaca-Ranson) decomposition	ALL: 17.67%	- - -	Employment, working hours, human capital
Garcia-Aracil (2007)	11 European countries	Higher education graduates in year 1995		GEG: (Logarithm of) gross annual earnings (standardized by purchasing power parity)	OLS, (Oaxaca-Ranson) decomposition	ALL: 17.67%	- - -	Employment, working hours, human capital

Source: Own compilation.

Notes: This is not a systematic review. We only consider European studies that have been either available in English or German. In the case of multiple studies per country, we listed them chronologically by date of publication in descending order. We did not find single country studies for Austria, Croatia, Czech Republic, Greece, Lithuania, and Malta.

^aThe study reports log points; for purposes of simplicity, we approximate them to percentage points.

^bNo further specification on the decomposition method available.

- - - indicates that the exact numbers are not provided in the article.

Previous research

GEG and GWG among higher education graduates

Overall, higher education graduates should experience smaller gender gaps in labor income because they are a more homogeneous group in terms of skills and knowledge than the general population. Moreover, women’s increased educational attainment has been a key driver of educational expansion (Sauer et al., 2022). Thus, the advancement of women with respect to educational attainment at the tertiary level should attenuate or even compensate their income disadvantages (Bar-Haim et al., 2018; Blau and Kahn, 2017). However, research demonstrated that gender gaps in income exist among tertiary graduates (e.g. Garcia-Aracil, 2007; Triventi, 2013) and even among PhD holders (e.g. Goldan, 2021; Passaretta and Triventi, 2021).

Table 1 summarizes the current state of research on gender gaps in earnings and wages among tertiary graduates in Europe. The table makes apparent two drawbacks of the existing literature. First, studies are single-country studies focusing on Germany. The specificities of the German labor market and welfare state make it impossible to draw conclusions on other countries based on the German findings. Second, there is only little evidence for most Eastern and Southern European countries, let alone in internationally comparative perspectives—with two exceptions (Garcia-Aracil, 2007; Triventi, 2013).

Triventi (2013) shows remarkable variation in the size of the gender earning gaps among higher education graduates in Europe. Based on data from the “Reflex survey”³ comprising 11 European countries, he found that the raw GEG varied between 32% in Germany and below 15% in the United Kingdom and Belgium; the average across all countries considered was about 23%. These patterns are in line with earlier evidence on European higher education graduates from the same set of countries based on older data from the “Cheers survey” (Garcia-Aracil, 2007).⁴ Overall, evidence from these comparative studies align with OECD’s country-rankings based on the extent of GEGs in the general population (OECD, 2018).

To the best of our knowledge, these are the only two studies that focused on higher education graduates by leveraging comparable international data. However, findings refer to cohorts that graduated more than 20 years ago (2000 in Triventi, 2013 and 1995 in Garcia-Aracil, 2007), when the new millennium higher education expansion had just started. At the time, women’s higher education graduation rate was still lower than men’s in most high-income countries (Sauer et al., 2022). This calls into question whether the same patterns will replicate in recent cohorts of European graduates.

The two comparative studies also offer insights on the factors that explain women’s disadvantage. Working hours stood out as the single strongest determinant of the GEG. Employment characteristics, such as occupation or organization type, and work-family-reconciliation factors, such as care responsibilities, also contributed to explaining women’s disadvantage in many countries (Garcia-Aracil, 2007; Triventi, 2013). The most significant human capital factor pertaining to higher education was the field of study (for similar findings on relevance of field of study see e.g. Leuze and Strauß, 2009, 2014; Machin and Puhani, 2003). On the contrary, self-perceived foreign language proficiency, ICT skills, and innovation skills played a minor role in explaining the GEG (Triventi, 2013).

Garcia-Aracil (2007) included self-reported levels of competencies required on the job as central explanatory factors in her decomposition analysis. She found that, overall, foreign language proficiency and computer skills contributed to reduce GEGs, while planning, coordinating and organization skills increased the GEGs. This indicates that the type of skill is relevant when explaining gender gaps in labor income. Still, this evidence is silent about whether potential mismatches between the level and the type of skills acquired by graduates and required by employers might explain women’s disadvantages.

Education-job mismatch and labor income

Educational expansion is accompanied by increased job mismatch if it does not coincide with an upgrade of the occupational structure (Brynin and Longhi, 2009; Kracke et al., 2018; Levels et al., 2014; Sattinger and Hartog, 2013). Job mismatch may take two different forms. On the one hand, a vertical mismatch occurs if the formal qualification or skill level required on the job is different than the own qualification (i.e. over- or undereducation) or skills (over- or underskilling). On the other hand, a horizontal mismatch occurs when the field of study is not aligned with the job content (e.g. Allen et al., 2013; Allen and van der Velden, 2001).

While being matched (vertically or horizontally) generally results in a wage premium among tertiary graduates (Bol et al., 2019), being mismatched comes along with a penalty. More generally, the overeducated suffer a penalty when compared to matched workers with the same level of education (McGuinnes, 2006). A similar penalty is found for mismatches with regard to the level and types of skills (e.g. Brynin and Longhi, 2009; Kracke et al., 2018; Levels et al., 2014; Sattinger and Hartog, 2013). Vertical and horizontal mismatches may contribute to gender inequality in labor income of tertiary education graduates because the probability of experiencing a mismatch is not necessarily the same for men and women (e.g. Addison et al., 2020; Petö and Reizer, 2021).

Based on data from the Programme for the International Assessment of Adult Competencies survey (PIAAC), Petö and Reizer (2021) analyzed gender differentials in skill use in 12 countries and found that, even within the same occupation, women report a lower usage of numeracy, literacy, and ICT skills compared to men. At the same time, skills and skill-task matches lead to income premiums even within the same occupation. Similarly, a country study for Austria shows that a substantial part of the GWG can be explained by gender-specific endowment and returns to skills and job tasks (Christl and Köppl-Turyna, 2020). For women, the usage of planning, reading, and writing skills leads to wage premiums while the usage of numerical and ICT skills does so for men. A match between the observed skills level and the skills used on the job is associated with a wage premium which, however, does not vary between men and women (ibd.).

A related factor is the field of study because women tend to graduate from less specific fields, which increases the chance of mismatch. Salas-Velasco (2021) shows that, in Spain, graduates from Sports Sciences and Business are at higher risk of vertical mismatch while graduates from Social Work, Labor Relations and Political Science are more likely to experience both vertical and horizontal mismatch. A comparative study by Rossen et al. (2019) concludes that overeducation varies for men and women across fields of study.

Drawing on internationally comparative evidence, gendered patterns of overeducation and field of study completions are robust. However, the degree of overeducation among the highly qualified differs across countries, being highest in Austria, France, and Italy and lowest in countries such as Estonia (Rossen et al., 2019). However, although overeducation results in wage penalties, Boll and Leppin (2014) found that overeducation does not contribute to explain income differences between men and women.

Two recent country studies conducted with data from the US (Addison et al., 2020) and Australia (Mavromaras et al., 2013) provide more detailed examinations of the role of overeducation and skill mismatch for gender gaps in income. Addison et al. (2020) stress that gender differences in horizontal mismatch, measured by comparing objective information on occupation-specific skills with multiple dimensions of “actual” skills, exist particularly among the highly educated. Women with children in flexible work arrangements are the ones most affected by mismatches. Mavromaras et al. (2013) study overskilling in conjunction with overeducation among Australian graduates and find that both forms of mismatch are slightly lower for women than for men. However, men experience earnings’ penalties only when they are both overeducated and overskilled, while being either overeducated or overskilled was sufficient to suffer from a wage penalty for women.

Altogether, the existing literature suggests that overeducation and horizontal mismatch are potentially important factors in explaining why highly educated women still lag behind men when it comes to labor income. And yet overeducation or horizontal mismatch has not received much attention as drivers of gender inequality, especially not in an internationally comparative perspective. The few studies that examine GEG and GWG among higher education graduates in a comparative setting did not focus on the potential role of overeducation and horizontal mismatch (Garcia-Aracil, 2007; Triventi, 2013).

Analytical framework

We draw on a descriptive mediation framework (Baron and Kenny, 1986; Shrout and Bolger, 2002) connecting gender gaps in labor income with overeducation and horizontal mismatch (see Figure 1). The previous literature unequivocally points toward the existence of gender gaps in labor income in Western societies (e.g. Bar-Haim et al., 2018), even among the tertiary educated (Triventi, 2013). Previous research has also shown associations between gender and overeducation and horizontal mismatch—path B—(e.g. Addison et al., 2020), and wage penalties for mismatched workers—path C (e.g. Mavormaras et al., 2013).

Figure 1.

Analytical framework: the expected mediating role of overeducation and horizontal mismatch for GEG and GWG. Source: Own illustration.

Gender disparities in the probability of being mismatched may arise because of differential preferences, endowments, or institutional and normative constraints. The existence of a penalty for mismatched workers is driven by the fact that education and skills generally pay off on the labor market (Hanushek et al., 2015). Thus, being mismatched should consequently be accompanied by an income penalty (e.g. Allen and van der Velden, 2001; Kleibrink, 2014; Kracke et al., 2018; Levels et al., 2014).

Based on this evidence, we posit that mismatches between the own educational level and level/type of skills and the educational level and skills’ level required on the job may represent a channel for the formation of gender gaps in labor income. Hence, we expect to find penalties in labor income for highly educated women across countries, and that such penalties will be partly channeled via overeducation and horizontal mismatch. That said, we also expect to find a residual gender gap – path A – which suggests a direct association between gender and labor income that is not channeled via overeducation and horizontal mismatch. This residual gap does not only capture the role of potential discriminatory behavior on the job but comprises all supply and demand factors not explicitly considered in the analyses (e.g. productivity, preferences, or time and cultural constraints).

This paper aims to provide a nuanced description of the simple set of relationships depicted in Figure 1 for a recent cohort of graduates from countries which were hardly considered by the previous literature.

Data and methods

Eurograduate pilot survey

We use data from the Eurograduate pilot survey⁵ which comprises information on two cohorts of recent higher education graduates in eight European countries—Austria (AT), Czech Republic (CZ), Germany (DE), Croatia (HR), Greece (GR), Lithuania (LT), Malta (MT), and Norway (NO) (for detailed information see Meng et al., 2020; Muehleck et al., 2020). The survey was carried out in 2018. The two cohorts comprise graduates who obtained a tertiary degree (BA, MA) from either a private or a public higher education institution in 2012/13 and 2016/17, respectively. Thus, the more recent cohort was interviewed 1 year after graduation while the earlier cohort was interviewed 5 years after graduation.

The Eurograduate pilot survey laid the ground for a sustainable European-wide graduate survey, which will be continued with 17 European countries in 2022 (Eurograduate consortium, 2022). The survey is targeted toward mapping the impact that multidimensional experiences of European graduates during their time as students have on their professional lives and their lives as European citizens. The survey comprises information on graduates’ secondary school career, study experiences, national and international mobility; first and current employment situation, income, as well as background information on marital status, social and ethnic background, and social and political values and attitudes.

The eight countries were selected according to a stratum comprising small, mid-, and large countries from the eastern, southern, western, and northern part of Europe. We provide detailed information on the eight countries in Appendix A, including survey information (e.g. country samples, selections, and response rates), background information on higher education systems, the general labor market situation, the gender income gap in the general population, and basic statistics on the extent of overeducation and horizontal mismatch.

Sample

Among the two Eurograduate cohorts, we selected the recent cohort, which graduated in 2016/2017 and interviewed in 2018 (1 year after graduation) for our analyses (N = 5502). This restriction (−4884 observations) was necessary because the variables used to construct horizontal mismatch were only collected in 2018, that is 5 years after graduation for the older cohort (1 year after graduation for the recent cohort), thus making the comparison between the two cohorts difficult. We opted for the younger cohort because the older cohort that graduate 5 years ago could have obtained additional academic degrees in the meantime, which is problematic because field-specific overskilling is measured as a self-reported information that compares the skills acquired during study time with the skills required at the current job. Since we would relate the skills students obtained in a study program from which they graduated 5 years ago with the skill requirements at their current job, we decided to concentrate our analyses on the recent cohort of graduates.

We excluded Germany because of the high selectivity and the critically low sample size. We provide extensive background information on the country samples, selections, and response rates in Appendix A. We also restricted our sample to graduates below the age of 50 (−248 observations) and excluded graduates from short-cycle and PhD programs (−107 observations). Our sample includes employed graduates only (dependent or self-employment) (−4331 observations).⁶ Table 2 reports the sample size in each country after list-wise deletion based on the analytical variables.

Table 2.

Distribution of analytical variables by country and gender.

	AT		CZ		GR		HR		LT		MT		NO		Total
	m	F	M	f	M	F	m	F	m	f	M	F	m	f	m	f
Labor income and working time (means)
Earnings^a	2838	2295	1601	1115	1303	961	1269	945	1420	962	1710	1560	4466	3904	2175	1652
Wages^a	18	16	10	7	8	6	7	6	9	6	10	10	28	24	13	11
Working hours^b	43	38	44	43	45	39	44	43	45	43	46	42	41	40	44	41
Mismatch^c
Over-education	18	21	13	22	24	30	31	28	29	25	30	31	20	20	24	25
Under-education	20	11	14	14	21	22	11	6	10	11	12	8	9	6	13	9
Field mismatch	14	16	24	22	17	36	15	13	19	26	14	16	14	13	16	17
Over-skilling (field)	40	41	42	47	24	29	24	27	38	36	20	29	28	37	31	35
Under-skilling (field)	16	17	8	14	24	22	25	27	14	15	23	14	24	24	20	21
Educational level^c
BA	32	42	29	30	68	64	30	25	57	63	58	59	25	32	38	37
MA	68	58	71	70	32	36	70	75	43	37	42	41	75	68	62	63
Field of study^c
Arts and humanities	1	3	11	5	3	2	11	3	3	4	5	1	4	2	6	4
Education	12	29	7	23	2	8	3	13	4	13	6	16	5	18	5	16
Health and welfare	6	6	7	9	2	11	5	6	4	12	13	18	10	6	6	8
Social science, BA, and law	31	36	25	31	42	44	25	39	21	33	27	37	33	38	28	36
Natural science and ICT	17	8	17	8	15	9	22	10	13	5	15	2	14	4	17	8
Engineering	26	8	24	9	33	19	25	7	52	18	18	5	25	7	31	10
Other	8	11	9	14	4	7	8	20	3	14	17	20	9	24	8	18
N observations/missings
Earnings (obs)	614	614	630	630	412	412	1932	1932	574	574	248	248	774	774	1002	1070
Earnings (missings)	62	62	52	52	28	28	196	196	126	126	17	17	21	21	96	104
Wages (obs)	589	589	589	589	386	386	1845	1845	547	547	245	245	724	724	954	1019
Wages (missings)	87	87	93	93	54	54	283	283	153	153	20	20	71	71	144	155

Source: Eurograduate Pilot Survey. Own calculations.

Notes: ^aIn Euro, weighted for purchase power parity.

^bWeekly working hours (with no overtime).

^cColumn percentages.

Measures

Our two outcomes are monthly earnings and hourly wages 1 year after graduation. Monthly earnings are top coded using the 99th percentile for Austria, Czech Republic, Greece, Croatia, Lithuania. For Malta and Norway, we use the 95th percentile since the increase from the 95^th to the 99^th percentile is relatively large. Hourly wages are computed based on monthly earnings and average weekly working hours reported by the respondents (without overtime). We apply the same top-coding procedure to hourly wages. Both variables are measured in Euros and logged to approximate a normal distribution.

There are several approaches to the measurement of education–job mismatches (e.g. Perry et al., 2016; Van Der Velden and Bijlsma, 2019). Objective measures are based on the comparison of observed education and skill levels⁷ with the level or type of skills required for the job, often derived from occupation-specific job descriptions (e.g. Addison et al., 2020). Subjective measures are based on respondent’s self-evaluation of job requirements, the own skills, or even the actual extent of mismatch (e.g. Figueiredo et al., 2015; McGuinness and Sloane, 2011).

Eurograduate contains data on self-reported measures of both overeducation and horizontal mismatch. Overeducation occurs when respondents consider the formal qualification required in their job lower than their actual degree. For example, when graduates from a MA degree work in jobs requiring no degree or short-cycle or BA degrees. Horizontal mismatch relates to the match between the current job and the type and level of education and skills. Field-of-study mismatch leverages respondents’ self-assessment regarding whether the field of study is appropriate for the current job. Respondents are considered mismatched if a completely different or no particular field was required (as opposed to exclusively own, own/related field). On the other hand, field-specific over-skilling refers to a situation in which field-specific competencies required by the job were lower than respondents’ self-assessed level.

We further examine over-skilling in three additional (more general) competence domains: ICT, problem solving, and communication skills (results can be found in the Appendix Table B1).

Figures 2 –4 provide descriptive statistics on the extent of overeducation and horizontal mismatch by gender in each country. Overeducation ranges from 13% for males in the Czech Republic to 30% for females in Greece. In Austria, Czech Republic, and Greece women are more likely to be overeducated compared to men, while the opposite is true for Croatia and Malta. The rather untypical male disadvantage in overeducation might be due to labor-market specificities of these 2 countries; being largely depended on tourism, the overall labor-market structure might give an advantage to female-typical occupations. Except for Greece, field-of-study mismatch tends to be lower than overeducation and less differentiated by gender. In contrast, large shares of respondents consider themselves being overskilled with regards to their field-specific competencies. This is particularly true for Austria and the Czech Republic.

Figure 2.

Overeducation. Source: Eurograduate Pilot Survey. Own calculations.

Figure 3.

Horizontal mismatch: field of study needed in current job. Source: Eurograduate Pilot Survey. Own calculations.

Figure 4.

Horizontal mismatch: field-specific skills. Source: Eurograduate Pilot Survey. Own calculations.

The incidence of overskilling is higher among women than among males in all countries; and yet differences can be relatively small. Generally, the descriptive statistics suggest that respondents are more likely to consider themselves overskilled than overeducated. As expected, undereducation and underskilling is less prevalent among higher education graduates than overeducation and horizontal mismatch, and no clear gender pattern emerges. For this reason, undereducated and underskilled individuals are included in the respective reference categories (matched workers) in our analyses.

Finally, we distinguish graduates according to the ISCED level of the study course completed the year before the interview, that is bachelor’s and master’s degrees. We also distinguish seven detailed fields of study. Table 2 reports summary statistics for all analytical variables.

Empirical strategy

The analyses are organized in 3 steps. In the first step, we describe and compare the raw GEG and GWG across the 7 European countries. The raw gaps are computed based ordinary least squares (OLS) models regressing log-monthly earnings and log-hourly wages on gender (female = 1), separately by country. We present the results in terms of percentage differences in wages/earning of women compared to men. Hence, for example, a coefficient of −0.2 implies that women’s average wages (earnings) are 18.1% lower compared to men’s average wages (earnings) ([ $\exp (- . 2) - 1] \times 100) .$

In the second step, we apply Kitagawa–Blinder–Oaxaca decomposition (Jann, 2008) to assess the share of GEG and GWG explained by overeducation and horizontal mismatch. At the same time, we also assess the contribution of the field of study itself and the course type (ISCED) to women’s disadvantage. The decomposition is implemented separately by country, but only for those countries in which we found substantially and statistically significant gender penalties in this first step of the analyses. The decomposition simulates the amount of gender inequality we would observe if we equalized the mismatch probability between women and men. Such simulation bears interesting implications from a policy perspective because it allows us to assess the extent to which policies affecting gender disparities in overeducation and horizontal mismatch have any potential to reduce gender inequality in labor income at the national level (although the approach does not allow for a micro-level interpretation of the effects of mismatch on the observed outcome).

In the third step, we test for gender differences in the propensity to experience overeducation and horizontal mismatch, and for potential wage/earning penalties associated with mismatch. We test for gender differences in mismatch via logistic regressions conditioning mismatch on gender. The mismatch penalty is assessed regressing wages (earnings) on overeducation and horizontal mismatch (separately) while conditioning for the ISCED level and the field of study (field). We used sampling design weights in all three steps of the analyses.

Results

Gender inequality in wages and earnings

Table 3 shows the average GEG and GWG in the 7 European countries. Gaps are computed as percentage differences of women’s earnings or wages compared to men’s. Women’s disadvantage is apparent in most countries. Women’s monthly earnings are substantially lower than men’s in all countries. The only exceptions are Malta and Norway, where estimates are comparatively small in magnitude and not generalizable beyond our samples. The GEG ranges from 18% in Austria to 23–25% in Czech Republic, Greece, and Croatia, and to a maximum of 33% in Lithuania. A similar ranking holds when looking at the GWG, which is smallest in Austria (10%) and largest in Lithuania (28%). Interestingly, the point estimate suggests that also in Norway suggests women get lower hourly wages compared to men. This is not the case in Malta, the only country with no sign of a women’s penalty either in terms of earnings or wages.

Table 3.

Raw gender gaps in earning (GEG) and wages (GWG) by country.

	GEG	SE	GWG	SE
Austria	−17.6***	3.7	−9.8***	3.1
Czech Republic	−24.7***	3.7	−25.2***	3.2
Greece	−25.3***	5.6	−17.2***	5.2
Croatia	−23.2***	2.3	−22.5***	2.3
Lithuania	−32.9***	3.3	−28.3***	4.3
Malta	−4.6	6.8	2.6	6.6
Norway	−9.1	5.1	−10.0**	3.2

Source: Eurograduate pilot survey. Own calculations. Notes: Gender gaps are computed as percentage differences of women’s earning/wages compared to men. *p < .05; **p < .01; ***p < .001.

Gaps in hourly wages are lower than gaps in monthly earnings in most countries. Figure 5 visualizes such differences by contrasting earning gaps (x-axis) and wage gaps (y-axis) in each country. The red line depicts a situation in which GWG and GEG are equal. Countries below the line have stronger GEG compared to GWG. The GWG is much lower than the GEG in Austria, Greece and Lithuania. This difference suggests that women’s penalty in monthly earnings are partly explained by working hours. Indeed, gender differences in working time are most pronounced in these countries (see Table 2). And yet in Croatia and Czech Republic—and perhaps even in Norway, where earning gaps are not statistically significant—working hours seem not to offer any explanation as for why women earn less than men monthly.

Figure 5.

Gender gaps in earnings versus gender gaps in wages. ed line: GEG = GWG. Source: Eurograduate Pilot Survey. Own calculations.

GWG and GEG shown in Table 3 and Figure 5 are surprisingly high considering that we are looking at tertiary graduates, a population that is believed to have meritocratic opportunities on the labor market. We are not in the position to ascertain whether women’s disadvantage is due to discrimination or other supply and demand factors. Nonetheless, these penalties are worrisome because they refer to highly educated women who spent a remarkable amount of time and effort in education.

Do overeducation and horizontal mismatch explain women’s penalties?

Figure 6 reports the main findings from the analyses decomposing the GEG (panel A) and the GWG (panel B) in the countries where these gaps were both substantially and statistically significant (hence, Malta and Norway [only GEG] are excluded).

Figure 6.

Percentage of the raw GEG (Panel A) and GWG (Panel B) explained by course type (ISCED), field of study, overeducation, and horizontal mismatch (field-specific mismatch). 95% confidence intervals shown. Source: Eurograduate Pilot Survey. Own calculations.

We report the contribution of overeducation and field-of-study mismatch together with the ISCED level and the field of study to women’s penalties. Positive values indicate that the relative characteristic contribute to explaining women’s penalty; negative values indicate that the characteristic works at women’s advantage.

We find striking results. Overeducation and horizontal mismatch do not explain neither GEG nor GWG between higher education graduates in any of the seven countries. Point estimates suggest that overeducation explains a minor share (3–6%) of the GEG and GEG only in Greece and Austria (also in Czech Republic when looking at the GWG). However, these point estimates are limited in magnitude and within the range of estimation error. Point estimates for field-of-study mismatch are close to 0 in all countries (except for the GWG in Greece, around 6%). The results using field-specific overskilling are almost identical to those presented in Figure 6 (see also Appendix Table B1, which also reports results using overskilling in the other domains – ICT, communication, and problem solving).

We conducted three further robustness checks on these findings. First, we excluded the undereducated from our sample; second, we used an alternative measure of severe overeducation;⁸ third, we ran the analyses pooling countries together and adjusting for country dummies in the decomposition. Results are unchanged: overeducation and horizontal mismatch do not explain why women get lower hourly wages and monthly earnings compared to men (see Appendix Figures C1, C2, and C3).

The field of study remains the most important factor behind women’s disadvantage among higher education graduates, in line with previous research. Although there is some uncertainty around the point estimates, possibly due to the small samples at hand, the field of study explains from 3% (in Czech Republic) to 23% (in Greece) of the GEG depending on the country (see panel A). These figures are generally lower and more uncertain in the case of the GWG (see panel B).

Explanations behind the absence of mediation

Our conceptual model makes it apparent that the contribution of overeducation and horizontal mismatch to the explanation of gender gaps in labor income hinges on 2 conditions. First, the existence of a difference in the propensity of experiencing overeducation and horizontal mismatch between men and women (B). Second, the existence of earnings/wage penalties related to the mismatch (C). Looking at these underlying relationships may provide some explanations for the negligible contributions to women’s penalty of both overeducation and horizontal mismatch revealed by our decomposition analysis.

Table 4 shows the odds ratios between the relative propensities be overeducated and mismatched with regards to field of study and skill level by gender (Columns B). Columns C show the average wage and earning penalty of a mismatch Appendix Tables B2 and B3 present results adding overskilling in all four competence domains simultaneously.

Table 4.

Associations between gender and mismatch (Path B in Figure 1) and between mismatch and earnings/wages (Path C in Figure 1).

Mismatch	Overeducation			Field-of-study mismatch			Overskilling
Country	B	C (earnings)	C (wages)	B	C (earnings)	C (wages)	B	C (earnings)	C (wages)
AT	1.15	−0.13*	−0.08*	1.05	−0.09	−0.09+	1.05	0.02	0.01
CZ	1.83**	−0.02	−0.08	1.00	0.09	0.11**	1.20	−0.07*	−0.06*
GR	1.31	−0.21**	−0.10	2.55*	−0.01	−0.04	1.29	0.22+	0.18*
HR	0.88	−0.14+	−0.14+	0.83	−0.09+	−0.03	1.14	−0.07**	−0.06+
LT	0.75	−0.18**	−0.23**	1.26	0.05	0.02	0.96	−0.05	−0.04
MT	1.08	−0.05	−0.01	1.08	−0.08	−0.07	1.76**	0.01	0.00
NO	0.97	−0.28***	−0.13**	0.89	−0.09	−0.05	1.45*	−0.06	−0.01

Source: Eurograduate Pilot Survey. Own calculations. B is the odds ratio from logistic regression models conditioning overeducation and horizontal mismatch on gender. C is the coefficient from linear regression models conditioning log-earnings (wages) on overeducation and horizontal mismatch (controlling for gender, level of education and field of study). When estimating the earnings penalty of overeducation we additionally condition on field-of-study mismatch. Results are qualitatively similar when using field-specific overskilling. ⁺p < .1; *p < .05; **p < .01; ***p < .001.

Overeducation is associated with a significant earnings penalty in Austria, Greece, Croatia, Lithuania, and Norway (coefficients ranging from −0.13 to −0.28). However, we do not find significant gender differences in the propensity to be overeducated in most countries (except for Czech Republic). Focusing on field-of-study mismatch, we found a negative association with income in Austria (wages) and Croatia (earnings) only. However, in any of those countries women have a higher propensity to be mismatched. Only female graduates in Greece are significantly more likely to report that their field of study is little or not required for conducting their current job, but this seems not accompanied by income penalties.

Table 4 also shows significant gender differences in field-of-study mismatch in Malta and Norway, but again this does lead to a significant earnings or wage penalty. Conversely, field-specific over-skilling is associated with lower income in the Czech Republic and Croatia, yet, here gender and over-skilling seems not associated. Appendix Tables B2 and B3 additionally show that, depending on the country, gender gaps in more specific skill domains exist in ICT (Greece), problem solving (Germany, Croatia, Malta, and Norway) and communication skills (Czech Republic, Germany, Croatia, and Norway); still, for those countries and domains we could not detect any negative association between mismatch and income.

Our results suggest that in neither of the countries the two necessary conditions for mediation are in place simultaneously – that is a gender-specific difference in mismatch (B) and an income penalty associated with mismatch (C). Either of the two conditions are in place depending on the country but never at the same time. The non-coexistence of gender gaps in mismatch and mismatch penalties help explaining the absence of mediation we found in the decomposition analyses.

Summary and conclusion

Gender gaps in labor income are puzzling in times of rapid higher education expansion. Women’s have nowadays surpassed men in terms of educational attainment in many countries. And yet women have lower labor income than equally, high-educated men. Against this backdrop, the aim of our paper was twofold. First, we aimed to provide evidence for gender gaps in labor income—monthly earnings (GEG) and hourly wages (GWG)—among higher education graduates in a variety of European countries. By using data from the Eurograduate pilot survey, we expanded the existing evidence to recent cohorts of graduates and to European countries that were less studied by the previous literature. Our results join a long-standing tradition of research showing significant gender gaps in monthly wages and hourly earnings. Gender gaps were smallest in Norway, in line with OECD figures for the general population, and Malta. Where existing, gender gaps in monthly earnings varied from a minimum of 18% in Austria and a maximum of 33% in Lithuania. The same ranking held when looking at wage gaps, which were smallest in Austria (10%) and strongest in Lithuania (28%). The general drop in gender differentials in hourly wages compared to monthly earnings confirm the importance of working hours in explaining gender disparities. In addition to working hours, the field of study is generally the most important factor explaining these gender differences in labor income.

Our second aim was to examine the contribution of overeducation and horizontal mismatch to gender disparities in labor income. We conducted separate mediation analyses for each country and used Kitagawa-Blinder-Oaxaca decomposition methods to assess the relative contribution of overeducation and field-of-study mismatch to explain gender gaps in income. We found that neither overeducation nor horizontal mismatch contribute to the explanation of women’s disadvantage. Also, in some countries women are more likely have jobs that do not reflect their education or skill level. However, in those countries overeducation and horizontal mismatch do not seem to carry an income penalty. In other countries such income penalty exists, but men and women do not differ in the probability of being overeducated or mismatched. These findings point toward the idea that, overall, education–job mismatch is not a crucial driver of gender inequality in labor income among higher education graduates.

That said, our findings have several limitations which should not go unmentioned. First, the lack of mediation of education–job mismatch maybe driven by measurement error. This is a crucial concern because Eurograduate data only provides self-reported assessment measures. Previous evidence on the validity of self-reported data varies largely with the skill domain and is rather mixed. However, Allen (in this special issue) demonstrates that measures of self-reported, field-specific skills of higher education graduates are quite valid. Moreover, if our overeducation and horizontal mismatch variables were completely off the target, it would be unlikely to observe gender difference in the probability of being mismatched.

We would also stress that our conclusions about the lack of mediation hinge on the point estimates (the proportion of gender gaps explained by mismatch is very close to 0 in most countries) and not to the absence of statistical significance of the estimates. Hence, the relatively low size of the country-samples seems not a good candidate for explaining the null finding (see also the decomposition results from the models pooling together the countries in Appendix Figure C3).

Another limitation stems from our focus on limited number of countries and from the adoption of a micro-level approach. That is, we did not attempt to model explicitly the association between the extent of gender gaps and other contextual or institutional characteristics as done in a macro-micro approach (Van der Lippe and Van Dijk, 2002). Provided the availability of data on a larger number of countries and periods, future research may explore the importance of welfare state contexts, labor market settings, and other institutional specificities for variations in gender disparities in labor income. The Eurograduate survey may be a good candidate for such avenue as a new round of data collection on 17 European countries has started in 2022 (Eurograduate Consortium, 2022).

A third concern regards the relatively small size of the individual country samples. This limitation prevented us to run more nuanced analyses, for example by testing for potential interactions between the field of study and mismatch in affecting wages and earnings. Further investigations could, for example, analyze whether overeducation and horizontal mismatch are equally irrelevant for graduates from female-dominated, male-dominated, or more balanced fields of study (Bol and Heisig, 2021; Leuze and Strauß, 2009, 2014).

Despite the limitations, our findings bear interesting implications for both policy and future research. We found that either education–job mismatches are not associated with wage and earning penalties, or that the propensity to experience education–job mismatches do not vary by gender. The first finding needs further examination. It is not fully clear why mismatch is not related to income disparities. This is the case in Norway and Malta and might indeed either hint toward measurement error or toward specific labor market features in these countries. The second finding is relevant for policy makers because it does indeed seem as if highly qualified men and women are less affected by labor market inequality in a lot of countries since they seem equally likely to be mismatched on the labor market.

Supplemental Material

Supplemental Material - The role of overeducation and horizontal mismatch for gender inequalities in labor income of higher education graduates in Europe

Supplemental Material for The role of overeducation and horizontal mismatch for gender inequalities in labor income of higher education graduates in Europe by Giampiero Passaretta, Petra Sauer, Ulrike Schwabe, and Katarina Weßling in Research in Comparative and International Education.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the CIDER (College for Interdisciplinary Educational Research) Micro Group “Gender inequality in graduates’ employment prospects”.

ORCID iDs

Giampiero Passaretta

Ulrike Schwabe

Katarina Weßling

Supplemental Material

Supplemental material for this article is available online.

Notes

Author biographies

Giampiero Passaretta is Assistant Professor at the Department of Political and Social Sciences at Universitat Pompeu Fabra, Spain. Previously, he held research positions at Trinity College Dublin, European University Institute, and Stockholm University. His research interests include education and labour market inequalities, comparative sociology, and quantitative methods of data analysis.

Petra Sauer is (LIS)^²ER Tony Atkinson research fellow at LIS and LISER in Luxembourg. She is also affiliated with the Research Institute `Economics of Inequality’ (INEQ, Vienna University of Economics and Business). She obtained her PhD in Economics from the Vienna University of Economics and Business in 2015. In her research, Petra is concerned with the causes and consequences of social and economic inequalities which stem from unequal educational opportunities and labour market experiences across socio-economic groups.

Ulrike Schwabe is working as a researcher in the research area Educational Careers and Graduate Employment at the German Centre for Higher Education Reserach and Science Studies (DZHW). Her research interests are sociology of education, sociology of labour markets, research in social stratification, empirical higher education research and international comparisons.

Katarina Weßling is a junior research group leader at the Federal Institute for Vocational Education and Training (BIBB), Germany and at ROA, Maastricht University. Previously she held positions at the University of Cologne and the University of Tuebingen, where she obtained her PhD in Sociology. Her research is concerned with social – in particular educational – inequalities over the life courses of individuals and in the relevance of (spatial, national and institutional) contexts. A focus of her empirical work is on linking large-scale survey data with geocoded information on contextual characteristics.

References

Achatz

Gartner

Glück

(2005) Bonus oder Bias? Mechanismen geschlechtsspezifischer Entlohnung. Kölner Zeitschrift für Soziologie und Sozialpsychologie 57: 466–493. DOI: 10.1007/s11577-005-0185-6

Addison

Chen

Ozturk

(2020) Occupational skill mismatch: differences by gender and cohort. ILR Review 73(3): 730–767. DOI: 10.1177/0019793919873864

Allen

Levels

Van Der Velden

(2013) Skill mismatch and skill use in developed countries: Evidence from the PIAAC study (No. 017. Research Memoranda). DOI: 10.26481/umaror.2013017

Allen

van der Velden

(2001) Educational mismatches versus skill mismatches: Effects on wages, job satisfaction, and on-the-job search. Oxford Economic Papers 53(3): 434–452. DOI: 10.1093/oep/53.3.434

Amilon

Persson

(2013) Scientific (wo)manpower – gender and the composition and earnings of PhDs in Sweden. International Journal of Manpower 34(6): 658–673.

Bar-Haim

Chauvel

Gornick

, et al. (2018) The Persistence of the Gender Earnings Gap: Cohort Trends and the Role of Education in Twelve Countries. (LIS Working Paper Series No. 737). Luxembourg Income Study (LIS). DOI: 10.1111/j.1467-629x.1980.tb00220.x.

Baron

Kenny

(1986) The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology 51(6): 1173–1182. DOI: 10.1037/0022-3514.51.6.1173

Blau

Kahn

(2017) The gender wage gap: extent, trends, and explanations. Journal of Economic Literature 55(3): 789–865. DOI: 10.2139/ssrn.2743114

Bol

Ciocca Eller

van de Werfhorst

, et al. (2019) School-to-work linkages, educational mismatches, and labor market outcomes. American Sociological Review 84(2): 275–307. DOI: 10.1177/0003122419836081

10.

Bol

Heisig

(2021) Explaining wage differentials by field of study among higher education graduates: Evidence from a large-scale survey of adult skills. Social Science Research 99: 102594. DOI: 10.1016/j.ssresearch.2021.102594

11.

Boll

Leppin

(2014) Formale Überqualifikation unter ost-und westdeutschen Beschäftigten. Wirtschaftsdienst 94: 50–57. DOI: 10.1007/s10273-014-1624-7

12.

Brynin

Longhi

(2009) Overqualification: major or minor mismatch? Economics of Education Review 28(1): 114–121.

13.

Budig

Misra

Böckmann

(2010) The Motherhood Penalty in Cross-National Perspective: The Importance of Work-Family Policies and Cultural Attitudes. (LIS Working Paper Series No. 542). Luxembourg Income Study (LIS).

14.

Chevalier

(2007) Education, occupation and career expectations: determinants of the gender pay gap for UK graduates. Oxford Bulletin of Economics and Statistics 69(6): 819–842. DOI: 10.1111/j.1468-0084.2007.00483.x

15.

Christl

Köppl–Turyna

(2020) Gender wage gap and the role of skills and tasks: evidence from the Austrian PIAAC data set. Applied Economics 52(2): 113–134. DOI: 10.1080/00036846.2019.1630707

16.

Eurograduate Consortium (2022). www.eurograduate.eu (retrieved 01-07-2022).

17.

Figueiredo

Rocha

Biscaia

, et al. (2015) Gender pay gaps and the restructuring of graduate labour markets in Southern Europe. Cambridge Journal of Economics 39: 565–598.

18.

Francesconi

Parey

(2018) Early gender gaps among university graduates. European Economic Review 109: 63–82. DOI: 10.1016/j.euroecorev.2018.02.004

19.

García-Aracil

(2007) Gender earnings gap among young European higher education graduates. Higher Education 53: 431–455. DOI: 10.1007/s10734-005-3864-3

20.

Goldan

(2021) Explaining the gender pay gap among doctoral graduates: analyses of the German labour market. European Journal of Higher Education 11(2): 137–159. DOI: 10.1080/21568235.2020.1835516

21.

Grusky

(ed) (2014) Social Stratification: Class, Race, and Gender in Sociological Perspective. 4th edition. Westview Press.

22.

Hanushek

Schwerdt

Wiederhold

, et al. (2015) Returns to skills around the world: Evidence from PIAAC. European Economic Review 73: 103–130. DOI: 10.1016/j.euroecorev.2014.10.006

23.

Jann

(2008) The Blinder–Oaxaca decomposition for linear regression models. The Stata Journal: Promoting Communications on Statistics and Stata 8(4): 453–479.

24.

Kleibrink

(2014) Sick of your job? – Negative health effects from non-optimal employment (No. 514. Ruhr Economic Papers). DOI: 10.4419/86788589

25.

Kracke

Reichelt

Vicari

(2018) Wage losses due to overqualification: the role of formal degrees and occupational skills. Social Indicators Research 139(3): 1085–1108.

26.

Leuze

Strauß

(2009) Lohnungleichheiten zwischen Akademikerinnen und Akademikern: Der Einfluss von fachlicher Spezialisierung, frauendominierten Fächern und beruflicher Segregation [Wage inequality between male and female university graduates: The influence of occupational segregation]. Zeitschrift für Soziologie 38(4): 262–281. DOI: 10.1515/zfsoz-2009-0401

27.

Leuze

Strauß

(2014) Female-typical subjects and their effect on wage inequalities among higher education graduates in Germany. European Societies 16(2): 275–298. DOI: 10.1080/14616696.2012.748929

28.

Levels

Van der Velden

Di Stasio

(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): 341–361.

29.

Machin

Puhani

P A

(2003) Subject of degree and the gender wage differential: evidence from the UK and Germany. Economics Letters 79: 393–400.

30.

Marginson

(2016) The worldwide trend to high participation higher education: dynamics of social stratification in inclusive systems. Higher Education 72(4): 413–434. DOI: 10.1007/s10734-016-0016-x

31.

Maurer

Unger

Jühlke

, et al (2021). EUROGRADUATE Pilot Survey 2018 . Data Collection: 2018/2019. Version: 2.0.0. Data Package Access Way: SUF: Download. FDZ-DZHW. Data Curation: Hoffstätter U Wallis

Niebuhr

. DOI: 10.21249/DZHW:egr2018:2.0.0

32.

Mavromaras

Mcguinness

O’Leary

, et al. (2013) Job mismatches and labour market outcomes: panel evidence on university graduates. Economic Record 89(286): 382–395. DOI: 10.1111/1475-4932.12054

33.

McGuinness

(2006) Overeducation in the labour market. Journal of Economic Surveys 20(3): 387–418.

34.

McGuinness

Sloane

(2011) Labour market mismatch among UK graduates: an analysis using REFLEX data. Economics of Education Review 30: 130–145.

35.

Meng

Wessling

Mühleck

, et al. (2020) Eurograduate Pilot Survey - Design and Implementation of a Pilot European Graduate Survey. European Commission - Directorate-General for Education, Youth, Sport and Culture https://op.europa.eu/en/publication-detail/-/publication/51f88c2e-a671-11ea-bb7a-01aa75ed71a1/language-en

36.

Mühleck

Meng

Oelker

, et al. (2020) Eurograduate Pilot Survey: Technical Assessment and Feasibility of a Full Rollout - Executive Summary. European Commission - Directorate-General for Education, Youth, Sport and Culture. https://www.eurograduate.eu/about/documents

37.

OECD (2018) Education at a Glance 2018: OECD Indicators. DOI: 10.1787/eag-2018-en

38.

Passaretta

Triventi

(2021) Inequality at the Top: The Gender Wage Gap Among the Italian Educational Elites. SocArXiv Papers. DOI: 10.31235/osf.io/updgw

39.

Perry

Martinez

Morris

, et al. (2016) Misalignment of career and educational aspirations in middle school: Differences across race, ethnicity, and socioeconomic status. Social Sciences 5(3): 1–10. DOI: 10.3390/socsci5030035

40.

Pető

Reizer

(2021) Gender differences in the skill content of jobs. Journal of Population Economics 34(3): 825–864. DOI: 10.1007/s00148-021-00825-6

41.

Piazzalunga

(2018) The gender wage gap among college graduates in Italy. Italian Economic Journal 4(1): 33–90. DOI: 10.1007/s40797-017-0069-8

42.

Reimer

Schröder

(2006) Tracing the gender wage gap: income differences between male and female university graduates in Germany. Zeitschrift für Arbeitsmarktforschung 39(2): 235–253.

43.

Research Centre for Education and the Labour Market (ROA) - Maastricht University (coordinator) (2005) REFLEX - REsearch into Employment and Professional FLExibility. DOI: 10.17026/dans-z3s-a2dh

44.

Rossen

Boll

Wolf

(2019) Patterns of overeducation in Europe: the role of field of study. IZA Journal of Labor Policy 9(1): 1–48. DOI: 10.2478/izajolp-2019-0003

45.

Salas-Velasco

(2021) Mapping the (mis)match of university degrees in the graduate labor market. Journal for Labour Market Research 55(14): 1–23. DOI: 10.1186/s12651-021-00297-x

46.

Sattinger

Hartog

(2013) Nash bargaining and the wage consequences of educational mismatches. Labour Economics 23: 50–56.

47.

Sauer

Van Kerm

(2021). Higher Education Expansion and Inequality in Labour Incomes: The Importance of a Gendered Perspective (Issue No. 18). https://www.lisdatacenter.org/newsletter/nl-2021-18-im-1/#

48.

Sauer

van Kerm

Checchi

(2022). Higher education expansion and labour income inequality in high-income countries: a Gender-specific perspective (No. 837; LIS Working Paper Series). https://www.lisdatacenter.org/wps/liswps/837.pdf

49.

Schomburg

Teichler

(2006) Higher Education and Graduate Employment in Europe: Results of Graduates Surveys from 12 Countries. Dordrecht: Springer.

50.

Shrout

P E

Bolger

(2002) Mediation in experimental and nonexperimental studies: new procedures and recommendations. Psychological Methods 7(4): 422–445. DOI: 10.1037/1082-989X.7.4.422

51.

Triventi

(2013) The gender wage gap and its institutional context: a comparative analysis of European graduates. Work, Employment and Society 27(4): 563–580. DOI: 10.1177/0950017012460322

52.

Van der Lippe

van Dijk

(2002) Comparative research on women’s employment. Annual Review of Sociology 28: 221–241.

53.

Van Der Velden

Bijlsma

(2019) Effective skill: a new theoretical perspective on the relation between skills, skill use, mismatches, and wages. Oxford Economic Papers 71(1): 145–165. DOI: 10.1093/oep/gpy028

54.

Weichselbaumer

Winter-Ebmer

(2005) A meta-analysis of the international gender wage gap. Journal of Economic Surveys 19(3): 479–511. DOI: 10.1111/j.0950-0804.2005.00256.x

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

1.17 MB