DISCUSSION PAPER SERIES

IZA DP No. 12407

The Impact of Horizontal Job-Education

Mismatches on the Earnings of Recent
University Graduates in Russia

Victor Rudakov
Hugo Figueiredo
Pedro Teixeira
Sergey Roshchin

JUNE 2019
DISCUSSION PAPER SERIES

IZA DP No. 12407

The Impact of Horizontal Job-Education

Mismatches on the Earnings of Recent
University Graduates in Russia

Victor Rudakov Pedro Teixeira

HSE and University of Porto IZA and University of Porto
Hugo Figueiredo Sergey Roshchin
University of Aveiro HSE

JUNE 2019

Any opinions expressed in this paper are those of the author(s) and not those of IZA. Research published in this series may
include views on policy, but IZA takes no institutional policy positions. The IZA research network is committed to the IZA
Guiding Principles of Research Integrity.
The IZA Institute of Labor Economics is an independent economic research institute that conducts research in labor economics
and offers evidence-based policy advice on labor market issues. Supported by the Deutsche Post Foundation, IZA runs the
world’s largest network of economists, whose research aims to provide answers to the global labor market challenges of our
time. Our key objective is to build bridges between academic research, policymakers and society.
IZA Discussion Papers often represent preliminary work and are circulated to encourage discussion. Citation of such a paper
should account for its provisional character. A revised version may be available directly from the author.

ISSN: 2365-9793

IZA – Institute of Labor Economics

Schaumburg-Lippe-Straße 5–9 Phone: +49-228-3894-0
53113 Bonn, Germany Email: publications@iza.org www.iza.org
IZA DP No. 12407 JUNE 2019

ABSTRACT
The Impact of Horizontal Job-Education
Mismatches on the Earnings of Recent
University Graduates in Russia*
This paper analyses the determinants and consequences of horizontal job-education
mismatches, an increasingly relevant topic in debates about education and labour markets.
This issue reflects the articulation of educational fields and occupations in the labour
market. We evaluate the determinants of job-education mismatches and their impact
on salaries of university graduates using comprehensive and representative national data
for Russia. The study is based in graduates’ assessment and statistical analyses. We find
that one-third of graduates in Russia work in a job that is not related to their field of
study. Moreover, graduates from fields that either generate more general human capital
(social sciences, business, law, services) or where low pay is common (agriculture) are
more likely to be in that situation. On the contrary, graduates from fields that generate
specific human capital (e.g.: medicine) are considerably more likely to be matched. We
find that mismatches negatively affect the earnings of university graduates and the higher
the degree of mismatch, the higher the penalty for the mismatch. The study depicts that
mismatch is penalized in the majority of fields except for low-paid ones (e.g.: agriculture).

JEL Classification: J24, J30, J31

Keywords: job-education mismatch, education-occupation mismatch,
horizontal mismatch, graduate salaries, labor market outcomes,
human capital

Corresponding author:
Pedro Teixeira
University of Porto
Praça de Gomes Teixeira
4099-002 Porto
Portugal
E-mail: pedrotx@fep.up.pt

* This work was supported by the Basic Research Program at the National Research University Higher School of
Economics in 2017 and Portuguese Foundation for Science and Technology (FCT), within the scope of the project
PTDC/CED-EDG/29726/2017.
 Introduction

The issue of horizontal job-education mismatch attracts considerable attention among

policy- makers and researchers as it reflects the congruency between higher education fields and
labor market requirements. It emphasizes the relations between graduate supply and employers
demand for human capital, specific fields and occupations, and shows how the labor market
value skills produced during university studies. On the one hand, the dynamics of labor markets
lead to rapid changes in the nature and structure of graduate jobs. Labor markets are frequently
more dynamic than educational systems, which cannot immediately react to the change in
technologies or demand for certain professions. This situation is defined as a race between
education and technology (Goldin and Katz, 2009). On the other hand, in some countries, labor
markets appear to not offer a sufficient amount of jobs requiring university education. The rapid
expansion of higher education during the last two decades was not accompanied by substantial
growth in high-skilled jobs, especially in Southern European and Post-Soviet countries
(Marginson, 2016; Green, Henseke, 2017). Mass higher education can therefore create an
oversupply of university graduates and result in labor market imbalances. These imbalances
create a possible decrease of the graduate wage premium (GWP) as well as increasing job-
education mismatches and the problem of overeducation.
Job mismatches among university graduates can be vertical and horizontal. Vertical
mismatch (overeducation) appears when university graduates work in non-graduate jobs.
Vertical mismatch is widely studied and most researchers found a significant wage penalty for
overeducated graduates in general (Wolbers, 2003; Chevalier, Lindley, 2009; Sloane, 2003;
McGuiness, 2006; Robst, 2007; Mavromaras, Mc.Guiness, 2012; Meroni, Vera-Toscano, 2017).
Our paper is mainly focused on horizontal mismatch, which is defined in different papers as job-
education mismatch (Robst, 2006) or education-occupation mismatch (Nordin, 2011). This
problem appears when graduates are employed in a job which is not related to their field of study.
For instance, graduates with a degree in Law or Engineering but who work in the service sector,
tourism, hospitality, retail can be considered as horizontally mismatched. Alternatively, they can
lack the appropriate skills needed for these jobs and then it is more connected with issues of
over- and under- skilling (Chevalier, Lindley, 2009). The problem of horizontal (job-education or
occupation-education mismatch) is less studied due to the lack of relevant data on self-evaluated
mismatched or incapability of building of statistical measures.
The dominant point of view on the effects of job-education mismatches on graduate
salaries is that job-education mismatches lead to the waste of the human capital accumulated
2
during graduates’ study years and brings negative consequences– wage penalties and lower job
satisfaction. This is confirmed by existing studies devoted to this problem (Robst, 2007; Nordin,
et.al, 2010; Boudarbat, Chernoff, 2012). Nevertheless, job-education mismatch can be positive
for some graduates as they can change their occupation after finishing their formal education
and find a better match between their abilities and the characteristics of the job 1.
Job-education mismatch can be also viewed as a tool for individual labor market
adjustment in the case of imbalances between demand and supply and the lack of vacancies for
particular jobs. From this point of view, we may expect different effects of job-education
mismatch for different fields of study. On the one hand, some fields of study (e.g.: Medicine,
Computer Science, STEM, Law) imply the accumulation of occupation-specific hard skills, most of
which are non-transferable to other sectors (Robst, 2007; Boudarbat, Chernoff, 2012). Graduates
of these fields are less likely to be horizontally mismatched as it is very non-beneficial to choose
another occupation due to a large waste of accumulated human capital. As a result, we may
expect a higher penalty for job-education mismatch for graduates with a degree in fields which
develop occupation-specific skills. On the other hand, there are some specializations which
mainly develop general skills (Social Sciences, Arts and Humanities, Services). These general skills
are easily transferable across sectors (Robst, 2007; Boudarbat, Chernoff, 2012). As a result,
graduates in these fields are relatively more likely to be mismatched, but the penalty for
mismatch could be very low, or insignificant, or we may expect even positive effects of mismatch.
The main purpose of the study is to evaluate the extent and determinants of job –
education mismatch and estimate their impact on the salaries of recent university graduates in
Russia. The study aims to answer the following questions: What is the proportion of mismatched
graduates? Which groups of graduates are more likely to be mismatched in the labor market,
regarding, in particular, their field of studies? How do such mismatches affect the salaries of
university graduates? How do the effects differ by fields of study?
Our study contributes to the literature on horizontal job-education mismatches, their
determinants and effects and increases attention to horizontal mismatches, which can be even
more influential than vertical. Researchers found that the wage penalty for horizontal mismatch
exceeds the penalty for overeducation (Robst, 2007; Budria and Moro-Egido, 2008; Domadenik
et al.,2013). Russia is an interesting case-study for the analysis of the horizontal job-education
mismatches due to the pace of higher education expansion and the size of its labor market, its

1
Such cases of non-optimal degree choice are also better captured by self-reported measures of mismatch. We get
back to this issue in the methods section of the paper.
3
high regional differentiation and, not the least, the post-communist transition that it
experienced, which resulted in significant shifts both in the labor market and the educational
system. The educational system created during the Soviet Era, on the one hand, had to adjust to
new labor market requirements in the market economy. On the other hand, it experienced a very
rapid expansion. The issue of the matching between education and job became and still is an
important educational policy issue for the country. The paper develops the methodology of
Robst (2006) and Nordin, et al (2010) to analyze the issue of horizontal mismatch in Russia using
a large comprehensive national representative graduate dataset.
The study is structured as follows. Section 1 focuses on the main theories explaining the
determinants and possible impact of horizontal job-education mismatches. Section 2 sets up the
institutional peculiarities of the Russian higher education system and labor market. Section 3
describes the data and methodology of the study. Section 4 is devoted to a preliminary
descriptive analysis of the data. Section 5 presents the results of the regression analysis of the
determinants of job-education mismatches and their impact on the salaries of university
graduates. The final section presents some main conclusions.

1. Horizontal job-education mismatches: theory and results of empirical studies

The problem of job-education mismatches is studied in multiple papers, but the majority
of them are devoted to vertical mismatches or the problem of overeducation rather than
horizontal mismatch. The literature on the problem of overeducation is thoroughly considered
in Hartog (2000) and Mc Guiness (2006) and there is a rich empirical literature on this topic
(Wolbers, 2003; Sloane, 2003; Lamo, Messina, 2010; McGuiness, Sloane, 2011; Mavromaras,
Mc.Guiness, 2012; Baert. et al, 2013; Meroni, Vera-Toscano, 2017). A number of papers are also
devoted to changes in the structure of graduate jobs, job polarization and its consequences for
graduate labor markets and graduates working in non-graduate jobs (Elias, Purcell, 2004; 2013;
Figueiredo et.al, 2017; Green, Henseke, 2017). The problem of overeducation and returns to
higher education takes a lot of attention due to the massification of higher education and
increasing government and household investments in higher education. It raises the question if
these investments are efficient and contribute to the development of a knowledge economy or,
due to a rigid employment structure, graduates work in non-graduate jobs and are overeducated.
The issue of horizontal mismatches is considerably less studied than vertical mismatch or
overeducation mainly due to the lack of relevant data on fields of studies of university graduates
that could contribute to the creation of statistical measures of the mismatch as well as due to a
4
lack of self-evaluation data on job-education matching (Gimpelson, et al. 2010). However, the
horizontal mismatch can be even more influential than vertical mismatch in terms of the effect
on salaries of university graduates and job satisfaction and may equally lead to a considerable
waste of accumulated human capital. The existence, determinants and the impact of job-
education mismatches on graduate salaries can be explained by different labor market theories,
including the human capital theory (Becker, 1964; Mincer, 1974), job matching theory
(Jovanovich, 1979) and job assignment theory (Sattinger, 1993).
Human capital theory states that workers are distributed between the jobs according to
the amount of accumulated human capital, which includes knowledge and skills, gained through
formal education and on-the job training. From this viewpoint, differences in salaries are
connected with differences in human capital, which depend on the years of schooling, quality of
education, innate abilities and other factors. According to this view, job-education mismatch can
be only temporary and seems inefficient both for workers and for employers. It raises the
question: why does the job-education mismatch exist? Higher education contributes to the
accumulation of general human capital, which is valuable regardless of industry of employment
and specific human capital, which can be fully used and bring returns in terms of salaries, if the
graduate is perfectly matched to the job. There can be significant differences in the relative share
of general and specific human capital between fields. Robst (2006) in one of the first papers,
which was devoted to horizontal mismatch, formulated the hypotheses that educational
mismatch is more likely among workers with degree fields that provide general skills (Humanities,
Social Sciences and Liberal Arts) and less likely among graduates of majors, which provide specific
human capital (e.g. Medicine, Computer Science, Engineering, Technology).
The issue of job-education mismatches is then closely related with the issue of the choice
of college major. We may assume that an individual chooses a particular field of education and
invests financial resources and time in acquiring skills for a given field of study with the
expectation of working in an occupation related to that field. From this point of view, job-
education mismatch is likely to be inefficient for workers and result in wage penalties. The
direction of the effect depends, however, on the nature of mismatch: if it is supply-related
(initiated by the individual) or demand-related (due to lack of vacancies, high job competition
etc.) (Robst, 2007; Nordin,2010).
A demand-related mismatch is likely to negatively affect salaries, due to the fact, that in
most cases it is connected with either a lack of demand for a given field in the economy or adverse
selection in mismatch status as graduates who are mismatched involuntary have lost competition
5
with their peers and are likely to have lower abilities then those who are matched. This may lead
to the endogeneity in the empirical analysis of impact of job-education mismatch on salaries due
to the fact that graduates with lower abilities are more likely to be mismatched and have lower
salaries. On the other hand, if the negative impact of job-education mismatch is just a reflection
of lower abilities, than the effect of job-education mismatch should be the same across different
fields (Robst, 2007). Moreover, if mismatch is supply-related – it is likely to be not connected
with sorting by abilities.
The effect of supply-related mismatch is likely to be ambiguous or even positive as
mismatch is initiated by the individual and may be connected with changes in preferences of the
individual after enrollment, acquisition of more information about job characteristics in the given
field and others (Robst, 2007; Nordin, 2010). Job matching theory proposes that due to
asymmetry of information, young workers and recent university graduates are unaware of the
matching between their skills, abilities and job characteristics and requirements (Jovanovich,
1979). As a result, they try to find the best matching between their skills and jobs, which result
in higher job mobility of university graduates and lead to voluntary job-education mismatch. This
outcome can be also predicted by job assignment theory (see Sattinger, 1993) according to
which, returns to education depend partially on the quality of assignment of heterogeneous
workers to heterogeneous jobs.
There are also considerable differences in the narrowness of educational fields. For
instance, in the US, education fields are very broad (Robst, 2007), while in Sweden fields are
detailed (Nordin, 2010). In Russia, educational fields are extremely narrow. There are different
ways to measure job-education mismatch, but most of them can be divided in two types: self-
evaluation measures (see Robst, 2007) or synthetic (statistical) mismatch variables (Nordin,
2010). In our paper we use both of them. Self-evaluated measures have considerable drawbacks,
which are connected with its subjective matter and endogeneity, which is related to the wages,
as it is stated by Nordin (2010) – ‘self-reported mismatch can be a rationalization of
disappointment with wages’. On the other hand, all statistical mismatch variables are less likely
to capture mismatch and are dependent on the method used as there are no common
classifications matching education and occupation (job) and the nature of occupations is also
rapidly changing.

6
 2. Institutional framework: the case of Russia

Russia traditionally has a very high level of participation in HE mainly due to the large
investment in education during the Soviet era. According to the OECD Education at a Glance
(2018), the Russian Federation has the fourth highest share (56%) of adults who have attained
tertiary education among all OECD and partner countries. The gross enrollment rate in tertiary
education has been growing steadily in Russia since 1995. During the 1995-2016 period, Russia
experienced a massive expansion of higher education further consolidating the massification of
higher education which has been common for many countries across the world (Trow, 2000;
Marginson, 2016). By the end of the Soviet era, Russia had a considerable stock of well-educated
individuals and strong higher education capacity due to the huge investments in the education
system in the Soviet Union. During the transition to a market economy, through radical economic
reforms and the economic recession of 1990-s, government expenditures on higher education
experienced a sharp decline. Despite the economic recovery in 2000-s the Russian government
still underinvest in higher education.
The expansion of higher education in Russia during the last two decades was mainly
driven by high demand for higher education and a large graduate wage premium, which arose as
a result of economic reforms and the creation of the market economy (Lukiyanova, 2010). High
demand for HE, low tuition fees and low expenditures on higher education has led, however, to
a significant differentiation of education by quality and a decrease of the average quality of HE.
HE expansion in Russia was achieved mainly by the appearance and growth of private
universities, the rapid growth of part-time programs, which were mainly focused on Business,
Administration and Law programs. These programs had very low quality and workload, low
requirements for enrollment and worked as ‘diploma factories’ (Roshchin, Rudakov, 2017).
However, the graduate wage premium has not decreased and holding a degree became a kind of
universal pass to access white-collar jobs. Its absence, in turn, became a strong negative signal
for employers (Andrushchak, Prudnikova, 2012).
However, the massification of HE in Russia has not led to the decrease of GWP and youth
unemployment, due to the demographic trends and absorption of graduates by service sector,
which has been growing since 1990-s (Rudakov, 2015). Regarding the composition of
employment, young Russian university graduates mainly concentrate in the service sector
(including trade, financial intermediation and retail), while the majority of Soviet university
graduates worked in manufacturing. In terms of payment, sectors such as Education, Agriculture,
to some extent Health and other mainly public sector industries are low paid, especially, in the
7
early stages of career and are less attractive to recent university graduates. While Education and
Health still have prestige in society despite the relatively low salaries, Agriculture seems to be a
sector with adverse selection of employees as the sector is both non-prestigious and low-paid:
high school graduates who failed to compete for state-funded education in other specialties, are
likely to choose agriculture due to relatively low competition for state funding for education in
this field.
The rapid expansion of higher education in Russia differs significantly by field of study:
the main increase was achieved by programs leading to a major in business, administration and
law. According to the OECD, in 2015, 38% of Russian tertiary students graduated with degrees in
business, administration and law, which is one of the highest shares across OECD and partner
countries (while the OECD average is 24%). However, this trend is gradually changing and among
new entrants in HEIs, engineering, manufacturing and construction are becoming the most
popular programmes in Russia (24%) although business, administration and law is still very
popular (22%) (OECD, 2018).
There is a broad discussion, debates and speculations among education authorities,
politicians, mass media and employers in Russia about overproduction of graduates in particular
spheres. For instance, there is a debate about the overproduction of specialists in economics,
management and business, whose education was financed by government while the state ‘needs
specialists in engineering’ for rapid development and economic growth. On the one hand, the
proponents of the overproduction hypothesis claim that most of these graduates could hardly
find a job related to their field of study and that this supply imbalance leads to a high job-
education mismatch and inefficiency of expenditures on higher education. On the other hand,
the results of previous research show that degrees in economics and law, together with degrees
in engineering bring higher returns in terms of wages (Denisova, Kartseva, 2007; Gimpelson, et
al 2009). It shows that in spite of the speculations about the overproduction of specialists in
economics and law, a degree in these fields is still a worthwhile investment and the market
absorbs these specialists and pays them good salaries, although they can be apparently job-
education mismatched. Research on this topic also warns us against using a static vision of
graduate labour markets and points out that many intermediate occupations have been
transformed or newly created (e.g: new managerial specialists) and that these make effective use
of graduate qualifications, not the least as an entry requirement (Elias, Purcell, 2013; Figueiredo
et.al, 2017).

8
 Another cliché is the importance of matching jobs with education trajectories that comes
as a legacy of the Soviet education system and an economy with a mandatory graduates’
distribution system among employees. From this point of view, job-education mismatch is seen
as an inefficiency of the higher education system and a waste of government funding. The listed
factors emphasize the importance of studying the issue of job-education mismatch and its labor
market consequences. Russia is an example of the extremely detailed system of educational
fields, which is a legacy of the Soviet educational system and economy with mandatory
distribution of graduates. In the Soviet period, multiple ministries and their affiliated enterprises
competed with each other for an ability to mandatory distribute graduates and some subfields,
especially in Engineering, were created to serve for particular enterprises. In that logics, the more
detailed and narrow is a subfield of study – the higher is the probability for the enterprise to
receive recent graduates (Gimpelson et al., 2010). After the transition to a market economy,
these subfields were too narrow and inadequate to labor market requirements, but the
education fields system has not been reformed yet. The extreme narrowness of the Russian
educational fields system is another argument for the relevance of the Russian case within the
international economic literature on this topic. On the basis of the considered theories and the
institutional peculiarities of the Russian higher education system, we formulate the following
hypothesis:
1. Job-education mismatches are more prevalent among graduates in fields, which
contribute to the accumulation of general human capital (Social Sciences, business, law,
Arts and humanities, Services)
2. Job education mismatches are less likely for graduates in fields, which contribute to
accumulation of specific human capital (Health, Engineering)
3. Job-education mismatches negatively affect salaries
4. The effect of job-education mismatches differs by fields of education:
- it is strongly negative for graduates in fields accumulating specific human capital and
insignificant for fields which accumulate general human capital.
- it is negative for low-paid sectors, but can be positive or insignificant for low-paid
sectors (e.g. Education and Agriculture).
3. Data and methodology
This study is based on the survey of Russian university graduates carried out in 2016 by
the Russian Federal Statistics Service as an additional module to the Russian Labor Force Survey.
The nation-wide (random) sample includes more than 20,000 observations on salaries as well as
9
other educational, socio-demographic and job characteristics of individuals who graduated from
HEIs between 2010 and 2015. All graduates were surveyed in 2016. This dataset contains data
on matching in the first job after graduation and in the current job of university graduates.
We calculate different measures of horizontal job - education mismatches following
existing research on this topic (Robst, 2007; Nordin, 2010; Boudarbat, Chernoff, 2012). We use
two different criteria: i) graduates’ self-evaluation of their job – education match; and ii) the
comparison of graduates’ ISCED97 fields of study with measures of the modal field(s) of study in
specific ISCO08 2-digit occupations with industry controls cells. The classification of the Russian
higher education system in accordance with the ISCED97 codes is presented in Table A1. We
calculated the modal ISCED field for each occupation, which was defined as a ISCO 2-digit code
*100 + code for industry 2. The statistical measure of mismatches reflects the actual distribution
of graduates of different fields of education by occupation. The graduate is considered as
matched if he/she has the same field of study as the mode of graduates in this job. Other
graduates are considered as mismatched. The self-evaluated measure of matching is used as the
main variable in our study.
We use logit regression (1) for the evaluation of the determinants of job-education
mismatches and OLS regression (2) for the calculation of the impact of mismatches on the
earnings of recent university graduates (based on the Mincerian wage equation).
𝑃𝑃𝑃𝑃(𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀ℎ𝑖𝑖 ) = 𝛽𝛽0 + 𝛽𝛽1 ∙ 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝑖𝑖 + 𝛽𝛽2 ∙ 𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝑖𝑖 + 𝛽𝛽3 ∙ 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑖𝑖 + 𝛽𝛽4 ∙ 𝑆𝑆𝑡𝑡𝑊𝑊𝑊𝑊𝑊𝑊𝑊𝑊 𝑖𝑖 + 𝜀𝜀 (1)
𝐿𝐿𝐿𝐿(𝑊𝑊𝑖𝑖 ) = 𝛽𝛽0 + 𝛽𝛽1 ∙ 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀ℎ𝑖𝑖 + 𝛽𝛽2 ∙ 𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝑖𝑖 + 𝛽𝛽3 ∙ 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝑖𝑖 + 𝛽𝛽4 ∙ 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑖𝑖 + 𝛽𝛽5 ∙ 𝑆𝑆𝑡𝑡𝑊𝑊𝑊𝑊𝑊𝑊𝑊𝑊 𝑖𝑖 + 𝛽𝛽6 ∙
𝐽𝐽𝐽𝐽𝐽𝐽𝑖𝑖 + 𝜀𝜀 (2)

Where:
Ln (Wi) is the logarithm of average monthly salary of university graduates
Matchi is a variable (dummy or categorical) which reflects either self-reported or statistical job-education match/mismatch
Fieldi is a categorical variable for the field of study of graduates
Acadi- is a set of academic factors (degree, type of institution (public/private), funding (self-funded, government-funded), type of
educational program (part-time, full-time))
SocDemi is a set of demographic factors (gender, age, marital status)
St_Worki is a dummy variable for combining study and work (during studies in university)
Jobi is a set of labor market factors (industry, congruence between first and current jobs);

2
As a robustness check, we also used detailed ISCO 3-digit jobs and 3-digit fields of study, but achieved almost the
same results as in the basic specification
10
 4. Descriptive analysis
In Table 1 we present the basic descriptive statistics. The average salary in our sample of
graduates is 24.069 Russian roubles (RR). Salaries are observed in 2016 for graduates of 2010-
2015. The average age of BA degree completion is 22, MA degree – 24 and Specialist degree- 23.
Taking into consideration the 5 year’s lag for graduates of 2010 and the methodology of the
Russian Federal Statistical Service, which defines youth as individuals aged between 15 and 30 3,
we restricted our sample to graduates aged between 20 and 30. The average age of graduates in
our sample is 26.1 years. Average tenure is 2.2 years with a maximum of 5 years, as due to the
survey methodology we account for tenure only after graduation. We do however include a
control variable on whether the graduate was combining study and work prior to graduation.
Table 1 - Basic descriptive analysis of the main continuous variables
Variable Obs Mean Std. Dev. Min Max
Salary 8177 24069 12044 5000 180000
Age 15155 26,1 2,1 20 30
Tenure 11525 2,2 1,6 0 5

The details about the sample are presented in table 2. The overall sample size is 15.155
graduates, who have completed university studies during 2010-2016. The number of graduates
in 2015 is considerably higher than in other years due to the emergence of the first wave of mass
graduation of BA students as a result of the reforms of 2011. As a result, 2015 was a year of
graduation of 5-years Specialists students and 4-years BA students. Average salary by year of
graduation grows with experience. Only 2014 graduates earn slightly less than the graduates of
2015, probably because of the economic crisis of 2014-2015. Differences between matching
status by year of graduation are insignificant and do not reflect any trend.

Table 2 – Descriptive analysis by year of graduation

Match (self-
Year of graduation N Percent Salary evaluation)
2010 2214 14,6 25906 68,5
2011 2252 14,9 25136 68,3
2012 2474 16,3 24658 68,0
2013 2509 16,6 24003 67,8
2014 2472 16,3 22570 66,4
2015 3234 21,3 22716 69,2
Total 15155 100,0 24096 68,1

3
Russian Federal Statistics Service definition of youth
11
 We use a number of different measures of job-education match: a categorical variable for
the self-evaluation of the current job-education matching status, a dummy variable for the same
measure, a categorical variable for the self-evaluation of the first job-education match, a variable
for statistical matching (on the basis of the mode ISCED97 fields for ISCO 2-digit occupations with
industry controls) (Table 3). More than half of university graduates (54%) report that they work
in a job which is related to their field of study, the partially matched are 14%. Partially
mismatched are 6% and absolutely mismatched in terms of their relatedness of their current job
to field of study are 26% of Russian graduates. The distribution of respondents by the matching
status is almost equal for the first and current jobs. Regarding salaries, there is a very well
observed trend, which shows that the higher the degree of matching of the job with education
obtained – the higher is the salary of university graduates.
Table 3 - Descriptive statistics of different matching variables
N Sample (%) Salary (RR) %
Current job-education match (self-evaluation)
Yes (perfect match) 6913 53,6 25481 18,4%
Rather, yes (good match) 1869 14,5 23702 10,1%
Rather, no (some mismatch) 802 6,2 22813 6,0%
No (absolute mismatch) 3320 25,7 21524 reference
Total 12904 100 24068
Current job-education match (self-evaluation)- dummy variable
Matched 8782 68,1 25103 15,3%
Mismatched 4122 31,9 21771 reference
First job-education match (self-evaluation)
Yes (perfect match) 7167 52,3 25458 17,0%
Rather, yes (good match) 1927 14,1 23620 8,5%
Rather, no (some mismatch) 874 6,4 23327 7,2%
No (absolute mismatch) 3,739 27,3 21767 reference
Total 13707 100 24068
Statistical job-education match (Mode ISCED fields for each industry)
Matched 8524 60,5 24109 0,4%
Mismatched 5560 39,5 24007 reference
Total 14084 100 24068
Vertical mismatch (statistical)
Horizontal match (specialists, managers) 9260 66,7 24840,8 2,10%
White-collar jobs (weak match) 2071 14,9 20388,8 -16,20%
Blue-collar jobs (mismatch) 2556 18,4 24321,1 reference
Total 13887 100

We created a dummy variable for the current job-education match, which is a stricter
version of the categorical variable: respondents, who answered that they are matched or ‘rather,

12
matched’ are considered as matched, and those who reported that they are ‘rather mismatched’
or ‘mismatched’ are considered as mismatched. According to this estimation, the share of
mismatched graduates by self-evaluation measure in Russia is 32%. The statistical measure of
mismatch shows that 39.5% of Russian graduates are mismatched, but the differences in average
salaries between matched and mismatched graduates are slight.
Additionally, we created the variable for vertical mismatches by using the ISCO
classification: university graduates who work as managers and professionals were considered as
horizontally matched, individuals in positions of ‘clerical support workers’, ‘services and sales
workers’ were considered as weakly matched (due to the fact that they are more likely to work
in ‘white-collar’ jobs). Individuals who worked as skilled workers, machine operators and in
elementary occupations were considered as mismatched. We found that 33,3% of Russian
graduates working in non-graduate jobs and are vertically mismatched.
Graduates who are vertically matched earn the highest salaries, but individuals who are
mismatched and work in blue-collar jobs earn 16.2% higher salaries than weakly matched
graduates in white-collar jobs, which is an unexpected result. This result provides some evidence
supporting the theory of compensating wage differentials (Rosen, 1986): university graduates,
who work as ‘blue-collars’, are more likely to achieve compensation for the nature of their job,
while those who are mismatched and decide to work as office clerks earn less. For the self-
evaluation vertical mismatch variable, we found that those graduates, who reported that their
education was useless in their first job, earn the lowest salaries.
The analysis of the correlation matrix shows that there is a very high correlation between
first job matching and current job matching.
Table 4. Correlation matrix of matching variables

First job- Statistical

Current job Vertical job
education job-
match (self- match in the
match (self- education
evaluation) first job
evaluation) match

Current job match (self-evaluation) 1.0000

First job-education match (self-evaluation) 0.8280*** 1.0000
Statistical job-education match 0.3618*** 0.2959*** 1.0000
Vertical job match in the first job 0.4541*** 0.3771*** 0.2029*** 1.0000
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

The correlation between the statistical job-education match and the self-evaluation
match is moderate (0.36), which shows that the usage of statistical measures of mismatch is

13
limited due to biases, which may happen during the construction of synthetic variables. The
correlation between vertical and horizontal mismatches, on the other hand, is rather strong
(Table 4).
Among the most popular degrees for Russian university graduates in our sample are those
in the fields of Social Sciences, Business and Law (35.3%), Arts and Humanities (16.4%), Education
(14.5%), Engineering, Manufacturing, and Construction (13.3%), Sciences, Math and IT (8.7%).
Other fields of studies are considerably less common (see Figure 1).

Engineering, Manuf &

Constr; 13,3
Education
Education; Agr&Vet; 3,2
Sciences, Maths and Agr&Vet
14,5
IT; 8,7 SS&Business&Law
Arts&Humanities
Services; 5,2
Health & Social Serv
Health & Social Serv; Services
3,5
Sciences, Maths and IT
SS&Business&Law;
35,3 Engineering, Manuf & Constr
Arts&Humanities; 16,4

Figure 1. Distribution of university graduates by ISCED fields of study (%).

The analysis of the probability of job-education mismatch by ISCED fields of education

shows that graduates with a degree in Health have the highest probability of being matched. The
probability of a match is also above average for graduates in Sciences, math and IT, Education,
Engineering, Arts and humanities, Social Sciences, business and law. The lowest probability of
being matched is for graduates in Services and in Agriculture (Figure 2). These results provide
some support for our hypothesis, that the probability of a mismatch is higher for disciplines which
provide general human capital (services, social sciences) and lower for disciplines which provide
very specific skills – Health. However, the probability of a mismatch for graduates in Arts and
Humanities is close to average, which contradicts our initial assumptions.
The high dispersion between the statistical and self-evaluated measures of mismatch is
connected to the imperfection of the statistical measure – it overestimates matching for fields
with high share of graduates (SS&Business and Law) and underestimate- for fields with low share
of graduates and/or that are dispersed between different occupations (Services).

14
 Health & Social Services 96,8
91,9
Sciences, Maths and IT 73,2
47,5
Education 70,4
60,6
Arts&Humanities 67,7
38,5
Engineering, Manuf & Constr 67,3
60,3
SS&Business&Law 65,9
83,2
Services 61,3
8,4
Agr&Vet 51,3
20,2

0,0 20,0 40,0 60,0 80,0 100,0 120,0

self-evaluation of matching statistical evaluation of matching

Figure 2. Share of matched graduates by ISCED fields of study (%)

Graduates in Engineering, Manufacturing and Construction and in Sciences, Math and IT
earn the highest salaries, while the lowest salaries are common for graduates in Education and
in Agriculture and Veterinary (Figure 3).

Engineering, Manuf & Constr 28,0

Sciences, Maths and IT 27,7
Services 26,4
Health & Social Serv 25,7
Average 24,1
Arts&Humanities 24,0
SS&Business&Law 22,6
Agr&Vet 21,9
Education 20,7

0,0 5,0 10,0 15,0 20,0 25,0 30,0

Figure 3. Average salary by ISCED fields of study (th.rub).

The descriptive analysis shows that the most popular industries for university graduates
are trade (16%), public administration and defence (14,5%), education (12,2%), manufacturing
(10,4%) and real estate, renting and business activities (9,7%). The highest salaries are earned by
graduates who work in mining and quarrying, the lowest – by those who work in education. The
highest probability of a self-evaluated job-education match is common for Health and Social
Work and Education (87-88%). The lowest – for graduates who work in trade (Table 5).

15
Table 5. Average salary and probability of job-education match by industry

Economic activity N % Salary (RR) Match(self) (%)

Mining and quarrying 225 1,6 36715 74

Construction 710 5 27741 64,7
Transport, storage and communication 1021 7,3 26891 59,5
Real estate, renting and business activities 1361 9,7 26313 77,9
Electricity, gas and water supply 373 2,7 24831 75,4
Manufacturing 1457 10,4 24648 67
Financial intermediation 902 6,4 24562 74,4
Public administration and defence 2042 14,5 24503 71,4
Health and Social work 808 5,7 23193 87,5
Trade 2242 15,9 22534 44
Other community, social and personal activities 583 4,1 22260 59,1
Hotels and restaurants 303 2,2 21865 44,8
Agriculture 344 2,4 20472 57,9
Education 1713 12,2 19605 87,9
Total 14084 100 24068 68,1

The descriptive analysis by professional groups shows that the majority of university
graduates work as professionals (59%) and managers (6.6%) and are vertically matched.
Graduates, who work as professionals also have the highest probability to be horizontally
matched. Vertically matched graduates who work as managers or professionals earn higher
salaries than weakly matched graduates who work in white-collar jobs. However, graduates who
are vertically mismatched but work in blue-collar jobs earn considerably high salaries, probably,
due to compensating wage differentials (Table 6).
Female graduates are more likely than male graduates to work in a job which is related to
their university field of study. There is a considerable gender wage gap: on average, male
graduate earnings are 27% higher than female graduates (see Table 7). Graduates who combined
their studies with a job are less likely to be job-education matched after graduation, which can
be endogeneous due to reversed causality: individuals who understood during their studies that
they will not work in the field related to the study start to combine study and work. On average,
graduates who combined studies with part-time jobs earn, after graduation, slightly more.
Graduates who combined studies with work related to their field of study are much more likely
to be job-education matched and earn higher salaries (see Table 7).

16
Table 6. Average salary and probability of job-education matching by ISCO professional groups.

Vertical matching Salary Match(self)

Major groups (ISCO-08) N %
status (RR) (%)

Managers Vertically 932 6,6 28746 61,5

Professionals matched 8328 59,1 24446 87,1
Technicians and Associate Professionals Weakly matched 1571 11,2 23685 53,1
Clerical Support Workers (white-collar jobs 608 4,3 20139 38,9
Services and Sales Workers mostly) 1463 10,4 20494 23,1
Skilled Agricultural, Forestry and Fishery
84 0,6 21869 9,6
Workers Vertically
Craft and Related Trades Workers mismatched 490 3,5 24753 34,5
Plant and Machine Operators and (blue-collar jobs
411 2,9 26220 21,7
Assemblers mostly)
Elementary Occupations 198 1,4 20841 0
Total 14085 100 24068 68,1

Table 7. Average salary and probability of job-education match by gender group and combining
study and work.
N % Salary (RR) Match(self) (%)
Gender
Female 9080 59,9 21526 70
Male 6075 40,1 27399 65,6
Total 15155
Сombining of study and work
Combined study with regular job 3124 20,6 24412 63,5
Combined study with part-time job 2692 17,8 24669 66,2
Did not work during studies 9339 61,6 23766 70,3
Total 15155
Combining of study and work by relatedness to field of study
Combined studies with work non-related
3062 52,7 23532 43,1
to field of study
Combined studies with work related to
2754 47,4 25571 87,7
field of study
Total 5816

The descriptive analysis shows that the optimal strategy of school-to-work transition is
connected to combining studies with work in a related field in order to accumulate necessary
work experience, which will be valued by employers after graduation. These results are
consistent with previous studies for the Russian labor market (Rudakov, Roshchin, 2017).
We also made a cross-tabulation of industries where graduates work and their ISCED
fields of study (see Table A1). According to our results, graduates in Health and Education fields
17
are considerably more likely to be matched: 80% of graduates with a degree in Health, work in
the Health and Social work industry and 59% of graduates with a degree in Education, work in
the Education industry. Graduates of Education and especially Health are likely to accumulate
specific human capital, which can be hardly transferred to other sectors, and graduates prefer to
work in a related industry. Although graduates in Engineering are quite dispersed by industries,
most of these industries are related to engineering: 25% work in Manufacturing, 15% - in
Construction and 9% - in Electricity, gas and water supply. However, 10% of graduates in
Engineering work in Trade and other 10% in Real estate, renting and business activities.
Graduates in Sciences, Math and IT are among the most dispersed between the fields: 23% work
in Real estate, renting and business activities, 13% work in Trade industry, 13% - in Manufacturing
and the rest – in multiple other industries. This may occur due to the fact that IT specialists are
needed in every industry and many fields require strong mathematical and IT skills, which are
getting universal for many businesses.
Graduates in spheres which are frequently classified as those which generate general
human capital are also likely to be dispersed between sectors and are frequently working in
industries which are far from their university specializations. For instance, 30% of graduates with
degrees in Arts and Humanities work in Public Administration, 15% - in real estate, renting and
business activities, 11% - in trade, 10% - in other social community, social and personal activities.
Graduates in Social Sciences, business and law frequently work in Trade (25%), Public
administration (15%), Financial intermediation (13%). Graduates in Services work in Transport,
storage and communication industry (27%), Trade (15%) and others (Table A1). In general, the
analysis shows that graduates in Health and Education prefer working in related industries,
graduates in Engineering also frequently work in industries connected to engineering, but also
work in other fields. Graduates of fields like Arts and Humanities, Social Sciences, Services,
Agriculture, Sciences, math and IT are considerably dispersed between different industries. This
evidence supports the idea of classification of ISCED fields on ‘general’, whose graduates work in
different areas, and ‘specific’, whose graduates work in related fields due to limited
transferability of human capital.

18
Table A1. Сross-tabulation of industry of current job and ISCED field of study (%)

Social Sciences Engineering,

Arts& Sciences, Agriculture&V Health &
Industry (OKZ)/ ISCED Code Education &Business and Manufacturin, Services Total
Humanities Math and IT eterinary Welfare
Law Construction

1 Agriculture and fishing 1,6 1,0 2,2 1,7 1,7 24,9 0,4 2,0 2,5
2 Mining and quarrying 0,3 0,5 0,9 7,8 2,5 1,1 0,0 1,6 1,6
3 Manufacturing 3,4 6,4 9,3 13,3 25,2 10,6 0,4 8,9 10,2
4 Electricity, gas and water
supply 0,7 1,2 1,9 3,1 9,3 3,2 0,0 1,2 2,7
5 Construction 1,9 3,1 3,9 3,9 15,1 4,5 0,0 6,5 5,0
6 Trade 8,2 10,8 24,7 13,3 10,3 15,3 13,8 15,1 15,9
7 Hotels and restaurants 1,2 1,7 2,8 1,1 2,5 2,3 0,0 4,5 2,2
8 Transport, storage and
communication 2,7 4,2 7,6 7,8 8,9 5,8 0,0 26,7 7,2
9 Financial intermediation 2,4 4,1 13,4 4,5 1,5 0,9 0,0 1,7 6,4
10 Real estate, renting and
business activities 3,8 14,9 7,8 22,7 10,0 3,8 0,0 8,9 9,6
11 Public administration and
defence 8,0 30,8 14,8 8,1 9,2 11,9 2,6 14,7 14,5
12 Education 58,8 8,1 3,8 8,1 1,5 2,3 0,8 2,4 12,2
13 Health and Social work 2,4 3,3 3,8 1,9 0,7 12,1 80,1 1,5 5,8
14 Other community, social
and personal activities 4,7 9,9 3,2 2,8 1,5 1,4 1,8 4,4 4,2
Total 100 100 100 100 100 100 100 100 100

19
 5. Results.
Determinants of job-education matching
The regression analysis of the determinants of job-education matching, based on the logit
regression, was conducted for self-evaluated matching. We found that graduates with a degree
in Health are considerably more likely to work in a related field (29%).

Table 8. Regression analysis of the determinants of job-education matching (marginal effects)

1
VARIABLES Self-evaluated matching
ISCED field of study: reference group 'Education'
Arts and Humanities -0.00531
(0.0151)
SS&Business&Law -0.0268**
(0.0132)
Sciences, Maths and IT 0.0518***
(0.0169)
Engineering, Manuf & Constr 0.000246
(0.0158)
Agriculture&Veterinary -0.166***
(0.0268)
Health & Social Services 0.289***
(0.0110)
Services -0.0629***
(0.0215)
Gender=male -0.0531***
(0.00907)
Age 0.00931***
(0.00287)
Combining study and work: reference group 'have not worked during studies'
Combined studies with regular job -0.0147
(0.0150)
Combined studies with part-time job 0.0109
(0.0142)
Funding = state-funded 0.0316***
(0.00914)
Degree= MA or Specialist 0.00543
(0.0116)
Full_time studies 0.0649***
(0.0144)
Public university 0.0001
(0.0157)
Observations 12845
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

This result supports the hypothesis, formulated in our study following Robst (2007), that
graduates who have accumulated specific human capital during their studies (e.g. in Medicine)
are less likely to be mismatched. This specific human capital cannot be easily transferred to other
20
sectors, and graduates in these spheres are less likely to search for a job in other sectors. They
are more likely to work in a job which is directly related to their field of study in order to use
specific human capital, which was accumulated during studies. The empirical analysis shows that
graduates in Agriculture and Veterinary, in Services and in Social Sciences, business and law have
the highest probability of being mismatched (see Table 8).
These results partially support our hypothesis that in fields which contribute mainly to the
accumulation of general human capital or in low-paid fields, the probability of a mismatch is
higher due to the easy transfer of human capital to other spheres which are not directly related
to the field of study and the lower opportunity costs. Mismatches for graduates in these fields
are less likely to lead to a waste of human capital and a wage penalty. Graduates in these fields
have an opportunity to find the best match between their field of study, abilities and job
requirements, and switch between different sectors of the economy.
Regarding other variables, female graduates are more likely to work in the jobs which are
related to their field of study (for a self-evaluated match). Graduates who studied on a state-
funded basis and those who studied in full-time programs are considerably more likely to be
matched in terms of job and education. The probability of a job-education match also slightly
increases with the age of graduates. We did not find any significant differences in the probability
of matching by degree and the type of university (public or private) (see Table 8).

The impact of the job-education match on graduates’ salaries

The regression analysis suggests that there is a significant and robust negative impact of
job-education mismatches on graduate salaries for the different self-evaluated and statistical
measures (Table 10). The regression analysis shows that the mismatch wage penalty increases
with the degree of mismatch. Graduates who reported a good match earn 7% lower salaries than
those who reported a perfect match, graduates who reported some mismatch experience an 11%
wage penalty, those who reported an absolute mismatch experience the most severe penalty –
15.5% compared to perfectly matched graduates. The impact of the mismatch in the first and in
the current job on current salaries is almost equal. The dummy variable for self-evaluation of
matching status shows that mismatched graduates earn 12% lower salaries than matched
graduates. The wage penalty for the statistical mismatch is twice as lower as a penalty for a self-
evaluated one: statistically mismatched graduates earn 6% lower salaries. This can be explained
by the higher accuracy of self-evaluated measures but also may indicate that graduates are more
likely to report that they are matched when they earn higher salaries.
21
Table 10. The impact of job-education mismatch on graduate salaries (OLS regression).
(1) (2) (3) (4)
VARIABLES lnwage lnwage lnwage lnwage
Matching in the current job- reference group (perfect match)
Good match -0.0708***
(0.0141)
Some mismatch -0.109***
(0.0208)
Absolute mismatch -0.155***
(0.0138)
Matching in the first job- reference group (perfect match)
Good match -0.0725***
(0.0142)
Some mismatch -0.0899***
(0.0205)
Absolute mismatch -0.131***
(0.0130)
Self-evaluated mismatch -0.122***
(0.0120)
Statistical mismatch -0.0609***
(0.0112)
Gender= male 0.173*** 0.173*** 0.173*** 0.170***
(0.0108) (0.0108) (0.0108) (0.0108)
ISCED field of study: reference group 'Education'
Arts and Humanities 0.0185 0.0247 0.0228 0.0401**
(0.0197) (0.0196) (0.0196) (0.0197)
SS&Business&Law -0.0219 -0.0115 -0.0217 -0.0322*
(0.0184) (0.0183) (0.0183) (0.0188)
Sciences, Maths and IT 0.0222 0.0304 0.0255 0.0378*
(0.0229) (0.0229) (0.0229) (0.0229)
Engineering, Manuf & Construction 0.0708*** 0.0792*** 0.0731*** 0.0785***
(0.0215) (0.0214) (0.0214) (0.0216)
Agriculture -0.0250 -0.0236 -0.0243 -0.0108
(0.0315) (0.0316) (0.0316) (0.0319)
Health 0.0951*** 0.103*** 0.106*** 0.111***
(0.0349) (0.0349) (0.0349) (0.0353)
Services 0.0626** 0.0705*** 0.0651** 0.0992***
(0.0260) (0.0260) (0.0260) (0.0263)
Age 0.0199*** 0.0199*** 0.0201*** 0.0205***
(0.00339) (0.00339) (0.00339) (0.00341)
Marital status 0.0165 0.0168 0.0148 0.0160
(0.0104) (0.0104) (0.0104) (0.0105)
Combining study and work: reference group 'have not worked during studies'
Combined studies with regular job 0.104*** 0.102*** 0.105*** 0.106***
(0.0168) (0.0168) (0.0168) (0.0169)
Combined studies with part-time job 0.0399*** 0.0423*** 0.0407*** 0.0413***

22
 (0.0129) (0.0130) (0.0130) (0.0130)
Funding = state-funded 0.00193 0.000278 0.00325 0.00272
(0.0104) (0.0104) (0.0104) (0.0105)
Degree= MA or Specialist 0.0125 0.0117 0.0133 0.0124
(0.0136) (0.0136) (0.0136) (0.0136)
Full_time studies 0.136*** 0.132*** 0.138*** 0.139***
(0.0166) (0.0166) (0.0166) (0.0166)
Public university -0.0823*** -0.0777*** -0.0841*** -0.0860***
(0.0191) (0.0192) (0.0192) (0.0193)
Vertical matching: reference group (horizontaly mismatched - white collars)
Horizontally matched 0.120*** 0.142*** 0.132*** 0.186***
(0.0160) (0.0155) (0.0159) (0.0148)
Horizontally mismatched (blue-collars) 0.0840*** 0.0897*** 0.0828*** 0.102***
(0.0172) (0.0172) (0.0172) (0.0172)
Year of graduation: reference group (2010)
2011 0.00365 0.00238 0.00470 0.00220
(0.0173) (0.0173) (0.0173) (0.0174)
2012 -0.00295 -0.00300 -0.00320 -0.00435
(0.0177) (0.0177) (0.0177) (0.0178)
2013 -0.0141 -0.0161 -0.0146 -0.0160
(0.0189) (0.0189) (0.0189) (0.0190)
2014 -0.0400** -0.0380* -0.0404** -0.0412**
(0.0203) (0.0204) (0.0204) (0.0205)
2015 -0.0304 -0.0284 -0.0300 -0.0311
(0.0220) (0.0220) (0.0220) (0.0221)
First job is not a current job -0.0338*** -0.0235* -0.0372*** -0.0472***
(0.0119) (0.0122) (0.0119) (0.0119)
Industry: reference group (manufacturing)
Agriculture, hunting and forestry -0.188*** -0.185*** -0.187*** -0.176***
(0.0370) (0.0371) (0.0371) (0.0372)
Mining and quarrying 0.326*** 0.329*** 0.330*** 0.342***
(0.0382) (0.0383) (0.0383) (0.0384)
Electricity, gas and water supply -0.0476 -0.0445 -0.0439 -0.0424
(0.0305) (0.0305) (0.0305) (0.0307)
Construction 0.0996*** 0.0958*** 0.0992*** 0.0970***
(0.0246) (0.0247) (0.0247) (0.0248)
Trade 0.0226 0.0153 0.0188 0.0104
(0.0193) (0.0193) (0.0193) (0.0194)
Hotels and restaurants 0.0109 0.0113 0.00975 0.0138
(0.0367) (0.0368) (0.0368) (0.0369)
Transport, storage and communication 0.113*** 0.113*** 0.114*** 0.118***
(0.0227) (0.0228) (0.0228) (0.0229)
Financial intermediation 0.0733*** 0.0753*** 0.0727*** 0.0775***
(0.0243) (0.0243) (0.0243) (0.0244)
Real estate, renting and business activities 0.0437** 0.0445** 0.0434** 0.0501**
(0.0208) (0.0209) (0.0209) (0.0210)
Public administration and defence; -0.0164 -0.0144 -0.0181 -0.00849
(0.0192) (0.0192) (0.0193) (0.0193)

23
 Education -0.201*** -0.193*** -0.195*** -0.191***
(0.0230) (0.0230) (0.0230) (0.0232)
Health and Social work -0.100*** -0.0977*** -0.0981*** -0.0882***
(0.0283) (0.0283) (0.0283) (0.0284)
Other community, social and personal activities -0.0447 -0.0453 -0.0479* -0.0538*
(0.0286) (0.0286) (0.0286) (0.0287)
Constant 9.293*** 9.262*** 9.136*** 9.196***
(0.103) (0.103) (0.103) (0.103)
Observations 8026 8026 8026 8026
R-squared 0.177 0.174 0.174 0.166
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

We found evidence of a considerable gender wage gap as male graduates earn 17% higher
salaries than their female counterparts. Graduates who combined studies with full-time or part-
time jobs earn respectively 10% and 4% higher salaries than those who did not get work
experience. There are no statistically significant differences in earnings between graduates who
studied on state-funded and fee-basis and those who got Specialists or Master and Bachelor
degrees. Graduates who studied full-time, earn 13-14% higher salaries than those who studied
part-time. Horizontal matching is a significant predictor of graduate salaries. Horizontally
matched graduates earn 9-19% higher salaries compared to horizontally mismatched graduates
who work in blue-collar jobs. Those graduates who are horizontally mismatched but work in blue-
collar jobs earn 9% higher salaries than those who are mismatched and work in white-collar jobs.
This is an interesting result, which can be partially explained by compensating wage differentials
– graduates who work in blue-collar jobs receive compensation for working in less-prestigious,
non-graduate jobs, which has negative non-pecuniary characteristics, and substitute graduates
of vocational courses, who should take these workplaces. Employment in a white-collar non-
graduate job is likely to be an adverse selection, which is not compensated for due to the lack of
negative non-pecuniary characteristics. Working in an office is less likely to result in health
damage. Moreover, these workplaces are more likely to be occupied by female graduates.
Taking employment in manufacturing as a reference group, we found that graduates who
are employed in Mining and Quarrying earn the highest salaries (32-33% higher), graduates who
work in Transport, storage and communication, in Financial intermediation and in Real estate,
renting and business activities earn respectively 11%, 7% and 4% higher salaries compared to
reference group. We also found a wage penalty for working in public sector industries, such as
Education (-20%) and Health (-10%) and penalty for employment in Agricultural sector (-18%).

24
 Differences in the impact of job-education matching on graduates’ salaries by fields of study
The overall effect of the ISCED fields of study can be hidden by the simultaneous inclusion
of controls for fields of study, industries and matching 1. In order to avoid this problem, we used
alternative specifications without industry controls. To reveal the differences in the effect of job-
education matching by fields of study we used specifications with interaction terms between
matching (self-evaluated and statistical) and the ISCED fields. The only difference between this
model and the previous one is the absence of industry controls, all other control variables are
the same. However, for simplicity we show the results only for the main outcome variables, as
the other results are almost the same.
Regarding our key variable, the magnitude of the negative impact of job-education
mismatches decreases slightly, from 12 to 11% for the self-evaluated matching status and from
6 to 4% in the case of the statistical matching. The results also indicate considerable differences
in the impact of the ISCED fields of study on graduates’ salaries compared to the full specification
with industry controls. In the basic specifications (1-2) all dummies for fields of study are
significant. Taking Education as a reference group, we found that graduates in all fields except
for Agriculture earn significantly higher salaries: 20% in Engineering, manufacturing and
construction, 20% in Services, 16%-18% in Sciences, math and IT, 14-16% in Health, 12% in Arts
and humanities and 10% in Social sciences, business and law (Table 11).
The inclusion of interaction terms for matching and ISCED fields of study in the regression
model (3-4) shows that self-evaluated mismatch is mostly penalized in Engineering, Services, Arts
and Humanities (relative to matched graduates in those fields). However, the differences
between the self-evaluated matched and mismatched graduates are insignificant for Education,
Social Sciences, business and law as well as Health and Agriculture (Table 11). The analysis of the
interaction effects for the statistical mismatch shows that mismatched Education graduates have
the lowest relative penalty and that penalties for mismatch exists in all other sectors except for
Agriculture. These results can be explained by the fact that Agriculture and Education are the
sectors with the lowest salaries, and those, who are mismatched, may have chosen other sectors,
which are better paid. These results support our hypothesis, that penalties for mismatch may
differ between low-paid and well-paid sectors.

1
VIF-statistics values for model estimated in Table 10 with industry controls are between 2.79 and 2.85, which
shows that multicollinearity is not a significant problem. After removal of industry controls VIF-statistics values
decreases to 1.68 - 1.72, showing even lower value of possible multicollinearity
25
 However, the results regarding the interaction terms do not support our hypothesis, that
penalties for mismatch differ between fields which contribute to the accumulation of general or
specific human capital. For instance, despite our prediction, that graduates in Health are likely to
have the most considerable penalty for mismatch because human capital accumulated during
health studies can be hardly transferred to other sectors without considerable waste was not
supported by results. Graduates in Health are not penalized for being mismatched in
specification, which used self-evaluation measure. However, these results for Health field could
be achieved because of low number of graduates in Health and very low number of mismatched
graduates in this field.
Table 11. The impact of job-education mismatches on graduate salaries (OLS regression) –
specifications without industry controls and with the inclusion of interaction terms
1 2 3 4
VARIABLES lnwage lnwage lnwage lnwage
Self-evaluated mismatch (dummy) -0.110*** -0.0453
(0.0121) (0.0292)
Statistical mismatch (dummy) -0.042*** 0.0538**
(0.0112) (0.0266)
ISCED field of study: reference group 'Education'
Arts and Humanities 0.118*** 0.132*** 0.147*** 0.165***
(0.0176) (0.0179) (0.0209) (0.0259)
SS&Business&Law 0.102*** 0.0952*** 0.116*** 0.139***
(0.0153) (0.0156) (0.0181) (0.0189)
Sciences, Maths and IT 0.166*** 0.175*** 0.191*** 0.247***
(0.0209) (0.0210) (0.0243) (0.0284)
Engineering, Manuf & Construction 0.202*** 0.206*** 0.235*** 0.239***
(0.0187) (0.0188) (0.0221) (0.0237)
Agriculture 0.0319 0.0412 0.00502 -0.0636
(0.0295) (0.0299) (0.0391) (0.0636)
Health 0.150*** 0.160*** 0.164*** 0.201***
(0.0285) (0.0287) (0.0299) (0.0314)
Services 0.204*** 0.229*** 0.242*** 0.461***
(0.0241) (0.0247) (0.0297) (0.0664)
Interactions of education and matching (self-evaluated and statistical)
Arts and Humanities##Mismatch -0.0978** -0.0877**
(0.0382) (0.0365)
SS&Business&Law ## Mismatch -0.0491 -0.134***
(0.0330) (0.0341)
Sciences, Maths and IT ##Mismatch -0.0879* -0.163***
(0.0460) (0.0412)
Engineering, Manuf & Constr ## Mismatch -0.107*** -0.0788**
(0.0393) (0.0370)
Agriculture##Mismatch 0.0351 0.0806
(0.0595) (0.0727)
26
 Health##Mismatch 0.0242 -0.146*
(0.133) (0.0861)
Services## Mismatch -0.117** -0.030***
(0.0504) (0.0727)
Gender= male 0.194*** 0.191*** 0.190*** 0.188***
(0.0108) (0.0108) (0.0108) (0.0110)
Combining study and work: reference group 'have not worked during studies'
Combined studies with regular job 0.106*** 0.107*** 0.105*** 0.105***
(0.0171) (0.0172) (0.0171) (0.0171)
Combined studies with part-time job 0.0411*** 0.0411*** 0.0415*** 0.0402***
(0.0132) (0.0132) (0.0132) (0.0132)
Vertical matching: reference group (horizontaly mismatched - white collars)
Vertically matched 0.108*** 0.162*** 0.113*** 0.172***
(0.0156) (0.0142) (0.0156) (0.0145)
Vertically mismatched (blue-collars) 0.0858*** 0.104*** 0.0886*** 0.108***
(0.0171) (0.0170) (0.0171) (0.0173)
Constant 9.030*** 9.073*** 9.073*** 9.340***
(0.103) (0.104) (0.104) (0.106)
Observations 8026 8026 8026 8026
R-squared 0.140 0.134 0.142 0.137
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

Conclusion

This study has focused on horizontal job-education mismatches (alternatively,

occupation-education mismatches), an issue which has become increasingly relevant given the
debates about mass higher education and its relevance to labor market requirements. By using
a comprehensive dataset on Russian university graduates we evaluated the determinants of job-
education mismatches and their consequences in terms of salaries. We used two different
measures of job-education mismatches: one (the main) based on a self-evaluated mismatch and
a statistical mismatch variable.
Both measures showed that there is a considerable wage penalty for being mismatched
and that the penalty rises with the extent of the horizontal mismatch. Our analysis showed that
32% of Russian graduates are mismatched by self-evaluation measures (a value that increased to
40% when we used the statistical measure). We also found that there are considerable
differences in the probability of job-education mismatches by fields of study. Graduates in fields
which generate rather general human capital (e.g., Social Sciences, business, law, Services) or in
relatively low-paid and non-prestigious fields (Agriculture) are more likely to be mismatched.
Graduates in fields which generate specific human capital (e.g. Health) are considerably more
likely to be matched. This can be explained by the relative non-transferability of specific human

27
capital to other sectors. On the contrary, general human capital, that can more easily be
transferred to other industries, makes mobility between fields beneficial and increases the
probability of a mismatch, though it reduces the penalty associated with it. Our analysis has
shown that job-education mismatches negatively affect salaries of university graduates and the
wage penalty for mismatches depends considerably on the degree of mismatch. The analysis of
interaction terms shows, that mismatch is not penalized in low-paid fields (Agriculture and
Education) and is penalized in more well-paid occupations.
Regarding vertical mismatches, we found that 33% of Russian graduates work in non-
graduate jobs. There is a strong penalty for working in white-collar non-graduate jobs compared
to white-collar graduate jobs, but those who work in blue-collar non-graduate jobs are not
penalized for vertical mismatches. The latter, on the contrary, earn more than those who work
in white-collar non-graduate jobs. This can be explained by compensating wage differentials, i.e.,
graduates that work in blue-collar jobs receive a compensation for working in less-prestigious
non-graduate jobs which have negative non-pecuniary characteristics. Graduates who work in
white-collar non-graduate jobs seem to be those who lost in the competition with graduates who
work in white-collar graduate jobs, and earn less due to adverse selection and lack of
compensation (as non-pecuniary job characteristics are neutral rather than negative).
Our results conform, in general, to human capital theory, which predicts that job-
education mismatches will be penalized in terms of salaries due to a potential waste of human
capital. The results mainly support the idea, formulated in Robst (2007) Nordin et al (2010) and
Boudarbat, Chernoff (2012), that graduates in fields which accumulate more specific human
capital are less likely to be mismatched relative to graduates in fields which contribute to the
development of general human capital. Unlike Robst (2007) and Nordin (2010), we found that
almost all fields, except for low-paid Agriculture and Education are penalized for mismatch.
We believe that our study will foster the discussion about horizontal dimension of job-
education mismatch, which is important as it shows the congruency between educational fields
of study and labor market requirements. The analysis proved that penalties for horizontal
mismatch are considerable and affect salaries to an even higher extent than just vertical
mismatch, a much better documented issue.

28
References:
1. Andrushchak, G., & Prudnikova, A. (2012). Gde luchshe uchitsya: differentsiatsiya
zarabotkov vypusknikov vuzov [The best universities to study in: Differentiation of
graduates’ salaries]. Vestnik RLMS, 2, 129-136.
2. Baert, Stijn, Bart Cockx, and Dieter Verhaest. "Overeducation at the start of the career:
stepping stone or trap?." Labour economics25 (2013): 123-140.
3. Beaudry, P., Green, D.A. and Sand, B.M., 2016. The great reversal in the demand for skill
and cognitive tasks. Journal of Labor Economics, 34(S1), 199-247.
4. Boudarbat, B., & Chernoff, V. (2012). Education–job match among recent Canadian
university graduates. Applied Economics Letters, 19(18), 1923-192
5. Chevalier, A., & Lindley, J. (2009). Overeducation and the skills of UK graduates. Journal
of the Royal Statistical Society: Series A (Statistics in Society), 172(2), 307-337.
6. Denisova, I. A., Kartseva, M. A. (2007). A premium for a degree in engineering: An
estimation of returns to the field-specific education in Russia. Prikladnaya ekonometrika=
Applied Econometrics, (1), 30-57.
7. Domadenik, P., Farcnik, D., & Pastore, F. (2013). Horizontal Mismatch in the Labour
Market of Graduates: The Role of Signalling.
8. Elias, P., & Purcell, K. (2004). Is mass higher education working? Evidence from the labour
market experiences of recent graduates. National Institute Economic Review, 190(1), 60-
74.
9. Figueiredo, H., Biscaia, R., Rocha, V., & Teixeira, P. (2017). Should we start worrying? Mass
higher education, skill demand and the increasingly complex landscape of young
graduates’ employment. Studies in Higher Education, 42(8), 1401-1420.
10. Gimpelson, V. E., Kapelyushnikov, R. I., Karabchuk, T. S., Ryzhikova, Z. A., & Bilyak, T. A.
(2009). Vybor professii: chemu uchilis’i gde prigodilis’ [A choice of profession: what and
where to learn and be useful?]. Ekonomicheskii zhurnal Vysshei shkoly ekonomiki=
Economic Journal of Higher School of Economics, 13(2), 172-216.
11. Green, F., and G. Henseke. "Graduates and ‘graduate jobs’ in Europe: a picture of growth
and diversification." Centre for Global Higher Education, London UK, working paper
series 25 (2017): 47.
12. Hartog, J. (2000). Overeducation and earnings. Where are we, where should we go?
Economics of Education Review, 19(2), 131–147.

29
13. Goldin, C, and L. Katz. The race between education and technology. Harvard university
press, 2009.
14. Lamo, A., & Messina, J. (2010). Formal education, mismatch and wages after transition:
Assessing the impact of unobserved heterogeneity using matching estimators. Economics
of Education Review, 29(6), 1086-1099.
15. Lukiyanova A. Returns to education in Russia: Evidence from meta-analysis: Working
paper WP3/2010/03.– Moscow: State University – Higher School of Economics, 2010. –
60 p.
16. Marginson, S. (2016). The worldwide trend to high participation higher education:
Dynamics of social stratification in inclusive systems. Higher Education, 72(4), 413-434.
17. Mavromaras, K., & McGuinness, S. (2012). Overskilling dynamics and education
pathways. Economics of Education Review, 31(5), 619-628.
18. McGuinness, S. (2006). Overeducation in the labour market. Journal of economic
surveys, 20(3), 387-418.
19. McGuinness, S., & Sloane, P. J. (2011). Labour market mismatch among UK graduates: An
analysis using REFLEX data. Economics of Education Review, 30(1), 130-145.
20. Meroni, E. C., & Vera-Toscano, E. (2017). The persistence of overeducation among recent
graduates. Labour Economics, 48, 120-143.
21. Nordin, M., Persson, I., & Rooth, D. O. (2010). Education–occupation mismatch: Is there
an income penalty?. Economics of education review, 29(6), 1047-1059.
22. Robst, J. (2007). Education and job match: The relatedness of college major and
work. Economics of Education Review, 26(4), 397-407.
23. Robst, J. (2007b). Education, college major, and job match: Gender differences in reasons
for mismatch. Education Economics, 15(2), 159–175.
24. Rosen, S. (1986). The theory of equalizing differences. Handbook of labor economics, 1,
641-692.
25. Rudakov, V. N. "Dynamics of salary of the youth in the labor market in Russia." Issues of
statistics 11 (2015): 43-57 (in Russian).
26. Trow, M. (2000). From mass higher education to universal access: The American
advantage. Minerva, 37(4), 303-328.
27. Wolbers, M. H. (2003). Job Mismatches and their Labour Market Effects among School
Leavers in Europe. European Sociological Review, 19(3), 249-266.

30
Table A1. Higher education fields in Russia encoded in accordance with ISCED97 codes.

Russian HE specialties encoded in accordance with ISCED97 codes

1 Education
Education and Pedagogics 65
2 Humanities and arts
Arts and Culture 67
Human Sciences 63
3 Social sciences, business and law
Social Sciences, Law 64
Economics and Management 68
4 Science
Physical and Mathematical Sciences 61
Natural Sciences 62
Information security 69
Informatics and Computers 622
Geology, Mineral Resources Exploring and Mining 628
5 Engineering, manufacturing and construction
Power Engineering, Power Engineering Industry and Electrical
Engineering 613
Metallurgy, Mechanical Engineering and Material Processing 614
Aviation and Rocket and Space Machinery 615
Weapons and weapons systems 616
Maritime Machinery 617
Instrument Making and Optical Equipment 619
Electronic Engineering, Radio Engineering and Communication 620
Automatic Devices and Management 621
Food and Consumer Goods Technology 624
Architecture and Construction 625
Chemical Technology and Biotechnology 627
6 Agriculture
Agriculture and Fishing Industry 611
Reproduction and Processing of Forest Resources 623
Geodesy and Land Management 612
7 Health and welfare
Public Health Service 66
8 Services
Service Trades 610
Transport 618
Personal and Social Safety, Environmental Engineering and Protection 626

31