Determinants of youth unemployment rate: case of Slovakia

Abstract

Research background: The employment rate of young individuals in the labour market has considerably decreased in developed countries recently. Due to lower labour capital, skills, and generic and job-specific work experience, youth consider finding suitable job challenging. If they fail to succeed in the labour market soon after graduation, it leads to long-term unemployment, unstable and low-quality jobs, and even social exclusion. Purpose of the article: This paper aims to analyse the unemployment rate of high school-graduated students and the factors impacting this unemployment rate, such as GDP per capita, total unemployment rate, apartment price per square meter and results from state exams. Identifying the determinants affecting youth unemployment is crucial for theoretical knowledge and for policymakers to ensure youth inclusion in the economic mainstream. As a result, society can reduce social and economic costs and avoid structural problems in the future. Methods: Data about 464 Slovak high schools from National Institute for Certified Educational. Data include the graduate unemployment rate for each high school in Slovakia. Furthermore, two logistic regression models have been developed to investigate the impact of selected factors on high school graduates? unemployment rate immediately after graduation and nine months after graduation. Findings & value added: This paper indicates the existence of statistical dependency between unemployment of high school graduates and overall unemployment rate in the region, GDP per capita in the region, quality of high school education and cost of living in the region immediately after graduation. Analysis of the period nine months after graduation has shown the important decline of education quality provided by high schools. To reduce youth unemployment, the state should focus primarily on improving overall unemployment itself by implementing a dual-learning system, simplifying business opportunities, making part-time work available, or introducing lifelong learning to help transform the economy into a knowledge base.

Full text

Equilibrium. Quarterly Journal of Economics and Economic Policy
Volume 17 Issue 2 June 2022
p-ISSN 1689-765X, e-ISSN 2353-3293
www.economic-policy.pl
Copyright © Instytut Badań Gospodarczych / Institute of Economic Research (Poland)
This is an Open Access article distributed under the terms of the Creative Commons Attribution License
(http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and repro-
duction in any medium, provided the original work is properly cited.
ORIGINAL ARTICLE
Citation: Papík, M., Mihaľová, P., & Papíková, L. (2022). Determinants of youth unemployment
rate: case of Slovakia. Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2),
391–414. doi: 10.24136/eq.2022.013
Contact to corresponding author: Mário Papík, mario.papik@gmail.com
Article history: Received: 19.11.2021; Accepted: 7.05.2022; Published online: 25.06.2022
Mário Papík
Comenius University in Bratislava, Slovakia
orcid.org/0000-0003-2563-77991
Paulína Mihaľová
Comenius University in Bratislava, Slovakia
orcid.org/0000-0002-9954-7203
Lenka Papíková
Comenius University in Bratislava, Slovakia
orcid.org/0000-0001-6525-6769
Determinants of youth unemployment rate: case of Slovakia
JEL Classification: E24; J64
Keywords: youth unemployment; unemployment rate for graduated students; work motivation;
determinants of unemployment; quality of high schools
Abstract
Research background: The employment rate of young individuals in the labour market has
considerably decreased in developed countries recently. Due to lower labour capital, skills, and
generic and job-specific work experience, youth consider finding suitable job challenging. If they
fail to succeed in the labour market soon after graduation, it leads to long-term unemployment,
unstable and low-quality jobs, and even social exclusion.
Purpose of the article: This paper aims to analyse the unemployment rate of high school-
graduated students and the factors impacting this unemployment rate, such as GDP per capita,
total unemployment rate, apartment price per square meter and results from state exams. Identify-
ing the determinants affecting youth unemployment is crucial for theoretical knowledge and for
policymakers to ensure youth inclusion in the economic mainstream. As a result, society can
reduce social and economic costs and avoid structural problems in the future.
Methods: Data about 464 Slovak high schools from National Institute for Certified Educational.
Data include the graduate unemployment rate for each high school in Slovakia. Furthermore, two
logistic regression models have been developed to investigate the impact of selected factors on

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
392
high school graduates’ unemployment rate immediately after graduation and nine months after
graduation.
Findings & value added: This paper indicates the existence of statistical dependency between
unemployment of high school graduates and overall unemployment rate in the region, GDP per
capita in the region, quality of high school education and cost of living in the region immediately
after graduation. Analysis of the period nine months after graduation has shown the important
decline of education quality provided by high schools. To reduce youth unemployment, the state
should focus primarily on improving overall unemployment itself by implementing a dual-
learning system, simplifying business opportunities, making part-time work available, or intro-
ducing lifelong learning to help transform the economy into a knowledge base.
Introduction
The labour market has changed recently for many different reasons, with
technological progress being one of the most important. On the other hand,
many positive aspects of this factor are compensated by certain negative
aspects. Technological development has changed labour force requirements
while increasing demand for skilled and educated workers, described as
skill-biased (Goldin & Katz, 2007). Also reflecting this trend, labour mar-
ket prospects for youth have deteriorated significantly in many developed
(as well as developing) economies over recent decades. At the same time, it
is important to keep in mind the repetitive findings that aggregate demand
is the key determinant of youth employment and unemployment
(O’Higgins, 2017).
The youth unemployment rate is represented by number of unemployed
15-24-year-olds expressed as a percentage of the youth labour force (over-
all amount of both employed and unemployed youth) (OECD, 2020a). It is
result of complex relationships among demographic trends and specific
economic, cultural, and political contexts. Data show that more young peo-
ple continue in education to higher levels, and school-to-work transition
takes longer, resulting in significant number of young people not complet-
ing this transition until their late twenties (O’Higgins, 2017).
A more serious problem for society are so-called NEETs (not in educa-
tion, employment or training). NEET indicator measures the sum of all
youth not in employment, education, or training. It measures the number of
unemployed, discouraged, and inactive as a percentage of all youth. Young
people who are neither in employment nor in education or training are at
risk of becoming socially excluded — individuals with income below the
poverty line and lacking the necessary skills to improve their economic
situation (OECD, 2020b). For both genders, the proportion of NEETs in
population increases with age. The percentage of NEETs among young
men peaks at age of 23, whilst for young women this increasing trend con-

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
393
tinues into late twenties (O’Higgins, 2017). During the first three years
after graduation, the unemployment phase increases the likelihood of exit-
ing employment in the subsequent time periods. Youth unemployment is
also posing a negative effect on the later employment chances of young
individuals. Based on different studies, it seems that employers consider
non-employment at the beginning of their career as a negative signal of an
employee´s capabilities. Macroeconomic development and the phase of the
economic cycle (e.g. economic boom, or recession, etc.) crucially affects
the level of unemployment in the country. For youth who left education in
a period of unfavourable macro-economic conditions, non-employment in
early work-life is probably less disadvantageous compared to those who
entered the labour market in a period of a favourable economic climate
(Luijkx & Wolbers, 2009).
In 2020, the youth unemployment rate achieved 17.6% in the EU and
18.1% in the Eurozone, which is persistently more than double of the over-
all unemployment rate (7.2% within the EU-27 and 8.0% within the Euro-
zone area in February 2020). (European Commission, 2020) Slovakia is, in
the long-term, facing higher rates of youth unemployment (20.4% in Feb-
ruary 2020) (European Commission, 2020). Many authors have studied the
extent and consequences of ongoing youth unemployment. These authors
are mostly analysing the reasons for that state, as well as consequences.
Authors like Bal-Domańska (2021), Bayrak and Tatli (2018), Buttler
(2019), Dagume and Gyekye (2016), Demidova and Signorelli (2012),
Dvouletý et al. (2020), Kang (2021), Stabingis (2020) and Tomić (2018)
have tried to identify the determinants of youth unemployment mainly
among microeconomic and macroeconomic, demographic or structural
factors across countries — mainly across the European Union countries or
OECD countries — or within one country (e.g. Hungary, Russia, Romania,
and South Africa). These studies indicate that youth unemployment largely
depends on local GDP or other macroeconomic factors like inflation. On
the other hand, recent studies (e.g. Brada et al., 2014; Cvecic & Sokolic,
2018; Dimian, 2011; Danacica, 2014; Kabaklarli et al., 2011; Dagume &
Gyekye, 2016) have studied possible effects which education quality im-
poses on youth unemployment only to a limited extend and instead, these
studies have focused on the highest education achieved by the youth. Nev-
ertheless, the results of these studies indicate that the degree of highest
achieved education does impact the youth unemployment rate. This study,
therefore, analyses the quality of individual high schools (upper secondary
education) via youth unemployment rates per individual school based on
achieved state exam results and selected macroeconomic factors. Evalua-
tion of the high school quality through the students' results achieved in state

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
394
exams is a novelty of this manuscript because existing studies perceive the
quality of the educational system through the proportion of university-
educated people in the population. Therefore, this study has the potential to
fill the gap in the Central and Eastern European region (CEE), where such
research has not yet taken place.
This manuscript aims to analyse the unemployment rate of high school-
graduated students and factors impacting this unemployment rate. Dataset
of this manuscript includes data about 464 high schools for period starting
2016 to 2017. To fulfil this aim, two logistic regression models have been
developed. The first regression model analysed impact of selected variables
on youth unemployment immediately after graduation. The second regres-
sion model has analysed this impact on youth unemployment nine months
after graduation. As mentioned above, the novelty of this manuscript is its
focus on the quality of high school and its effect on youth unemployment.
Buttler (2019) has advised this point of view in his study as a possibility for
future research. The other novelty is the analysis of the relationship be-
tween selected individual factors and youth unemployment during various
time series (immediately after graduation and nine months after gradua-
tion). This study might be relevant due to its innovative approach, mainly
for CEE. The upper secondary level of education structure is similar across
many post-communist countries integrated into European structures (e.g.
Croatia, the Czech Republic, Hungary, Poland or Slovenia). For non-CEE
countries, this approach can also be innovative, provided other countries
monitor unemployment at the school level and their school system is final-
ized by standardized state exams applicable to the whole country.
The manuscript is structured into the following sections: Section 1 con-
tains a literature review of relevant prior studies in youth unemployment.
Section 2 describes the research methodology and data preparation process.
Section 3 contains description statistics and results of conducted analysis.
Section 4 includes a discussion, and the last section represents the conclud-
ing remarks.
Literature review
The status of youth in the labour market is, in general, linked to the existing
education system. Greater emphasis on specific skills of technological
changes and trends and a closer link between schools and employers has
led to easier transition from education to the labour market. As already
mentioned, the school-to-work transition process is moving to mid and late
twenties, as in general young people use to spend longer time in the educa-

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
395
tion process. Understandably, an easy transition from education (either
secondary or tertiary) to work is necessary to avoid youth unemployment.
Education may affect individual labour market vulnerability in several
ways. For example, the educational system confers attained skills (general,
as well as specific vocational), degree attained at school is also signal of
“trainability” of the person. In general, people with higher education are in
a more favourable position in the labour market (Korpi et al., 2003).
However, transition from education to work should be seen as a se-
quence of various procedures. This transition includes several phases in the
life of young individual, finding a satisfying job can take much time, and
this is the period of other important decisions such as leaving the house-
hold, forming a family etc. It is also evident that youth react more sensitive-
ly to socio-economic changes in labour markets. This is because the status
of young people in the labour market is less protected and more vulnerable
(Brzinsky-Fay, 2007). Once youth are unemployed, they are more exposed
to long-term unemployment, unstable and low-quality jobs, and social ex-
clusion. There is direct effect on the personal well-being and social devel-
opment of youth on individual level.
Countries use different approaches to find balance between education
and labour market needs. Developed countries use mainly vocational upper
secondary education to solve the mismatch between supply and demand of
the labour market (Estévez-Abe et al., 2001). It can be assumed that the
specific content of upper vocational education reflecting labour market
needs is necessary to reduce youth unemployment. Besides education, as
Tåhlin and Westerman (2020) point out, previous experience is a crucial
determinant of productive capacity and on-the-job learning rather than for-
mal schooling (Tåhlin & Westerman mention research of, e.g. Mincer,
1984). Many studies proving positive correlation between education re-
quirements and experience exist (Tåhlin & Westerman, 2020). Since young
individuals are relatively inexperienced, they are (given education) less
competitive in high skill jobs than in low-skill jobs. Availability of low-
skill jobs is crucial for youth to start their working careers.
A study by Šafránková and Šikýř (2017) has confirmed that part of
young individuals born between the early 1980s and mid-1990s in the
Czech Republic have problems finding a job on the labour market. This
inability to find a job is caused by high work expectations of these individ-
uals, like high income or rapid career growth combined with a lack of nec-
essary experience and work habits. Similar results have been achieved in
other studies from Poland as well. This Polish study has determined that
employers had problems with youth employment due to lack of experience,
initiative and entrepreneurship skills, and learning skills and courses

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
396
(Kobylińska et al., 2017). Similar results have been achieved in study of
Saxunová and Chorvatovičová (2018). Their study has determined that
secondary schools and universities could prepare graduated students for the
labour market requirements in a more effective way.
Brada et al. (2014) defined three major determinants of youth unem-
ployment such as cyclical conditions, social and structural conditions, and
government policies. Similarly, Bayrak and Tatli (2018) have shown that
GDP, inflation, and domestic gross savings have a negative effect on youth
unemployment whilst labour productivity affects youth unemployment
positively. The study by Tomić (2018) indicates that the most important
factors affecting the youth unemployment rate are real GDP growth along
with the size of the national construction sector. On the other hand, the
Tomić (2018) study also suggests that the amount of public debt in GDP is
the variable that is vastly associated with youth unemployment in Europe.
Studies by Kabaklarli et al. (2011), Arslan and Zaman (2014) and Bal-
Domańska (2021) have identified the relationship between youth unem-
ployment and GDP, inflation and labour market.
Danacica (2014), like Buttler (2019), has used the cases of Romania and
Hungary to show that individuals with higher education leave their jobs
more easily. However, they are also more often working in inappropriate
positions for their skill set and level of education. There are differences
among countries, but in developed economies it is not rare that young peo-
ple are “over-qualified” for the offered job positions (Ngai et al., 2016).
Contrary, Robayo and Estévez (2019) have shown that higher achieved
education might not lead to a higher probability of employment but rather
a higher probability of a better-quality job. Moreover, the duration of youth
unemployment also correlates with legal period for receiving unemploy-
ment allowances in Romania. Cvecic and Sokolic (2018), Danacica (2014),
and Dolado et al. (2013) have also pointed out the negative relationship
between the level of highest achieved education and youth unemployment.
On the other hand, Egessa et al. (2021) on a sample from Uganda and Car-
oleo et al. (2022) on a sample from Italy, Romania, and Bulgaria showed
that unemployment affects young people with lower levels of education.
Therefore, if the employment rate needs to be improved, the actual educa-
tion system needs to be improved. The impact of education on employment
has been reduced in Finland, where Pitkänen et al. (2021) showed that
youth unemployment currently depends more on the socio-economic back-
ground and adverse childhood experiences than on education. However,
these findings do not apply to the country´s senior population, which start-
ed their education within the previous educational system. According to

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
397
Kang (2021), countries with more work-oriented study programs have low-
er youth unemployment than other countries.
Dvouletý et al. (2020) and Marelli and Vakulenko (2016) showed
a higher unemployment rate for younger individuals, Dvouletý et al.
(2020), Lallukka et al. (2019), Lim and Lee (2019), Petrakis (2021) and
Schioppa and Lupi (2002) for females and Dvouletý et al. (2020) for people
from ethnic minority groups. Similar results were also achieved by Dagume
and Gyekye (2016), Maynou et al. (2022) and Wesseling (2021). They
have shown that among all characteristics of an individual (e.g. gender, age,
education), training and previous work experience have the biggest impact
on employment. The importance of training programmes and their effect on
employment has been confirmed by a Polish study by Styczyńska (2013).
Demidova and Signorelli (2012) have shown that youth unemployment is
higher than overall unemployment in Russia from 2000 to 2009. Across all
regions, it has been shown that youth unemployment can be improved by
migration, institutional and labour policies (both macroeconomic and mi-
croeconomic), permitting sustainable economic and social development of
young people. It is therefore important to focus on permanent monitoring of
these attributes across regions. Dimian (2011) has compared the impact of
selected demographic, institutional and business cycle factors on youth
unemployment. This study has shown that skilled youth unemployment,
unlike unskilled youth unemployment, depends on structural factors like
unemployment benefits or taxes on labour. It has also shown that youth
unemployment is negatively related to GDP per capita and positively relat-
ed to the unemployment rate from the previous period. Impacts of the un-
employment benefit system (negatively for passive labour market policies
and positively for the duration of unemployment benefits) were observed
by Ductor and Grechyna (2020).
A factor impacting unemployment rate is also financial security of indi-
viduals, e.g., in the form of savings, liquid assets and homeownership, and
lack of debts. Several studies have indicated that having savings allows
individuals to find a new job faster as they have the means to network and
travel for job interviews (Bayrak & Tatli, 2018; Tapsin, 2011). On the other
hand, having life savings also decreases the motivation of some individuals
to look for a new job (Kanfer et al., 2001; Solove et al., 2015). This finan-
cial security is created by parents’ financial support or living in parents’
houses for young individuals. According to a study by Aquilino (2005),
young individuals may also benefit from continuous parent support in diffi-
cult life situations. In addition, Ngai et al. (2016) have indicated that young
individuals with lower dependency on their parents have higher motivation
to look for a job and remain unemployed for shorter time period.

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
398
Youth unemployment is one of the key determinants of market income
inequalities, which leads to costs and structural problems for both states
and society and individuals and their families. Unemployment and related
loss of income, skills, motivations or dignity are closely linked to inequality
in all its different “spaces”. Many experts in different parts of the world
raise concerns about increasing income disparities and persistent unem-
ployment. Special attention is also given to youth unemployment, because
it could serve as a potential cause of political and social unrest (Huber &
Stephens, 2014; Ngai et al., 2016).
Income inequalities seem to have also occurred between technological
changes and education, a division of benefits of economic expansion is not
clear. Positive progress of economy based on technological advances is
indicated, but income increase is not equal within the labour market and the
economy in general. If workers have flexible skills and if educational infra-
structure expands sufficiently, then the supply of skills will increase due to
increased demand. Goldin and Katz, based on their findings, conclude that
when it comes to changes in the wage structure and returns to skill, supply
changes are critical, and education changes are by far the most important on
the supply side (Goldin & Katz, 2007).
The majority of the existing studies (Bayrak & Tatli, 2018; Bal-
Domanska, 2021; Buttler, 2019; Hasan & Sasan, 2020; Kang, 2021; Tomić,
2020) performed their analysis for a group of countries, e.g. EU countries
or OECD countries and a longer period. Therefore, panel data models (es-
pecially general linear models) were developed in the existing studies. This
method was also used by Hasan and Sasana (2020), who worked with the
Association of Southeast Asian Nations (ASEAN) countries, and Demidova
and Signorelli (2012), who analysed several regions within one country —
Russia. Ductor and Grechyna (2020) used the Bayesian model averaging
approach as an alternative to the panel data models. However, studies using
data for a single period and usually from one country, such as Dagume and
Gyekye (2016), Dimian (2011), Egessa et al. (2021) and Luijkx and Wol-
bers (2009), use logistic regression in their analysis.
Research method
Students start upper secondary education in Slovakia at the age of 15, simi-
larly to many CEE countries (Czech Republic — 15, Slovenia — 15, Hun-
gary — 14, Poland — 16, Croatia — 14). Secondary education lasts from 4
to 5 years and ends with a state exam. Secondary schools can be organiza-
tionally divided into general, vocational and art schools. Data of 464 Slo-

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
399
vak high schools (out of 800 Slovak high schools) from the National Insti-
tute for Certified Educational Measurements place during the 2015/2016
academic year have been collected to fulfil the aim of this manuscript. The
2015/2016 academic year was the last academic year prior to announcing
a significant change in the educational system in Slovakia and was, there-
fore, the last academic year unaffected by the preparations for this reform.
Data contain graduate unemployment rate for each of the high schools for
A) September 2016 and B) period nine months afterwards (May 2017).
High schools and other institutions use this data to determine their gradu-
ates' chances to succeed in the labour market.
Moreover, dataset from National Institute for Certified Educational
Measurements has been extended by results from the Slovak Language
State Exams, which represent the quality of education factor provided by
a particular high school. Analysis of high school quality through the stu-
dents' results from the state exams is a new approach in this study. Existing
studies (Brada et al., 2014; Cvecic & Sokolic, 2018; Dagume & Gyekye,
2016; Danacica, 2014; Dimian, 2011; Kabaklarli et al., 2011) mostly per-
ceive the quality of the educational system through the proportion of uni-
versity-educated people in the population. State exams from many courses
took place during the 2015/2016 academic year. However, only Slovak
Language State Exam was mandatory for all high schools in Slovakia, mak-
ing it one of the independent variables of this manuscript. Furthermore,
Slovak Language State Exam is not compulsory only at those high schools
with tuition language other than the Slovak language. The mentioned data
are publicly available on the webpage of the National Institute for Certified
Educational Measurements.
Other independent variables identified in this manuscript are unem-
ployment rate and gross domestic product (GDP) per capita, both on re-
gional level and cost of living, determined by apartment price per square
meter on regional level. These variables have been chosen as they provide
ideas about economic growth and inflation in each of the Slovak regions
(Buttler, 2019; Dagume & Gyekye, 2016; Demidova & Signorelli, 2012).
When graduates do not have enough job opportunities in their home regions
due to slow or no economic growth, neither graduation from high-quality
high school nor above-average graduation exam results are of any use when
looking for a job in these regions. The same situation occurs with the
apartment price per square meter when students who reside in regions with
higher living costs tend to be less motivated to find a job as they cannot be
entirely independent on their parents. Even being employed due to the high
cost of living, they have to share accommodation with their relatives mak-
ing it non-motivational to have a proper job. Especially in the eastern Slo-

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
400
vak regions, individuals prefer having their apartment instead of renting it.
It leads to a vicious circle, with young individuals not having their places
because of the high cost of living. Thus, they live with their parents, mean-
ing they have lower monthly expenses then they would have if living inde-
pendently. The result is that these young individuals are not under pressure
to look for a job, and when they do not have money, they cannot afford
their apartment.
As a basic economic theory describing the relationship between these
transformations, the Philips curve can be used as it expresses the relation-
ship between inflation (based on economic growth) and unemployment.
According to the Phillips curve, a decline in unemployment caused by eco-
nomic growth increases prices, leading to increased inflation and thus in-
creased living costs.
Therefore, the test unit is a high school, and the predicted variable is the
unemployment of graduates of a particular secondary school. The macroe-
conomic variables of a given region or the state test results at a given
school are then selected as explanatory variables. The models are simplified
because it is impossible to record macroeconomic indicators at a lower
level than the regional level (e.g., city level). Based on one macroeconomic
value, they try to predict multiple unemployment opportunities for second-
ary schools belonging to the region. This approach can be considered
a limitation of the applied methodology and the developed regression mod-
els. All these data are publicly available on the webpage of Slovak Statisti-
cal Office.
Studied logistic model (1) describing relationship between unemploy-
ment rate of graduated students and independent variables described above
has the following vector form






(1)
where 

represents unemployment rate of graduated high school stu-
dents in September 2016 for high schools included in data sample, 

is
unemployment rate in the region,  !" is gross domestic product per
capita in the region, !! represents apartment price per square meter in the
region and SE stands for results of the Slovak Language State Exams.
#

$ #
%
$ #
&
'()#
*
'+,estimated coefficients by the OLS method and -
corresponds to residuals. Coefficient #
.
is intercept of logistic regression
model.
Whilst model (1) analyses relationship between unemployment rate of
high school graduates and quality of economic or education factors imme-

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
401
diately after graduation, and another model has been developed to describe
relationship between these factors nine months after graduation, in May of
the following year. Model (2) has used results of model (1) as another in-
dependent variable. It is expected that high schools with higher graduate
unemployment rate in September 2016 would also report higher graduate
unemployment rate in May 2017. This would help to explain the theory of
difficulties in employment of young individuals in regions.
During the analysis, second
logistic
model (2) has developed with the
following form

/01



23
(2)
where all variables are identical to variables from Model (1)
and
/0145
represents unemployment rate for graduated high school
students in May of the following year for analysed high schools and #
6
rep-
resents estimated coefficient.
Parameters of both models (1, 2) have been estimated by statistical
software R-studio, namely its build-in function for generalized linear mod-
els (glm()) (Chambers, 1992). The selection of final models was conducted
through a stepwise regression function using R squared, Akaike Infor-
mation Criterion and Bayesian Information. Multicollinearity has been
detected by variance inflation factor (VIF). Linearity between predictors
and response variable has been tested by Box-Tidwell test. Outliers have
been analysed by Cook’s distance plot (Draper & Smith, 2014; Faraway,
2002).
Results
Table 1 shows number of high schools in each of the individual regions.
Data from 464 high schools in Slovakia were collected in overall. Most of
the schools in the data sample are located in the Bratislava, Trnava or Nitra
regions. These three regions cover high acreage, have high number of in-
habitants, and therefore proportion of the high schools is higher there.
Table 1 also indicates the average unemployment rate of high school
graduates measured immediately after graduation (in September) and after
nine months afterwards (May of the following year). The region with the
highest graduate unemployment rate is the Košice region located in the
easter part of Slovakia. This region is characterised by generally low num-
ber of vacant job opportunities. The high school graduate unemployment

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
402
rate in Košice region is well over 12% immediately after graduation, which
is more than 2 p.p. higher than the average. From September to May, how-
ever, the highest decline in the unemployment rate was observed in this
region - the unemployment rate decreased from 14.24% to 6.28%.
The average result achieved by Slovak students from the Slovak Lan-
guage State Exams in sample is 53.55%. The worst result was achieved
mainly in regions with a high presence of the Hungarian minority in the
south of Slovakia, such as the Banská Bystrica, Nitra and Trnava regions.
Worse results from Slovak Language State Exams in these regions do not
indicate lower quality of these high schools, but rather problems with inte-
gration of Hungarian minority into Slovak majority. On the other hand,
regions located in the north and west of Slovakia, like the Žilina, Trenčín,
Prešov and Bratislava region, have achieved above-average results of over
53.55%.
Table 2 includes the following macroeconomic information related to
each of the individual regions: unemployment rate, GDP per capita and
apartment price per square meter in September 2016 and May 2017. The
unemployment rate during the analysed period has decreased by almost two
percent, with this decline in the unemployment rate taking place in all Slo-
vak regions. In the case of GDP per capita, growth has been reported in the
Bratislava and Žilina regions. However, other regions have reported a de-
cline in GDP per capita.
The price of real estate did not change significantly during the analysed
period. As was the case with GDP per capita, some regions have reported
a decline in apartment prices or flat period-over-period prices (e.g. Brati-
slava, Košice, Prešov, and Žilina regions). Cost of living slightly increased
in the remaining regions. The overall cost of living remained constant, as
was the case for GDP per capita during same analysed period.
Analysis of model (1) has identified the following independent variables
as statistically significant at level 0.1%: GDP per capita in the region and
Slovak Language State Exam results of particular high school and at the
level 1%: unemployment rate in the region, as shown in Table 3. Model (1)
confirmed that should the unemployment rate in the region be rather high,
then also rate of high school graduates is rather high as well. The employ-
ment rate of high school graduates is thus dependent on the labour market
of the region. Moreover, Model (1) showed a negative relationship between
GDP and youth unemployment rate. Young people living in regions with
higher economic growth have a higher chance to employ after graduation.
The last statistically significant variable, which was identified by model
(1), is the result from Slovak Language State Exam of particular high
school. However, the Slovak Language State Exam results indicated rather

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
403
insufficient for high schools with Hungarian tuition language. Despite this
fact, analysis of model (1) indicated that quality education has an impact on
high school graduates. the better the study results of high school graduates,
the higher the probability of finding a job immediately after graduation.
Pseudo R-squared is 79.6%, so the model fits well to the research sam-
ple. Akaike information criterion is -1 833.59, and the Bayesian infor-
mation criterion is -1 806.93. Variance inflation factors are under 2.1,
which indicate weak correlations among predictors. Moreover, there is also
linearity, tested by Box-Tidwell test, between predictors and response vari-
able was. Cook’s distance plot shows that there are no outliers.
Model (2) of this manuscript studies variables affecting high school
graduate unemployment rate nine months after graduation and results are
shown in Table 4. Analysis of model (2) has identified the following inde-
pendent variables as statistically significant at level 0.1%: the unemploy-
ment rate in the region and the unemployment rate of high school graduates
from September 2016 and at the level 5%: GDP per capita. Thus, the de-
pendency between the unemployment rate in the region and the unemploy-
ment rate of high school graduates is decreasing in nine months. However,
it is still correct to say that regions with higher unemployment rates and
lower GDP per capita also report higher high school graduate unemploy-
ment rates. Most regions' results indicate that the high school graduate un-
employment rate is decreasing below the regional unemployment rate after
nine months. This would indicate that young individuals are able, when
given a few more months of searching on the labour market, to find job
easier than a majority of unemployed population. On the other hand, results
also show that should a high school report higher graduate unemployment
rate in September, it would still report higher graduate unemployment rate
after nine months.
Model (1) indicates that if a graduate studied at quality high school, they
would succeed on the labour market sooner. However, this positive impact
of education quality diminishes at the course of time as is shown in model
(2). The quality of education is considered important benefit for high
school graduates which allows them to find a job in shorter time period. As
time goes, other factors like the macroeconomic situation (measured by the
unemployment rate or GDP per capita) in the region become more im-
portant than the quality of education.
Model (2) has pseudo R-squared equal to 83.2% and also, this model fits
well to observed data. Akaike information criterion is -3 178.97, and
Bayesian information criterion is -3 147.45. Variance inflation factors are
under 1.4, which indicate even weaker correlations among predictors than
model (1). Linearity between predictors and response variable was con-

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
404
firmed by Box-Tidwell test. Moreover, Cook’s distance plot does not show
any extreme values.
Discussion
This manuscript has indicated that high school graduates´ unemployment
rate is affected by several factors. The unemployment rate of high school
graduates immediately after graduation depends on the macroeconomic
situation (overall unemployment rate and GDP per capita in the region) and
quality of education. Studies identified the positive impact of overall un-
employment by Bal-Domańska (2021) and Dimian (2011), whilst the nega-
tive impact of GDP was identified by studies of Arslan and Zaman (2014),
Bal-Domańska (2021), Bayrak and Tatli (2018), Buttler (2019) and Dimian
(2011) and Tomić (2018). Although according to these studies, it is enough
to boost the economic growth to reduce youth unemployment, the reality of
many European economies shows that during the economic growth also,
other various macroeconomic and non-macroeconomic indicators affect
unemployment. The findings from this study are also in line with Tomić
(2018), who showed that the results from the application of individual-level
data could be beneficial not only for the country which was analysed. Un-
like existing studies (Brada et al., 2014; Cvecic & Sokolic, 2018; Dimian,
2011; Danacica, 2014; Kabaklarli et al., 2011; Dagume & Gyekye, 2016),
this manuscript has measured quality of education not by the highest
achieved level, but rather by the results achieved in the state exams. The
quality of education measured in this way directly impacted the lower un-
employment rate immediately after graduation. The positive impact of edu-
cation in the short term was also observed by Dvouletý et al. (2020), May-
nou et al. (2022) and Tomić (2018). They indicate that education of good
quality and training programs contribute to school-to-work transition. De-
spite the minor difference in used methods, the results of this study from
high school-level perspective are similar and comparable to studies from
other CEE countries like Romania or Hungary, but also with Russia
(Danacica, 2014; Demidova & Signorelli, 2012) and Western Europe (Bal-
Domańska, 2021; Buttler, 2019; Dvouletý et al., 2020) in short-term.
Analysis of data nine months later after graduation has shown that the
statistical significance of education quality decreased with time. Contrary
to the findings of this manuscript are the findings by Lallukka et al. (2019).
They observed lower unemployment among young people with better so-
cial determinants such as education in the long run. The difference among
the findings may be caused by the time interval when both studies exam-

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
405
ined employment rate — nine months after graduation in this study and 12
months after graduation in the Lallukka et al. study (2019). Since there is
common practice in Slovakia to provide limited-term contracts (usually one
year) to people entering the labour market. It is possible that after 12
months, some graduated students who were initially employed did not have
their contracts extended and became unemployed — a possibility this re-
search did not capture. Regardless, the impact change of in the educational
process on employment over time could be classified as a major finding
resulting from this manuscript.
The results of this paper confirm expected results that the youth unem-
ployment rate depends on the quality of provided education. The employ-
ment rate of graduates from higher-quality high schools is significantly
higher than the rate of graduates from lower-quality high schools immedi-
ately after graduation. Education support for students from regions with
higher unemployment is hence crucial. However, this support would de-
crease youth unemployment mainly in the short run. If we aim to decrease
youth unemployment in the medium run, we shall decrease overall unem-
ployment. The relationship between youth unemployment and overall un-
employment was significant also after nine months after graduation.
The collected data might be considered as a limitation of this study.
A more extended period analysed would provide higher-quality results
smoothing abnormalities of the analysed year. Using several years in the
analysis would allow the application of a panel data model within one
country, as some existing studies have done (Bayrak & Tatli, 2018; Bal-
Domanska, 2021; Buttler, 2019; Hasan & Sasan, 2020; Kang, 2021). More-
over, the panel data model could also contain a region or a type of high
school as factors, which might help distinguish specific behaviour across
these units. Another limitation of this paper could be the lack of infor-
mation about the job vacancies available to graduates. This limitation could
be resolved by considering job offers pools in the prediction model. Job
vacancies are an important factor, as they show better young people's
chances of finding employment in a given region than the economy's per-
formance measured by GDP. Based on the methods used, a limitation can
be observed in the granularity of predictors, which are not at a lower level
than regions. Insufficient detailed data can then skew the prediction to es-
timate the unemployment of graduates of individual schools through high-
level information from a given region.

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
406
Conclusions
Even though this study was conducted only on data from Slovak high
schools, its results are applicable also for other countries since these results
are in line with those from other studies from CEE. Results of this manu-
script show that should state want to achieve lower youth unemployment, it
should concentrate on improvement of education quality and overall mac-
roeconomic situation in the region. The quality of education has the highest
impact on youth unemployment immediately after graduation, whilst the
macroeconomic situation in the region measured by the overall unemploy-
ment rate impacts youth unemployment rate in a few months after gradua-
tion.
There are several ways in which governments and policymakers could
reduce youth unemployment. Experience from Western Europe (e.g., Fin-
land) shows that the improvement of the education system contributes to
a reduction in youth unemployment. Such improvements to the education
system include various training programs or the implementation of a dual
education system, which aims to combine theoretical knowledge from
schools with the practical experience from the work environment — anoth-
er aspect where policymakers can reduce the barriers arising during this
process. Experience from abroad also shows that the possibility of part-time
contracts leads to a lower youth unemployment rate. In the long run, how-
ever, policymakers should focus on making lifelong learning more accessi-
ble and thus contribute to the transformation of their countries to
knowledge-based economies.
From the national perspective, the main contribution of this paper is that
it has analysed the vast majority of Slovak high schools. Furthermore, this
study did not focus on the highest achieved education, but rather on the
quality of education expressed as achieved results in the state exams. Fu-
ture studies could concentrate on analysis of university student’s unem-
ployment as majority of youth has started studying at universities. Hence
the challenges of unemployment are shifted from high school graduates to
university graduates. Possible future factors that could be analysed are in-
dividual personal attributes of youth, environment where youth grew up,
and personal motivation to look for a job. Even though system parameters
are not set in favour of young job applicants, several examples from other
countries show that youth can find a job despite unfavourable system con-
ditions.
From the international perspective, the main contribution of this paper is
the methodological perspective. The novel approach of this paper relates to
the measurement of the upper secondary school quality through the

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
407
achieved results of students from state exams, in contrast to the existing
studies, which usually measure the proportion of university-educated peo-
ple in a population. Therefore, it contributes to a more comprehensive and
homogenous comparison. Future research abroad should try to perform
similar analyses if similar data is available within a given country. In the
context of CEE, it is possible to extend this comparison to an international
perspective and comparison as some countries already collect similar data
(e.g., a comparison of Slovakia and the Czech Republic).
References
Arslan M., & Zaman R. (2014), Unemployment and its determinants: a study of
Pakistan economy (1999-2010). Journal of Economics and Sustainable
Development, 5(13), 20–24. doi: 10.2139/SSRN.2515075.
Aquilino, W. S. (2005). Impact of family structure on parental attitudes toward the
economic support of adult children over the transition to adulthood. Journal of
Family Issues, 26(2), 143–167. doi: 10.1177/0192513X04265950.
Bal-Domańska, B. (2021). The impact of macroeconomic and structural factors on
the unemployment of young women and men. Economic Change and
Restructuring. Advance online publication. doi: 10.1007/s10644-021-09341-9.
Bayrak, R., & Tatli, H. (2018). The determinants of youth unemployment: a panel
data analysis of OECD Countries. European Journal of Comparative
Economics, 15(2), 231–248. doi: 10.25428/1824-2979/201802-231-248.
Brada, J. C., Marelli, E., & Signorelli, M. (2014). Young people and the labor
market: key determinants and new evidence. Comparative Economic Studies,
56(4). 556–566. doi: 10.1057/ces.2014.30.
Brzinsky-Fay, C. (2007). Lost in transition? Labour market entry sequences of
school leavers in Europe. European Sociological Review, 23, 409–422. doi:
10.1093/esr/jcm011.
Buttler, D. (2019). Employment status and well-being amongst youth: explaining.
In B. Hvinden, J. O’Reilly, M. A. Schoyen & C. Hyggen (Eds.). Negotiating
early job insecurity. Cheltenham: Edward Elgar Publishing, 19–45. doi: 10.433
7/9781788118798.00008.
Caroleo, F. E., Rocca, A., Neagu, G., & Keranova, D., (2022). NEETs and the
process of transition from school to the labor market: a comparative analysis of
Italy, Romania, and Bulgaria. Youth & Society, 54, 109–129. doi:
10.1177/00441 18X211056360.
Cvecic, I., & Sokolic, D. (2018). Impact of public expenditure in labour market
policies and other selected factors on youth unemployment. Economic
Research-Ekonomska Istraživanja, 31(1), 2060–2080. doi: 10.1080/1331677X
.2018.1480970.

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
408
Danacica, D.-E. (2014). Determinants of youth unemployment spells and exit des-
tinations in Romania and Hungary. Acta Oeconomica, 64(3), 335–365. doi:
10.1556/AOecon.64.2014.3.4.
Dagume, M. A., & Gyekye A. (2016). Determinants of youth unemployment in
South Africa: evidence from the Vhembe district of Limpopo province.
Environmental Economics, 7(4). doi: 10.21511/ee.07(4).2016.06.
Demidova, O., & Signorelli, M. (2012). Determinants of youth unemployment in
Russian regions. Post Communist Economies, 24(2), 191–217. doi: 10.1080/14
631377.2012.675155.
Dimian, G. C. (2011). The role of education in explaining youth labour market
imbalances in CEE countries. Journal on Efficiency and Responsibility in
Education and Science, 4(3), 105–115.
Dolado J. J., Jansen, M., Felgueroso, F., Fuentes, A., & Wölf, A. (2013). Youth
labour market performance in Spain and its determinants. OECD Economics
Department Working Papers. doi: 10.1787/18151973.
Draper, N. R., & Smith, H. (2014). Applied regression analysis. Wiley series in
probability and statistics. John Wiley & Sons. doi: 10.1002/9781118625590.
Ductor, L., & Grechyna, D., (2020). Determinants of ins and outs of unemploy-
ment. Applied Economics Letters, 27, 892–898. doi: 10.1080/13504851.201
9.1646858.
Dvouletý, O., Lukeš, M., & Vancea, M. (2020). Individual-level and family back-
ground determinants of young adults’ unemployment in Europe. Empirica, 47,
389–409. doi: 10.1007/s10663-018-9430-x.
Egessa, A., Nnyanzi, J. B., & Muwanga, J. (2021). Determinants of youth unem-
ployment in Uganda: the role of gender, education, residence, and age. IZA
Journal of Labor Policy, 11, 20210008. doi: 10.2478/izajolp-2021-0008.
European commission (2020). Unemployment statistics. Retrieved from
https://ec.europa.eu/eurostat/statistics-explained/index.php/Unemployment_stat
istics#Unemployment_in_the_Member_States.
Estévez-Abe, M., Iversen, T., & Soskice, D. (2001). Social protection and the for-
mation of skills: a reinterpretation of the welfare state. In P. A. Hall & D.
Soskice (Eds.). Varieties of capitalism: the institutional foundations of compar-
ative advantage. Oxford/New York: Oxford University Press, 145–183. doi:
10.1093/0199247757.003.0004.
Faraway, J. J. (2002). Practical regression and Anova using R. Retrieved from
https://cran.r-project.org/doc/contrib/Faraway-PRA.pdf.
Goldin, C., & Katz, L. F. (2007). The race between education and technology: the
evolution of U.S. educational wage differentials, 1890 to 2005. NBER Working
Paper, 12984. doi: 10.3386/w12984.
Hasan, Z., & Sasana, H., (2020). Determinants of youth unemployment rate in
Asean. International Journal of Scientific & Technology Research, 9(6), 6687–
6691.
Huber, E., & Stephens, J. D. (2014). Income inequality and redistribution in post-
industrial democracies: demographic, economic and political determinants.
Socio.Economic Review, 12, 245–267. doi: 10.1093/ser/mwu001.

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
409
Chambers, J. M. (1992). Linear models. In J. M. Chambers, T. J. Hastie & T. Has-
tie (Eds.). Statistical models in S. Wadsworth & Brooks/Cole.
Kabaklarli, E., Hazeler, P., & Buhus, A. (2011). Economic determinants of Turkish
youth unemployment problem: cointegration analysis. In Proceedings of the
international conference on applied economics. Scientific Research. An Aca-
demic Publisher, 267–272.
Kanfer, R., Wanberg, C. R., & Kantrowitz, T. M. (2001). Job search and employ-
ment: a personality-motivational analysis and metaanalytic review. Journal of
Applied Psychology, 86. 837–855. doi: 10.1037/0021-9010.86.5.837.
Kang, Y. (2021). Determinants of youth unemployment: empirical analysis of
OECD and EU member countries in 2000-2017. Journal of Economic Devel-
opment, 46(3), 109–131. doi: 10.35866/caujed.2021.44.3.006.
Kobylińska, U., Rollnik-Sadowska, E., & Samul, J. (2017). Young people on the
labour market in Poland — the point of view of the employer. Oeconomia
Copernicana, 8(4), 553–568. doi: 10.24136/oc.v8i4.34.
Korpi, T., de Graaf, P., Hendrickx, J., & Layte, R. (2003). Vocational training and
career employment precariousness in Great Britain, the Netherlands and Swe-
den. Acta Sociologica, 46, 17–30. doi: 10.1177/0001699303046001002.
Lallukka, T., Kerkelä, M., Ristikari, T., Merikukka, M., Hiilamo, H., Virtanen, M.,
Øverland, S., Gissler, M. & Halonen, J. I., (2019). Determinants of long-term
unemployment in early adulthood: a Finnish birth cohort study. SSM
Population Health, 8, 100410. doi: 10.1016/j.ssmph.2019.100410.
Lim, J.-Y., & Lee, Y.-M., (2019). Exit duration and unemployment determinants
for Korean graduates. Journal for Labour Market Research, 53(5), 2–14. doi:
10.1186/s12651-019-0255-2.
Luijkx, R., & Wolbers, M. H. J. (2009). The effects of non-employment in early
work-life on subsequent employment chances of individuals in the Netherlands.
European Sociological Review, 25(6), 647–660. doi: 10.1093/esr/jcp002.
Marelli, E., & Vakulenko, E. (2016). Youth unemployment in Italy and Russia:
aggregate trends and individual determinants. Economic and Labour Relations
Review, 27(3), 387–405. doi: 10.1177/1035304616657959.
Maynou, L., Ordóñez, J., & Silva, J. I. (2022). Convergence and determinants of
young people not in employment, education or training: an European regional
analysis. Economic Modelling, 110, 105808. doi: 10.1016/j.econmod.2022.10
5808.
Mincer, J. (1984). Human capital and economic growth. Economics of Education
Review, 3(3), 195–205. doi: 10.1016/0272-7757(84)90032-3.
Ngai, S. S. Y., Cheung, C. K., Yuan, R., & Lin, S. (2016). Work motivation of
unemployed youths: moderating effects of financial dependence on parents.
Children and Youth Services Review, 71, 157–165. doi: 10.1016/j.childyouth.2
016.11.005.
O’Higgins, N. (2017). Rising to the youth employment challenge - new evidence on
key policy issues. International Labour Office – Geneva: ILO.
OECD (2020a). Youth unemployment rate. Retrieved from https://data.oecd.org/u
nemp/youth-unemployment-rate.htm.

Equilibrium. Quarterly Journal of Economics and Economic Policy, 17(2), 391–414
410
OECD (2020b). Youth not in employment, education or training (NEET). Re-
trieved from https://data.oecd.org/youthinac/youth-not-in-employment-educat
ion-or-training-neet.htm.
Petrakis, I., (2021). Determinants of female labour force participation: evidence
from Greece. Labour, 35, 538–567. doi: 10.1111/labr.12206.
Pitkänen , J., Remes, H., Moustgaard, H., & Martikainen, P. (2021). Parental soci-
oeconomic resources and adverse childhood experiences as predictors of not in
education, employment, or training: a Finnish register-based longitudinal study,
Journal of Youth Studies, 24(1), 1–18. doi: 10.1080/13676261.2019.1679745.
Robayo, C. D. C., & Estévez, J. G., (2019). Desempleo juvenil en Colombia ¿La
educación importa? Revista Finanzas Y Política Económica, 11(1), 101–127.
doi: 10.14718/revfinanzpolitecon.2019.11.1.7.
Šafránková, J. M., & Šikýř, M. (2017). Work expectations and potential employa-
bility of millennials and post-millennials on the Czech labor market.
Oeconomia Copernicana, 8(4). 585–599. doi: 10.24136/oc.v8i4.36.
Saxunová, D., & Chorvatovičová, L. (2018). Management of labour force move-
ment applied in Slovakia. Social and Economic Revue, 16(2), 35–43.
Schioppa, F. K. P., & Lupi, C. (2002). Family income and wealth, youth unem-
ployment and active labour market policies. International Review of Applied
Economics, 16(4), 407–416. doi: 10.1080/02692170210161147.
Solove, E., Fisher, G. G., & Kraiger, K. (2015). Coping with job loss and
reemployment: a two-wave study. Journal of Business and Psychology, 30(3),
529–541. doi: 10.1007/s10869-014-9380-7.
Stabingis, L. (2020). Impact of youth guarantee iniciatives on NEETs population
changes in EU. Management Theory and Studies for Rural Business and
Infrastructure Development, 42(2), 145–156. doi: 10.15544/mts.2020.15.
Styczyńska, I. (2013). Enhancing youth opportunities in employment: determinants
and policy implications. Intereconomics, 48(4), 196–235. doi: 10.1007/s10272-
013-0464-9.
Tåhlin, M., & Westerman, J. (2020). Youth employment decline and the structural
change of skill. European Societies 2020, 22(1), 47–76. doi: 10.1080/146166
96.2018.1552981.
Tapsın G. (2011). The relation between national savings, economic growth and
employment. İstanbul: Istanbul University Social Sciences Institute.
Tomić, I. (2018). What drives youth unemployment in Europe? Economic vs non-
economic determinants. International Labour Review, 157(3), 379–408. doi:
10.1111/ilr.12113.
Wesseling W. I. E. (2021). Towards a more all-inclusive evaluation of interven-
tions for unemployed youth: a longitudinal investigation of participant, pro-
gramme, relational and contextual factors. Social Policy & Administration,
55(7), 1–17. doi: 10.1111/spol.12705.

Annex
Table 1. Description table for high school attributes for regions
Region Number of
high schools
Unemployment rate
Sep.
-
graduated
Unemployment rate
May.
-
graduated
Results
state exams
BA
122
8.23%
3.25%
54.85%
BB
41
12.14%
5.92%
51.98%
KE
33
14.24%
6.28%
52.93%
NR
60
10.79%
4.39%
50.59%
PO
48
10.87%
5.43%
56.58%
TN
48
9.70%
3.65%
54.01%
TT
64
10.38%
3.97%
52.24%
ZA
48
10.95%
5.43%
53.92%
Total
464
10.34%
4.44%
53.55%
Note: (Meaning: BA — Bratislava region, BB — Banská Bystrica region, KE — Košice
region, NR — Nitra region, PO — Prešov region, TN — Trenčín region, TT — Trnava
region and ZA — Žilina region).
Source: own calculation based on data from Statistical Office of the Slovak Republic and
National Institute for Certified Educational Measurements.
Table 2. Description table for regional data
Region Unemployment
rate Sep.
Unemployment
rate May
GDP
Sep.
GDP
May
Property
price Sep.
Property
price May
BA 5.00% 3.68% 235.25 240.29 1.38 1.37
BB 13.70% 11.31% 73.77 72.99 0.52 0.55
KE 11.15% 9.86% 83.16 78.44 0.71 0.71
NR 7.66% 5.14% 87.46 86.38 0.46 0.48
PO 14.08% 11.66% 59.85 60.64 0.60 0.60
TN 6.26% 4.06% 87.46 86.38 0.50 0.53
TT 5.12% 3.15% 108.62 109.04 0.66 0.67
ZA 7.37% 5.71% 85.34 86.24 0.63 0.62
Total 9.26% 7.32% 103.50 103.56 0.71 0.71
Note: (Meaning: BA — Bratislava region, BB — Banská Bystrica region, KE — Košice
region, NR — Nitra region, PO — Prešov region, TN — Trenčín region, TT — Trnava
region and ZA — Žilina region).
Source: own calculation based on data from Statistical Office of the Slovak Republic and
National Institute for Certified Educational Measurements.

Table 3. Results of logistic model for
unemployment rate — high
school
graduates in September 2016
Variables Estimate Std. Error t value Pr(>|t|) Sign.
(Intercept) 0.3946 0.04004 9.857 < 2e-16 ***
Unemployment rate 0.2412 0.09098 2.652 0.0082 **
GDP per capita -0.0012 0.00034 -3.452 0.0005 ***
Apartment price for square meter
0.0182 0.02794 0.652 0.5148
Results from state exams -0.0042 0.00021 -19.734 < 2e-16 ***
Box-Tidwell test MLE of lambda t value Pr(>|t|) Sign. vif
Unemployment rate
0.4786
-
0.5683
0.5698
2.0157
GDP per capita
-
63.92397
-
1.9517
0.0510
2.0218
Apartment price for square meter
8.57818
-
1.9077
0.0564
1.0191
Results from state exams
0.9060
0.4012
0.6883
1.0070
Pseudo R-squared 0.796
Residual standard error 1.946
Degrees of freedom residuals 460
Akaike Inf. Criterion -1 833.59
Bayesian Inf. Criterion -1 806.93
Note: Significance: * corresponds to statistical significance at 5%, ** corresponds to
statistical significance at 1% and *** corresponds to statistical significance at 0.1%
Source: own calculation in R-studio based on data from Statistical Office of the Slovak
Republic and National Institute for Certified Educational Measurements.
Table 4 Results of logistic model for unemployment rate
—
high school
graduates in May 2017
Variables Estimate Std.
Error
t value Pr(>|t|) Sign.
(Intercept)
0.0162
0.01366
1.182
0.2375
Unemployment rate
0.1504
0.03768
3.992
0.0001
***
Unemployment rate – high
school graduates in September
0.2712
0.01257
21.571
< 2e
-
16
***
GDP per capita
-
0.0002
0.00011
-
2.265
0.0238
*
Apartment price for square
meter
0.0029
0.01237
0.233
0.8156
Results from state exams
0.0002
0.00010
1.696
0.0904
Box-Tidwell test MLE of
lambda
t value Pr(>|t|) Sign. vif
Unemployment rate -1.15476 -1.8532 0.0638 1.3767
Unemployment rate – high
school
graduates in September
0.9348 -0.3965 0.6917 1.3903

Table 4. Continued
Box-Tidwell test MLE of
lambda
t value Pr(>|t|) Sign. vif
GDP per capita 32.3105 -1.3787 0.1682 1.3814
Apartment price for square
meter
277.9150 -0.6584 0.5103 1.0107
Results from state exams 2.3785 -1.4023 0.1608 1.3527
R-squared 0.832
Residual standard error 0.326
Degrees of freedom residuals 459
Akaike Inf. Criterion -3 178.97
Bayesian Inf. Criterion -3 147.45
Note: Significance: * corresponds to statistical significance at 5%, ** corresponds to statistical
significance at 1% and *** corresponds to statistical significance at 0.1%
Source: own calculation in R-studio based on data from Statistical Office of the Slovak Republic and
National Institute for Certified Educational Measurements.
Figure 1. Cook’s distance for model (1)
Source: own calculation in R-studio based on data from Statistical Office of the Slovak Republic and
National Institute for Certified Educational Measurements

Figure 2. Cook’s distance for model (2)
Source: own calculation in R-studio based on data from Statistical Office of the Slovak Republic and
National Institute for Certified Educational Measurements.