UNDERSTANDING THE NEET IN TURKEY

Abstract

The purpose of this paper is twofold. First, drawing on data from the Household Labor Force Surveys over the period 2004-2013, it examines the determinants of the NEET (Not in Employment, Education or Training) status for the Turkish youth. This is particularly important for Turkey as it has the highest NEET rate among the OECD countries. Second, it describes the
movement of the youth across four states: education, employment, unemployment and inactivity. Probit results indicate that gender and educational attainment are key factors for
explaining the NEET status. Findings also show that a greater number of household members that are in employment is associated with a lower likelihood of NEET. Transition analyses reveal that the state of inactivity remains highly persistent despite the substantial fall over the sample period. In addition, the rise in the persistence of education between 2007 and 2009 underlines
the choice of the youth to stay in education in response to the fall in labor market prospects.

Full text

Eurasian Journal of Economics and Finance, 4(2), 2016, 42-57
DOI: 10.15604/ejef.2016.04.02.004
EURASIAN JOURNAL OF ECONOMICS AND FINANCE
http://www.eurasianpublications.com
UNDERSTANDING THE NEET IN TURKEY
Z. Bilgen Susanli
Isik University, Turkey. Email: bilgen.susanli@isikun.edu.tr
Abstract
The purpose of this paper is twofold. First, drawing on data from the Household Labor Force
Surveys over the period 2004-2013, it examines the determinants of the NEET (Not in
Employment, Education or Training) status for the Turkish youth. This is particularly important
for Turkey as it has the highest NEET rate among the OECD countries. Second, it describes the
movement of the youth across four states: education, employment, unemployment and
inactivity. Probit results indicate that gender and educational attainment are key factors for
explaining the NEET status. Findings also show that a greater number of household members
that are in employment is associated with a lower likelihood of NEET. Transition analyses reveal
that the state of inactivity remains highly persistent despite the substantial fall over the sample
period. In addition, the rise in the persistence of education between 2007 and 2009 underlines
the choice of the youth to stay in education in response to the fall in labor market prospects.
Keywords: Youth, NEET, Turkey
1. Introduction
Labor market experience of the youth has recently been a main concern in both the developed
and the developing countries due to the sharp increase in youth unemployment rates
throughout the global recession (OECD, 2010). It is well known that youth unemployment is
more sensitive to fluctuations in economic activity than adult unemployment.1 The global youth
unemployment rate rose to 12.9 per cent in 2009 from 11.7 per cent in 2007 and 12.7 per cent
in 2008. It stagnated around 13 per cent since 2012. 2 Many young individuals experienced long
spells of joblessness during the economic recovery from the global recession (European
Commission, 2010). During the same period, the rise in unemployment was accompanied by a
withdrawal of young people from the labor market. Between 2008 and 2014, the youth labor
force participation rate fell steadily from 49.8 per cent in 2008 to 47.3 per cent in 2014. While
many young individuals chose to stay longer in education, in many countries the decline in
falling youth labor force participation was due to discouragement (ILO, 2015).
Although the unemployment rate is a good measure of the difficulties faced by young
people in the labor market, it does not fully reflect the situation of inactive young people who are
not in education or training. A measure that captures both unemployment and inactivity is the
1 Choudry et al. (2012) document that the impact of financial crises on youth unemployment is greater than
the effect on overall unemployment.
2 Youth unemployment rates display substantial regional disparities. In the wake of global recession, the
youth unemployment rate ranged between 26 per cent and 9.3 per cent in the Middle East and in East
Asia, respectively.

Z. Bilgen Susanli / Eurasian Journal of Economics and Finance, 4(2), 2016, 42-57
43
share of youth neither in employment nor in education and training in the youth population – the
so-called NEET rate. This rate is calculated as the share of individuals aged between 15 and 24
years that are not in employment, education or training, as a percentage of the total population
of that age group.3 Hence, the NEET rate is a more comprehensive alternative to the narrower
measures such as unemployment and labor force participation rate. Since it includes the
conventionally unemployed group as well as the involuntarily jobless, sick and disabled, and
young carers, the NEET definition represents a wider range of vulnerabilities.
Unlike employment and unemployment, the NEET category lacks an international
standard definition. Therefore, the definition of this group varies across countries. For example,
in the UK the NEETs are defined as the individuals aged between 16 and 18 years who are not
in education, employment or training. The age group 16-18 corresponds to the completion of
compulsory school education (Coles et al. 2002). In Japan, the definition of the NEET status is
expanded to include the 15-34 year-olds who are not married; who are not in employment and
in education; and who are not searching for a job (Yuji, 2007). In Korea, this definition is broader
as it refers to the 15-34 year-old individuals who are not married, not in employment, not
attending school or a job training program, not searching for a job, and not handling family
responsibilities (OECD, 2007).
Between 2007 and 2012, the number of youth aged between 15 and 29 years in NEET
status has increased steeply by 2.5 million (by 7 per cent) to 38.4 million in the OECD
amounting to 16 per cent of the youth population (Carcillo et al. 2015). The NEET rate peaked
in 2010, reaching 13.1 per cent among the 15-24 year-olds. The growth of the NEET category
spurred a large body of academic research as well as European Commission and OECD
reports (Carcillo et al. 2015; Eurofound, 2012; OECD, 2010; O‟Higgins, 2012; Scarpetta et al.
2010). Evidence from the European Union indicates that the regional NEET rates are persistent,
and the persistence rises over the crisis period (Bruno et al. 2014). In addition to regional
disparities in the size of the NEET category, the NEETs within a country may itself be a
heterogeneous group (Finlay et al. 2010; Tamesberger and Bacher, 2014; Yates and Payne,
2006). Accordingly, the risk factors for becoming NEET are diverse ranging from low household
income and remote living areas to family background and immigrant status (Eurofound, 2012).
Research suggests that there are scarring effects from becoming NEET in the sense that
exposure to NEET status leads to unfavorable subsequent labor market outcomes (Bynner and
Parsons, 2002; Crawford et al. 2010). In addition, empirical evidence shows that spells of
unemployment while young lead to lower levels of happiness, job satisfaction, wages and health
in the future (Bell and Blanchflower, 2010). Due to the high prevalence of the NEETs, the public
health literature has recently shown interest in this phenomenon. A number of studies document
a positive association between the NEET status and mental health issues (Baggio et al. 2015;
Benjet et al. 2012). Making the direction of causality from NEET status to health unclear, other
studies find that individuals with prior mental health problems such as depression and anxiety
disorders are more likely to become NEETs (Herbig et al. 2013; Waghorn and Chant, 2005).
Lastly, at the macroeconomic level, the NEET youth can be considered as an unutilized
productive capacity, and hence a constraint to economic growth (Kovrova and Lyon, 2013).
Understanding the determinants of the NEET category is crucial as these individuals
are at a high risk of exclusion because they have given up both studying and looking for a job.
The growth of the NEET group may present a more difficult policy challenge than
unemployment, as it represents disconnection from the labor market as well as the society in
general (Bell and Blanchflower, 2015).
While the above mentioned studies focus on developed countries, little is known about
the situation of the NEETs in the developing world. In this regard, Kovrova and Lyon (2013)
show that the probability of NEET status moves with the business cycle in Brazil and Indonesia.
They also find that education, particularly primary school, is an important determinant in both
countries. Ranzani and Rosati (2013) document that the NEET status is persistent in Mexico,
and persistence is higher for poorer, less educated youth and for women.
3 Note that the NEET rate is not computed for the labor force and has a different denominator from the
youth unemployment rate.

Z. Bilgen Susanli / Eurasian Journal of Economics and Finance, 4(2), 2016, 42-57
44
The youth labor market in Turkey is characterized by high levels of unemployment and
inactivity. The youth unemployment rate rose steeply from 13.1 per cent in 2000 to 16.2 per
cent in 2001 due to the economic crisis of 2001. Between 2002 and 2008 it stagnated around
19 per cent and 20 per cent. In line with the contraction of the economy in 2009, the youth
unemployment rate jumped to 25 per cent, well above the OECD average of 16.7 per cent. It
declined to 21.7 per cent in 2010 and to 18.4 per cent in 2011. Since 2013, the youth
unemployment rate has stagnated around 18 per cent. With persistently higher rates than the
OECD average, youth unemployment in Turkey has recently received much attention (e.g.,
Condur and Bolukbas, 2014; Sayin, 2011; Tas, 2014). However, a less investigated subject is
the prevalence of the NEET status which also captures the inactivity. Despite the steady decline
from 42 per cent in 2004 to 27 per cent in 2013, in 2014 the NEET rate among 15-24 year-olds
was 28.4 per cent - the highest among the OECD countries. This is equivalent to about 3.5
million young individuals considering that there are 12.8 million individuals in the 15-24 age
group. The importance of the NEET problem is better understood if one considers the
demographic structure of Turkey. As of 2014, Turkey‟s population is estimated to be 77.7
million, and 16.5 per cent of this consists of persons between the ages of 15 and 24 years
(TUIK, 2015). While Turkey currently has the highest share of youth among the EU-28 and
candidate countries, with the increase in life expectancy and the fall in the total fertility rate the
population will start ageing in about 10 years (TUIK, 2014). According to the demographic
transition process, Turkey will undergo a demographic period called the “Window of
Opportunity” from 2000 to 2025. During this period, the working age population is expected to
reach its peak, which will provide more suitable conditions for economic growth.4 Therefore, the
youth population amounts to an important resource for the society, and it is crucial that the
youth are productively integrated into the labor market.
While many studies investigate the problem of youth unemployment in Turkey, very few
of them take into account the NEETs. Two exceptions are Yanik-Ilhan and Tunali (2009) and
Kilic (2014). Yanik-Ilhan and Tunali (2009) examine the evolution of the transition from school to
first permanent jobs during the 1988–2006 period and document that more than half of the
women are NEETs. Similarly, using data from the 2012 wave of the Household Labor Force
survey, Kilic (2014) documents that the category of NEET is predominantly female and more
likely to have no prior work experience.
This study has two purposes. First, using micro data from the Household Labor Force
surveys for the period 2004-2013, it estimates a probit model to examine the determinants of
the probability of being in the NEET category. Second, it describes the patterns of movement of
the youth across the following states: education, employment, unemployment, and inactivity.
Findings from probit estimations indicate that women and individuals aged between 20-24 years
are significantly more likely to be in the NEET category. Higher levels of education and a
greater number of employed individuals in the household are associated with a significantly
lower likelihood of being NEET, and this is stronger for the sample of women. Analyses of
transitions reveal useful insights. First, the state of inactivity is highly persistent and persistence
is higher for women. Second, the choice of young individuals to stay in education reflects the
difficulties with finding employment during the global recession period. Third, the difficulty of
finding a job is manifested also by the rise in the persistence of unemployment and the fall in
the persistence of employment. Lastly, transitions from education into employment remain
below the pre-recession levels even years after the global recession, which highlights the
difficulties in the school-to-work transitions.
The structure of the paper is as follows. Section 2 summarizes the dataset and the
variables used in the analyses. Section 3 describes the empirical methodology. Section 4
presents the main results. Section 5 draws some conclusions.
4 See Hosgor and Tansel (2010) and Icduygu (2012) for a more detailed discussion of Turkey‟s
demographic transition.

Z. Bilgen Susanli / Eurasian Journal of Economics and Finance, 4(2), 2016, 42-57
45
2. Data
This paper explores micro-level data from Household Labor Force Surveys (HLFS) between
2004 and 2013. HLFS are nationally representative surveys carried out yearly by the Turkish
Statistics Institute (TUIK).5 These surveys are designed to collect information on demographic,
economic and occupational characteristics of individuals aged 15 or above. In particular, they
contain information on economic activity, earnings, occupation, status in employment, tenure,
and hours worked as well as past employment status for employed persons; and information on
the duration of unemployment, occupation sought, past work experience, and job search
methods for the unemployed. Of particular interest in this study is the labor market status of the
individuals that are between the ages of 15 and 24. Among this age group, the focus of this
study is the group of individuals that are not employed and not in education or training. The
survey distinguishes between three labor market states: employed, unemployed and out-of-
labor force. The employed covers all individuals 15 years old or older who were economically
active for at least one hour during the reference period.6 Individuals who are temporarily not
working but who have a job attachment are also defined as employed. The unemployed
includes all individuals who are not in employment but who report having looked for a job during
the reference period and who are available to work within two weeks if a job is found. The
remaining individuals aged 15 or above are categorized into out-of-labor force. For this study,
individuals aged between 15 and 24 years are grouped into the following four states regarding
labor market behavior:
i. Education: students and trainees that are not employed.
ii. Employment: employed individuals, excluding the students.
iii. Unemployment: non-students who report having searched for a job in the reference
period, and who are available to work in two weeks.
iv. Inactivity: individuals who are not attending school at the time of the survey, who do not
have a job, and who report not having searched for a job in the reference period.
The last two states constitute the NEET group. The variable of interest in this study is a
binary variable that takes on the value 1 if the individual is currently in one of the last two states
and 0 otherwise.
The sample consists of 738,386 observations with a complete set of variables. Table 1
presents the summary statistics of the variables used in the analyses. Individuals in the NEET
category constitute 37 per cent of the sample. About 56 per cent of the individuals in the sample
are between the ages of 15 and 19; 54 per cent of the sample are females; and 18 per cent of
the sample are married. Around 13 per cent of the sample have less than primary school
education; and 47 per cent have secondary education. On average, there is one household
member in employment (excluding the respondent).
5 The surveys employ two-stage stratified sampling methodology. In the first stage blocks which consist of
100 households are sampled. In the second stage addresses are selected within the blocks. Stratification
is done at the Nuts 2 and urban-rural level.
6 This includes unpaid family workers that receive no pay.

Z. Bilgen Susanli / Eurasian Journal of Economics and Finance, 4(2), 2016, 42-57
46
Variables Mean Std. Dev.
Labor market status in the current period
Education 0.37 0.48
Employment 0.26 0.44
Unemployment 0.06 0.25
Inactivity 0.31 0.46
Age
Age: 15-19 0.56 0.50
Age: 20-24 0.44 0.50
Female 0.54 0.50
Married 0.18 0.38
Education
Less than primary 0.13 0.33
Primary 0.11 0.32
Secondary 0.47 0.50
High School 0.16 0.36
Vocational or Technical High School 0.09 0.28
University or more 0.05 0.21
Number of other household members in employment 1.23 1.05
Urban 0.70 0.46
Year:
2004 0.09 0.29
2005 0.11 0.31
2006 0.10 0.31
2007 0.11 0.31
2008 0.10 0.30
2009 0.10 0.30
2010 0.10 0.30
2011 0.10 0.30
2012 0.10 0.30
2013 0.09 0.29
Observations 738,386
Table 1. Summary Statistics
Source: Household Labour Force surveys 2004-2013.
Figure 1 presents the trends in the NEET rate and the four states defined above.
Between 2004 and 2013, the share of individuals in the NEET category declined substantially
from 42 per cent to 28 per cent. The decline in the NEET category can be attributed to the
remarkable increase in the share of individuals in education from 28 per cent in 2004 to 48 per
cent in 2013; and to the decrease in the share of individuals in inactivity from 35 per cent to 22

Z. Bilgen Susanli / Eurasian Journal of Economics and Finance, 4(2), 2016, 42-57
47
per cent during the same period. The share of individuals in employment decreased from about
29 per cent to 24 per cent over the sample period.
Figure 2 depicts the change in the NEET rates by gender and the region of residence.
The NEET rates for both men and women display a downward trend. Given the historically low
labor force attachment of women in Turkey, the NEET rates are substantially higher for women
than for men in both areas. The resulting gender gap is remarkably high. As of 2013, the
gender gap in the NEET rate is 23 per cent and 20 per cent in rural and urban areas,
respectively. Despite the prevalence of family enterprises in agriculture in the rural areas, the
total NEET rate in rural areas exceeds the NEET rate in urban areas in all years except 2004.
Figure 2 shows also that the NEET rates for men in rural areas increased from 24 per cent to
28 per cent between 2007 and 2009 in response to the global recession. In urban areas, it
stagnated around 21 per cent during 2007-2009 before declining to 19 per cent in 2010.
Figure 1. NEET rate and labor market states
Figure 2. NEET rates by gender in urban-rural areas
0.1
0.2
0.3
0.4
0.5
0.6
0.7
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
Rural
Urban
Women
Total
Men
0.42
0.42
0.41
0.41
0.39
0.38
0.34
0.31
0.30
0.27
0%
20%
40%
60%
80%
100%
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
Employment
Unemployed & Searching
Education
Inactivity
NEET

Z. Bilgen Susanli / Eurasian Journal of Economics and Finance, 4(2), 2016, 42-57
48
3. Empirical Methodology
The following latent variable model explains the outcome of being in the NEET category for
individual i. Let 
 be the unobserved variable determined by:

=++,
 = 1  
> 0,   = 0 .
 = 1|=
> 0)=  [>-(+ )] = +
(1)
where  is the binary outcome variable that takes on the value 1 if individual i is
observed in the NEET status, and 0 otherwise; and  = 1| is the probability that individual
i is observed in the NEET category conditional on . Assuming that the error term  follows the
standard normal distribution, . is the cumulative distribution function of the standard normal.
The vector  consists of individual and household characteristics that are expected to influence
the probability of the outcome. Among these variables are age, gender, marital status and
education. The HLFS provide age data in brackets: 15-19 and 20-24. The base category for age
is 15-19 years. Gender is controlled for by using a dummy variable, Female, that equals 1 for
females. To control for marital status, a dummy variable, Married, that equals 1 for married
individuals, is included. Educational attainment is controlled for by using a categorical variable
that is measured in terms of the highest level of education completed, with less than primary
school as the reference category. Also included in  is the number of other household
members are employed. This variable is expected to capture the labor market connections of
the household, and hence, to have a positive effect on labor force participation (Ponzo and
Scoppa, 2010). A further control variable is the dummy variable, Urban, that equals 1 if the
individual resides in an urban area, and zero otherwise. All estimations control for year
dummies to account for business cycle effects.
The parameters and  will be estimated using maximum likelihood. The marginal
effect of the continuous variable , which captures the effect of  on the probability of being
NEET, is obtained from the following partial derivative:
 = 1|

=(+)
(2)
where (. ) is the standard normal density evaluated at +. Clearly, this evaluation
depends on the particular values of the explanatory variables. Some studies report the marginal
effects evaluated at the means of explanatory values. However, this yields nonsensical
interpretations when those other controls are dummy variables. Therefore, average marginal
effects which are derived from sample averages the will be reported in this study.7 For a dummy
variable, the marginal effect is evaluated by taking the difference in the predicted probability by
setting the dummy variable at 0 and 1.
Separate probit regressions are conducted for men and women and for the urban and
rural samples to examine differences in labor market behavior across genders as well as across
urban and rural areas.
To examine the individual transitions and the role the NEET status plays in the
transitions, information regarding the labor market status of the individual in the same month the
previous year is exploited by using retrospective questions. In particular, the HLFS include a
question “What was your situation in that month one year before the survey?” and the
respondents choose from 1: Working, 2: Working at another job, 3: Retired, 4: Looking for a job,
5: Housekeeping, 6: Studying, 7: Ill, 8: Serving in the military, 9: Elderly (for 65+ year-olds), and
7 See Bartus (2005) for details on marginal effects at the means and average marginal effects.

Z. Bilgen Susanli / Eurasian Journal of Economics and Finance, 4(2), 2016, 42-57
49
10: Other.8 Using this information, the individuals are grouped into the following four states that
describe their labor market status in the previous year:
i. Education: individuals that report studying.
ii. Employment: individuals that report working at the same or at another job.
iii. Unemployment: jobless individuals that report searching for a job.
iv. Inactivity: all the remaining individuals except for those serving in the military.
Whether an individual is in education, employed, unemployed or inactive in the current
period depends on his or her status in the previous status. For this reason, following Tansel and
Kan (2012), a Markov chain model will be used in order to describe mobility across different
states using transition matrices. Assuming a homogenous Markov process, St is a random
variable defined over a discrete state space E={1,..,4}. St is called a first-order discrete Markov
chain if
=1,,1) = (=1
(3)
where z={1,…,4}. If St is a Markov chain, and  and  are elements of the state space
E, the conditional probability

 ==1=) = (1=  =)/(1=)
(4)
is called the transition probability of moving from state  to state  and can be estimated
by the maximum likelihood estimator 

=Nxy/Nx where Nxy is the number of individuals who
were in state  and moved to state  between t-1 and t; and Nx is the number of individuals who
were in state  in period t-1. The states in the current period are as defined in section 2. Each
entry in a transition matrix gives the probability of finding an individual in state y at time t given
that the individual is observed state x at time t-1. The share of youth in each state at time t is
given by P.y. By construction, the sum of elements in each row of the transition matrix equals
one. To look at differences in transitions across genders, analyses are conducted separately for
men and women. In addition, to uncover changes in movements from one year to the next,
transitions are separately estimated for each year.
4. Findings
4.1. Probit Model
Table 2 presents the probit estimation results from the set of equations (1). Estimation results
report the average marginal effects. Results in column (1) indicate that individuals aged
between 20-24 years are significantly more likely to become NEETs relative to individuals aged
between 15-19 years. This finding will remain robust across urban and rural areas as well as
across genders. Results also show that females are significantly more likely to be NEETs
relative to men. The intuition behind the average marginal effect of the gender dummy is
equivalent to comparing the average of the likelihoods of NEET status for two hypothetical
populations - one all male and one all female - that have the exact same values of all the
remaining independent variables. Accordingly, on average, women are 19.4 per cent more likely
to be in the NEET status. As expected, there is a negative and significant association between
education and the likelihood of NEET status. Column (2) proceeds by introducing the number of
other household members in employment. In all estimations, average marginal effects are
computed by setting the number of employed members in the household at the sample median
8 Information on whether an individual was in training in the same month the previous year is not collected.
The share of individuals in training amounts to 6 per cent of the sample. Analyses (not reported) by
excluding the individuals in training yields qualitatively similar results.

Z. Bilgen Susanli / Eurasian Journal of Economics and Finance, 4(2), 2016, 42-57
50
value, 1. Findings indicate that the number of employed household members has a negative
and statistically significant average marginal effect. Results also show that individuals that live
in urban areas are on average about 4.3 per cent less likely to be NEETs compared to
individuals in rural areas.
Given persistently higher NEET shares among women than among men, one can
surmise that the effect of each variable may vary across these groups. For example, it can be
expected that marriage or having a child will raise the likelihood of becoming NEET for women,
but not for men. Therefore, probit estimations are conducted separately for male and female
samples. Results in Table 3 confirm that marital status has opposite effects on the labor market
behavior of men and women. While the average marginal effect of being married is negative
and significant for men, it is positive and significant for women. Residing in an urban area and
more household members in employment are associated with a significantly lower likelihood of
being NEET for both men and women.
Table 4 presents the results from estimating equation (1) separately for the urban and
rural samples. Results are qualitatively the same for both samples. Findings indicate that age,
gender and marriage have positive average marginal effects. The average marginal effects of
education levels carry the expected signs. The impact of employed household members is
negative and statistically significant in both samples but the magnitude is much larger in the
rural sample. This finding may be explained by the prevalence of unpaid family work in rural
areas in addition to the labor market connections of the employed members of the household.
While a greater number of household members in employment may imply greater household
labor earnings in urban areas, it may not necessarily translate into greater household income in
rural areas given the prevalence of unpaid family work. Therefore, a higher household income
associated with a greater number of family members in employment may generate a stronger
income effect and hence less pressure for the jobless youth to find a job.
Dependent Variable: Pr(Neet=1) Avg. Marg. Effect Std. Error Avg. Marg. Effect Std. Error
Age group: 20-24 0.061** (0.005) 0.059** (0.005)
Gender 0.194** (0.012) 0.201** (0.012)
Marital status 0.210** (0.016) 0.212** (0.016)
Education
Primary -0.131** (0.013) -0.130** (0.011)
Secondary -0.276** (0.014) -0.282** (0.013)
General High School -0.208** (0.013) -0.222** (0.012)
Vocational or Technical High School -0.196** (0.018) -0.206** (0.017)
University or more -0.169** (0.020) -0.183** (0.020)
Urban -0.017+ (0.010) -0.042** (0.007)
No of other household members in employment -0.042** (0.007)
Observations 735,237 735,237
Log-likelihood -4.07e+05 -4.00e+05
Pseudo R-Squared 0.167 0.194
(1)
(2)
Table 2. Determinants of NEET status
Note: Probit estimation results from equation (1) are presented. Standard error in parentheses. All estimations control for year dummies.
+p<0.10, *p<0.05, **p<0.01. Omitted categories are 15-19 years old, male, unmarried, less than primary school education, and rural.

Z. Bilgen Susanli / Eurasian Journal of Economics and Finance, 4(2), 2016, 42-57
51
Dependent Variable: Pr(Neet=1) Avg. Marg. Effect Std. Error Avg. Marg. Effect Std. Error
Age group: 20-24 0.054** (0.006) 0.061** (0.004)
Marital status -0.089** (0.003) 0.325** (0.020)
Education
Primary -0.104** (0.013) -0.139** (0.021)
Secondary -0.197** (0.010) -0.333** (0.024)
General High School -0.111** (0.016) -0.304** (0.022)
Vocational or Technical High School -0.114** (0.009) -0.267** (0.027)
University or more -0.029** (0.008) -0.293** (0.031)
Urban -0.039** (0.007) -0.053** (0.008)
No of other household members in employment -0.068** (0.007) -0.021** (0.010)
Observations 341,639 396,747
Log-likelihood -1.67e+05 -2.21e+05
Pseudo R-Squared 0.051 0.194
Men
Women
Note: Probit estimation results from equation (1) are presented. Average marginal effects are reported. Standard errors in parentheses. All
estimations control for year dummies. +p<0.10, *p<0.05, **p<0.01. Omitted categories are 15-19 years old, not married, less than primary
school education, and rural.
Table 3. Determinants of NEET status, by gender
Dependent Variable: Pr(Neet=1) Avg. Marg. Effect Std. Error Avg. Marg. Effect Std. Error
Age group: 20-24 0.059** (0.005) 0.064** (0.006)
Gender 0.187** (0.011) 0.233** (0.019)
Marital status 0.251** (0.012) 0.114** (0.019)
Education
Primary -0.116** (0.010) -0.147** (0.018)
Secondary -0.301** (0.018) -0.256** (0.019)
General High School -0.237** (0.014) -0.169** (0.019)
Vocational or Technical High School -0.217** (0.019) -0.198** (0.022)
University or more -0.192** (0.021) -0.179** (0.026)
No of other household members in employment -0.008+ (0.005) -0.081** (0.008)
Observations 514,621 223,765
Log-likelihood -2.67e+05 -1.27e+05
Pseudo R-Squared 0.199 0.153
Urban
Rural
Note: Probit estimation results from equation (1) are presented. Standard errors in parentheses. All estimations control for year dummies.
+p<0.10, *p<0.05, **p<0.01. Omitted categories are 15-19 years old, male, not married, and less than primary school education.
Table 4. Determinants of NEET status, by urban-rural residence

Z. Bilgen Susanli / Eurasian Journal of Economics and Finance, 4(2), 2016, 42-57
52
4.2. Transitions
Tables 5 to 7 present the movements of the individuals across the four labor market states by
using equation (4). Table 5 presents the average flows using pooled data over the sample
period for all youth. Inactive youth constitute the majority of the NEET in the sample, and the
state of inactivity is highly persistent. About 81 per cent of the individuals that were inactive in
the previous year remain in the same state. Employment also displays a high level of
persistence: about 86 per cent of the youth employed in the previous year stay in the same
state. Individuals in unemployment exhibit a rather slow transition into employment as only
about 39 per cent of the individuals that were unemployed in the previous year moved into
employment, with 34 per cent continuing to search.
Table 6 displays the average flows by gender. The transitions from unemployment into
employment are quite similar for men and women, but persistence of unemployment is higher
for women. However, while about 25.1 per cent of the unemployed men quit searching and
move into inactivity, this is only 17.6 per cent for women. Persistence of inactivity for women is
Education Employment Unemployment Inactivity Total
Education 82.8 4.7 2.6 9.9 100
Employment 1.9 86.2 5.8 6.2 100
Unemployment 3.2 39.4 34.0 23.4 100
Inactivity 5.1 9.0 4.6 81.3 100
P.y 36.8 26.2 6.3 30.7 100
t-1
t
Table 5. Average flows, all youth
Source: Household Labour Force surveys 2004-2013.
Male
Education Employment Unemployment Inactivity Total
Education 82.1 5.7 2.6 9.6 100
Employment 1.7 88.1 6.2 4.0 100
Unemployment 2.9 39.7 32.3 25.1 100
Inactivity 4.7 37.9 19.3 38.0 100
P.y 41.2 37.3 9.0 12.6 100
Female
Education Employment Unemployment Inactivity Total
Education 83.6 3.5 2.6 10.3 100
Employment 2.2 83.0 5.1 9.7 100
Unemployment 4.3 38.6 39.6 17.6 100
Inactivity 5.2 3.9 2.1 88.9 100
P.y 33.1 16.7 4.0 46.3 100
Table 6. Average flows, male and female youth
Source: Household Labour Force surveys 2004-2013.
t-1
t-1
t
t

Z. Bilgen Susanli / Eurasian Journal of Economics and Finance, 4(2), 2016, 42-57
53
about 89 per cent, more than twice of that for men. In addition, men and women differ
substantially in the transitions from inactivity into employment. While only about 4 per cent of
the women that were inactive in the previous year moved into employment, this is about 38 per
cent for men.
Finally, Table 7 displays the annual transitions separately for each year in the sample.
Four interesting patterns emerge. First, the persistence of education displays a notable increase
starting from 2004-2005. This increase is more pronounced between 2007-2008 and 2008-
2009. In addition to this, during the same period the share of outflows from education into
inactivity declined from 14.1 per cent to 6.3 per cent. These may indicate the choice of the youth
to stay longer in education due to the impact of the global recession. Second, the effect of the
global recession is also manifested in lower transitions from education to employment. The
transitions from education into employment decreased from 5.6 per cent in 2007-2008 to 4.1 per
cent in 2008-2009. Although the outflows from education into employment rose to 4.7 per cent
in 2011, they fell to 4.3 per cent between 2011 and 2012, and to 3.8 per cent between 2012 and
2013. Third, the effect of the global recession is also reflected by the rise in the persistence of
unemployment. Between 2007-2008 and 2008-2009 the share of the unemployed youth that
remained unemployed increased from 34.1 per cent to almost 37.4 per cent highlighting the
difficulties faced in finding jobs. Accordingly, between 2007-2008 and 2008-2009 the
persistence of employment fell (from about 85 per cent to 81.7 per cent) accompanied by a
greater share transitioning into unemployment. From 2010 onwards, as the impact of the global
recession faded, the persistence of unemployment started to decline, and transitions into
employment rose sharply to 42.5 per cent. Finally, the state of inactivity remains highly
persistent despite the considerable decline from about 87 per cent to 74 per cent over the
sample period.
2003-2004
Education Employment Unemployment Inactivity Total
Education 83.4 4.7 4.0 7.9 100
Employment 0.4 89.8 4.6 5.1 100
Unemployment 1.5 35.7 39.1 23.6 100
Inactivity 2.1 6.2 4.7 87.1 100
P.y 28.3 29.4 6.9 35.3 100
2004-2005
Education Employment Unemployment Inactivity Total
Education 78.0 5.0 2.9 14.1 100
Employment 0.7 89.8 4.4 5.1 100
Unemployment 2.4 37.7 37.7 22.2 100
Inactivity 4.6 7.4 4.4 83.6 100
P.y 30.7 26.9 6.4 36.0 100
2005-2006
Education Employment Unemployment Inactivity Total
Education 78.8 5.6 2.8 12.9 100
Employment 1.0 88.5 4.8 5.7 100
Unemployment 2.2 35.1 34.8 27.9 100
Inactivity 3.3 8.0 4.1 84.6 100
P.y 31.7 27.0 6.1 35.2 100
Table 7. Annual transition tables
t-1
t-1
t
t-1
t
t

Z. Bilgen Susanli / Eurasian Journal of Economics and Finance, 4(2), 2016, 42-57
54
2006-2007
Education Employment Unemployment Inactivity Total
Education 78.0 5.0 2.9 14.1 100
Employment 0.7 89.8 4.4 5.1 100
Unemployment 2.4 37.7 37.7 22.2 100
Inactivity 3.0 7.3 4.2 85.5 100
P.y 31.0 27.5 6.4 35.1 100
2007-2008
Education Employment Unemployment Inactivity Total
Education 80.4 5.6 2.8 11.2 100
Employment 1.7 85.0 6.6 6.8 100
Unemployment 3.0 37.1 34.1 25.8 100
Inactivity 3.8 9.0 4.5 82.7 100
P.y 33.9 26.6 6.6 32.9 100
2008-2009
Education Employment Unemployment Inactivity Total
Education 85.0 4.1 2.6 8.3 100
Employment 2.4 81.7 8.9 7.0 100
Unemployment 3.8 35.7 37.4 23.1 100
Inactivity 5.2 8.8 5.6 80.3 100
P.y 37.6 24.6 7.8 30.0 100
t-1
t
t-1
t-1
t
t
2009-2010
Education Employment Unemployment Inactivity Total
Education 84.7 4.2 2.4 8.7 100
Employment 2.8 83.4 6.9 6.8 100
Unemployment 4.0 42.5 33.0 20.4 100
Inactivity 6.5 10.7 5.3 77.5 100
P.y 40.7 25.0 6.6 27.6 100
t-1
t
2010-2011
Education Employment Unemployment Inactivity Total
Education 84.7 4.7 2.2 8.4 100
Employment 2.9 85.0 5.6 6.6 100
Unemployment 3.6 45.7 29.4 21.4 100
Inactivity 7.5 11.9 4.8 75.7 100
P.y 42.6 26.1 5.7 25.6 100
t-1
t

Z. Bilgen Susanli / Eurasian Journal of Economics and Finance, 4(2), 2016, 42-57
55
5. Conclusion
This paper aims to examine the determinants of the NEET status for the Turkish youth between
2004 and 2013. Findings from probit estimations indicate that higher levels of education and a
greater number of household members in employment are associated with a significantly lower
likelihood of being NEET; and this is stronger for women. In addition, marriage seems to be a
significant determinant of the NEET status for women. The trends in the mobility of the youth
across different states over the period of 2004-2013 are also examined. Despite the
shortcomings of the available data that allow focusing on movements across two points in time,
the transition analyses provide useful insights. First, the majority of the NEET group are the
inactive individuals; and the state of inactivity is highly persistent for women. Second, after
2007-2008 the persistence of education rose sharply indicating the choice of the youth to stay in
education as their labor market prospects fell. Third, while the outflows from unemployment into
employment recovered after the recession period, transitions from education into employment
remain below the pre-recession period levels which underscore the need for examining the
school-to-work transition experience.
The findings of this study suggest that labor market policies and future research that
address the NEET problem should take into consideration the gender dimension of the issue by
taking into account the reasons for inactivity. Education is a key factor in lowering the likelihood
of NEET status. Nevertheless, while the youth's choice of staying in education is favorable,
additional years of schooling will not necessarily translate into better labor market outcomes
unless followed by an efficient school-to-work transition. Therefore, understanding the job
search behavior becomes essential for developing better policies to address the challenges
faced by the youth in the labor market.
2011-2012
Education Employment Unemployment Inactivity Total
Education 85.4 4.3 2.0 8.2 100
Employment 3.3 84.0 5.8 6.9 100
Unemployment 4.4 43.6 26.4 25.7 100
Inactivity 8.8 11.2 4.3 75.7 100
P.y 44.9 24.7 5.1 25.4 100
2012-2013
Education Employment Unemployment Inactivity Total
Education 87.8 3.8 2.1 6.3 100
Employment 4.2 82.5 6.3 7.1 100
Unemployment 5.0 44.7 28.5 21.8 100
Inactivity 10.2 11.9 4.5 73.5 100
P.y 48.2 24.2 5.2 22.4 100
t-1
t
t-1
t
Source: Household Labour Force surveys 2004-2013.

Z. Bilgen Susanli / Eurasian Journal of Economics and Finance, 4(2), 2016, 42-57
56
References
Baggio, S., Iglesias, K., Deline, S., Studer, J., Henchoz, Y., Mohler-Kuo, M., and Gmel, G.,
2015. Not in education, employment, or training status among young Swiss men.
Longitudinal associations with mental health and substance use. Journal of Adolescent
Health, 56(2), pp.238-243. http://dx.doi.org/10.1016/j.jadohealth.2014.09.006
Bartus, T., 2005. Estimation of marginal effects using margeff. The Stata Journal, 5(3), pp.309-
329.
Bell, D.N.F. and Blanchflower, D.G., 2010. Youth unemployment: Déjà vu?. IZA Discussion
Paper, No.4075.
Bell, D.N.F. and Blanchflower, D.G., 2015. Youth unemployment in Greece: Measuring the
challenge. IZA Journal of European Labor Studies, 4(1), pp.1-25.
http://dx.doi.org/10.1186/2193-9012-4-1
Benjet, C., Hernandez-Motoya, D., Borges, G., Mendez, E., Medina-Mora, M.E., and Aguilar-
Gaxiola, S., 2012. Youth who neither study nor work: Mental health, education and
employment. Salud Publica de Mexico [Public Health Mexico], 54(4), pp.410-417.
http://dx.doi.org/10.1590/S0036-36342012000400011
Bruno, G., Marelli, E., and Signorelli, M., 2014. The rise of NEET and youth unemployment in
EU regions after the crisis. Comparative Economic Studies, 56(4), pp.592-615.
http://dx.doi.org/10.1057/ces.2014.27
Bynner, J. and Parsons, S., 2002. Social exclusion and the transition from school-to-work: The
case of young people not in education, employment, or training (NEET). Journal of
Vocational Behavior, 60(2), pp.289-309. http://dx.doi.org/10.1006/jvbe.2001.1868
Carcillo, S., Fernandez, R., Konigs, S., and Minea, A., 2015. NEET youth in the aftermath of the
crisis: Challenges and policies. OECD Social, Employment and Migration Working
Papers, no.164. Paris: OECD Publishing.
Choudry, M.T., Marelli, E., and Signorelli, M., 2012. Youth unemployment rate and the impact of
financial crises. International Journal of Manpower, 33(1), pp.76-95.
http://dx.doi.org/10.1108/01437721211212538
Coles, B., Hutton, S., Bradshaw, J., Craig, G., Godfrey, C., and Johnson, J., 2002. Literature
review of the costs of being "Not in Education, Employment or Training" at age 16-18.
Department for Education and Skills Research Report, No.347.
Condur, F. and Bolukbas, M., 2014. Türkiye'de iş gücü piyasası ve genç işsizlik büyüme ilişkisi
üzerine bir inceleme [An analysis of the relationship between labor market, youth
unemployment an growth in Turkey]. Amme Idaresi Dergisi [Public Administration
Journal], 47(2), pp.77-93.
Crawford, C., Duckwoth, K., Vignoles, A., and Wyness, G., 2010. Young people‟s education and
labor market choices aged 16/17 to 18/19. UK Government Department of Education
Research Report, No.182.
Eurofound, 2012. NEETs - Young people not in employment, education or training:
Characteristics, costs and policy responses in Europe. Luxembourg: Publications Office
of the European Union.
European Commission, 2010. Employment in Europe 2009. Brussels: European Commission.
Finlay, I., Sheridana, M., McKayb, J., and Eudora, H., 2010. Young people on the margins: In
need of more choices and more chances in twenty-first century Scotland. British
Educational Research Journal, 36(5), pp.851-867.
http://dx.doi.org/10.1080/01411920903168532
Herbig, B., Dragano, N., and Angerer, P., 2013. Health in the long-term unemployed. Deutsches
Arzteblatt International [German Arzteblatt International], 110(23-24), pp.413-419.
Hosgor, S. and Tansel, A., 2010. 2050’ye doğru nüfus bilim ve yönetim: Eğitim, işgücü, sağlık
ve sosyal güvenlik sistemine yansımalar [Demography and Management towards 2050:
Education, labor force, and social security]. Publication No. TUSIAD-T/2010/11/505.
İstanbul: Türkiye Sanayicileri ve İşadamları Derneği ve Birleşmiş Milletler Nüfus Fonu
[Turkish Industry and Business Association and United Nations Population Fund].

Z. Bilgen Susanli / Eurasian Journal of Economics and Finance, 4(2), 2016, 42-57
57
Icduygu, A., 2012. Demography and immigration/emigration. In: M. Heper and S. Sayarı, eds.
2012. The Routledge Handbook of Modern Turkey. Abingdon: Routledge. pp.328-337.
International Labor Organization (ILO), 2015. Global employment trends for youth 2015: Scaling
up investments in decent jobs for youth. Geneva: International Labor Office.
Kilic, Y., 2014. Young people in Turkey who are not in education, employment or training
(NEET). Education and Science, 39(175), pp.121-135.
Kovrova, I. and Lyon, S., 2013. NEET youth dynamics in Indonesia and Brazil: A cohort
analysis. Understanding Children’s Work (UCW) Programme Working Paper Series,
March 2013.
O‟Higgins, N., 2012. This time it‟s different? Youth labor markets during the „the Great
Recession‟. Comparative Economic Studies, 54(2), pp.395-412.
http://dx.doi.org/10.1057/ces.2012.15
OECD, 2007. Jobs for Youth/Des emplois pour les jeunes: Korea 2007. Paris: OECD
Publishing.
OECD, 2010. Off to a good start? Jobs for youth. Paris: OECD Publishing
Ponzo, M. and Scoppa, V. 2010. The use of informal networks in Italy: Efficiency or favoritism?.
The Journal of Socio-Economics, 39(1), pp.89-99.
http://dx.doi.org/10.1016/j.socec.2009.07.007
Ranzani, M. and Rosati, F., 2013. The NEET trap: A dynamic analysis for Mexico.
Understanding Children's Work (UCW) Project Working Papers, March 2013.
Sayin, F., 2011. Türkiye‟de 1988-2010 döneminde eğitim ve büyümenin genç işsizliğine
etkisinin analizi [The analysis of the impact of education and growth on youth
unemployment in Turkey between 1988-2000]. Dokuz Eylül Üniversitesi Sosyal Bilimler
Enstütüsü Dergisi [Dokuz Eylul University Journal of Graduate School of Social
Sciences], 13(4), pp.33-53.
Scarpetta, S., Sonnet, A., and Manfredi, T., 2010. Rising youth unemployment during the crisis:
How to prevent negative long-term consequences on a generation?. OECD Social,
Employment and Migration Working Papers, No.106.
Tansel, A. and Kan, E.O., 2012. Labor Mobility across the formal/informal divide in Turkey:
Evidence from individual level data. IZA Discussion Papers, No. 6271.
Tamesberger, D. and Bacher, J., 2014. NEET youth in Austria: A typology including socio-
demography, labor market behavior and permanence. Journal of Youth Studies, 17(9),
pp.1239-1259. http://dx.doi.org/10.1080/13676261.2014.901492
Tas, H.Y., 2014. İşkur‟un mesleki eğitim faaliyetlerinin istihdam üzerine etkileri: Yalova İşkur
örneği [The effects of vocational training course on the employment: The case of Yalova
Employment Agency]. Journal of Social Policy Conferences, 61(2), pp.153-176.
TUIK, 2014. İstatistiklerle gençlik, 2013 [Youth in Statistics]. Haber Bulteni [Bulletin], No.16055.
Ankara: TUIK.
TUIK, 2015. İstatistiklerle gençlik, 2014 [Youth in Statistics]. Haber Bülteni [Bulletin], No.18625,
Ankara: TUIK.
Waghorn, G. and Chant, D., 2005. Labor force activity by people with depression and anxiety
disorders: A population-level second-order analysis. Acta Psychiatrica Scandinavica,
112(6), pp.415-424. http://dx.doi.org/10.1111/j.1600-0447.2005.00600.x
Yanik-Ilhan, B. and Tunali, I. 2009. Türkiye'de ı̇ş piyasasına geçiş aşamasındaki gençler
[Turkish youth in transition to the job market]. ERF Economic Research Forum,
Research Note: 09-1.
Yates, S. and M. Payne. 2006. Not so NEET? A Critique of the use of NEET in setting targets
for interventions with young people. Journal of Youth Studies, 9(3), pp.329–344.
http://dx.doi.org/10.1080/13676260600805671
Yuji, G., 2007. Jobless youths and the NEET problem in Japan. Social Science Japan Journal,
10(1), pp.23-4. http://dx.doi.org/10.1093/ssjj/jym029