Exploring the Diversity of Young People Not in Employment, Education or Training (NEET): The Gender Profile of NEETs in Georgia and Armenia1 South Caucasus Gender Program (P160432) April 2019 Poverty and Equity Global Practice Europe and Central Asia Region 1 Team: Paola Buitrago-Hernandez (E.T. Consultant, Poverty Global Practice), Alan Fuchs (Senior Economist, Poverty Global Practice, co-TTL) and Cesar Cancho (Economist, Poverty Global Practice, co-TTL). The team is grateful for the valuable comments received from Jonna Maria Lundvall (Social Scientist, GPV01) and Natalia Millan (Economist, GSP03). Table of Contents Summary ............................................................................................................................................................... 3 Motivation ............................................................................................................................................................. 5 Data Sources and Methodology ............................................................................................................................. 7 The Gender Profile of NEETs in Georgia and Armenia ............................................................................................ 8 General Picture of Young People...............................................................................................................................8 The Profile of Male and Female NEETs ...................................................................................................................13 Women are more likely to be NEET than men, and this disadvantage is more pronounced in rural areas ......14 The gender gap in NEET rates starts at age 20 and increases with age .............................................................14 In Georgia the NEET condition is more urban, while in Armenia it is both urban and rural; yet, gender differences in the likelihood of being NEET persist in both areas......................................................................15 NEETs are heterogeneous: Inactivity is more prevalent among NEET women whereas unemployment is more the condition among NEET men ........................................................................................................................16 Educational attainment does not necessarily protect men and women against being NEET............................18 Both male and female NEETs belong to poorer households..............................................................................20 Previous work experience among NEET is scarce, particularly for women .......................................................21 NEET rates are higher for those men and women from a minority group or with migration background .......22 Gender Differences in the Pathways to Becoming NEET ...................................................................................... 24 While most young men transition from school to work, most young women remain economically inactive and out of the labor force for longer periods ...........................................................................................................24 Family formation plays a role: compared to the average of young women (15-29), female NEET marry at higher rates ........................................................................................................................................................26 Social norms on care, work and motherhood, and the typical gendered roles that steer women towards household work, help shaping the gender differences in the transition pathways...........................................27 What Characteristics are correlated to Being NEET? A Regression Analysis ......................................................... 29 Conceptual Framework ......................................................................................................................................29 Household composition and marital status stand out as important correlates of being NEET for women but not men; for all other correlates, the direction of the effect is the same but the size is not ............................30 Concluding Remarks ............................................................................................................................................ 36 References ........................................................................................................................................................... 39 Annex: Additional Figures and Tables .................................................................................................................. 41 Summary Georgia and Armenia have considerably high shares of youth who are detached from both the labor market and the education and training systems. Around one third (31 percent in Georgia and 33 percent in Armenia) of the young population in the 15-29 age group are not in employment, education, or training, commonly referred as NEET. International comparisons among people in the same age group, indicate that NEET rates in the Czech Republic, Slovenia, and Lithuania are less than 10 percent, in Romania 18 percent, in Turkey 28 percent, and the proportion in the EU28 is about 13 percent (Eurostat, 2017). Furthermore, the existing gender differences in NEET rates indicate that young women are at a higher risk than men to fall under this condition in both countries. High NEET rates indicate that young people -specially women- in Georgia and Armenia are not building up skills and are disproportionately at risk of labor market exclusion, despite having high educational attainment. Moreover, those who are NEET today run a high risk of remaining outside the labor market in the future, resulting in an inefficient allocation of resources.2 This is particularly problematic in the context of two countries confronted by demographic forces of population shrinkage and aging active labor resources, which is also the case in most of the ECA region. In the case of women, the conflicting demand of their time for care and work activities represents a fundamental barrier to economic participation and generates a vicious circle of low labor market attachment and prominence of the care provider role that leads to increased vulnerability and gender- based inequalities (World Bank, 2019). These barriers manifest earlier in the life of youth as the different school-to-work transition pathways between men and women show: while most young women in Georgia and Armenia will not enter at all or exit the labor force (i.e. become economically inactive) after finishing school, most young men will be either working or looking for a job. This note analyses the gender profile of NEETs in Georgia and Armenia according to individual and household characteristics, the gender differences in the school-to-work pathways, and the characteristics or risk factors that are correlated to being NEET and their differential effect between men and women. The key findings from this note can be summarized as follows: 1. NEET rates in Georgia and Armenia are high compared to other ECA and EU peers, and this is true for both men and women. However, there is a consistent disadvantage towards young women in falling under this condition, particularly in Armenia where the share of women who are NEET is almost twice the share among men (42 percent women vs. 23 percent men in Armenia; 37 percent women vs. 25 percent men in Georgia). 2. The NEET condition is more prevalent in urban areas in Georgia (explained in part by high unemployment among urban young men), whereas in Armenia is as an urban as rural phenomenon. Yet, despite these differences, the gender gaps in NEET rates between rural women and rural men are consistently higher than among their urban counterparts in both countries; this, partly driven by lower observed NEET rates among rural men (compared to urban men) due to subsistence employment in the agriculture sector. 3. The gender gap in NEET rates emerges at about age 20 -which coincides with the beginning of the peak productive and reproductive ages (20 to 35-year-old age group)- and widens with age. While 2 Bynner and Parsons (2002); Furlong (2006); ILO (2012). 3 at age 15 the NEET rate is below 10 percent in both countries and for both sexes (as most youngsters are still in school), by age 29 as many as half young women will have fallen under this condition. 4. The NEET indicator involves heterogeneous conditions and therefore heterogeneous populations. Economic inactivity (which excludes those in education or training) is the most prevalent condition for non-employment among NEET women, whereas unemployment is the most common status among NEET men. Yet, there is a sizable share of NEET -men and women- who are discouraged or unavailable. 5. In Georgia and Armenia, the school-to-work transition is very different for males and females: while most young men transition from school to employment, most young women remain inactive and out of the labor force for longer periods as they stay home. NEET women marry at higher rates than average young women which indicates that family formation -and the norms attached to it- play an important role in the different observed transition pathways. 6. The perceived association of education in lowering the numbers of NEETs is not necessarily guaranteed. Georgia and Armenia are cases where graduates of upper secondary/TVET succeed less in finding jobs than those with lowest and highest levels of education. This is also an indication of the importance that skills mismatch plays for job placement in both countries. 7. Household income and the NEET condition are strongly correlated, and this is true for men and women. Around 60 percent of NEET youth live in households in the bottom 40 percent of welfare aggregate. Similarly, the average NEET -man or woman- lives in a household of a lower income quintile compared to their peers in school, employed, or active and studying. 8. Regression analysis shows that, among individual and household correlates, household composition (i.e. presence of children in the household) and marital status stand out as important correlates to being NEET among women (but not for men). For all other correlates, the direction of the effect is the same but not the magnitude. Neither country have yet focused on the NEET phenomenon and no policy measures specifically targeting NEETs exist. Rather, general youth policies have been developed to address the different problems affecting young people. As it is the case across the ECA region, the most common target group in these policies is ‘unemployed youth’, but other important vulnerable groups such as family caregivers, discouraged workers, disabled or unavailable ‘drop off the radar’. Tackling information gaps regarding these vulnerable groups is one first step in developing appropriate and targeted policy interventions. While young people under the NEET classification suffer from some kind of exclusion, from either education and training or employment, the reasons for their exclusion might be completely different. National policies require a tailored yet articulated approach to mitigate failed school-to-work transition and the NEET phenomenon, considering the heterogeneous sub-groups within the NEET category. The policy response should include policies and support structures in relevant fields such as childcare3, the VET system, employment, healthcare, housing, transport and cultural (e.g. social norms) considerations. 3For instance, there is rich evidence that increased availability of formal childcare options results in improved labor force participation of women in many different contexts—in Brazil (Deutsch 1998; Paes de Barros et al. 2011); in rural Colombia 4 Motivation Youth’s quality of life is largely determined by how successfully they make the transition from school to work (ILO, 2012). In most of the East Europe and Central Asia region (ECA) this transition is burdened by specific challenges that result in relatively low employment coupled with high unemployment and high economic inactivity; the South Caucasus is not the exception. In Georgia and Armenia job opportunities for young people are limited, with between one fifth and one third of employed youth working in subsistence agriculture – a low-productivity, low-growth sector. Moreover, about one in three youth age 15-29 in each country (31 percent in Georgia, 33 percent in Armenia) does not have a job and is not participating in any in education or training, commonly referred as NEET. (Georgia, Armenia LFS 2017). Young women in both countries are at a higher disadvantage of falling under the NEET condition: in Georgia, 37 percent of young women in the 15-29 age group are NEET compared to 25 percent of young men; in Armenia the shares are even higher among women, and the gender disparity is wider with the female NEET rate being double that of men (42 percent and 23 percent, respectively). These gender differences reflect the fact that transition patterns from education to work are very different by sex. In fact, while women in these two countries will mostly transition into inactivity after finishing school, men will either get a job or be looking for one. Furthermore, relative to the regional peers, international comparison of NEET rates (for the 15-24 age group) show that Armenia and Georgia, along with Kosovo, rank among the highest in ECA (25 percent Georgia, 27 percent Armenia) (Figure 1).4 Looking at a wider age group, the 15-29, Armenia and Georgia also stand as countries with the highest NEET rates when compared to peers such as the Czech Republic, Slovenia, and Lithuania where rates are less than 10 percent, Romania (18 percent), Turkey (28 percent), and the EU28 average of 13 percent (Eurostat, 2017). Figure 1. Percent of NEET (youth 15-24) in selected comparator countries, 2017 Kosovo Armenia Georgia North Macedonia Bosnia and Herzegovina Turkey Europe & Central Asia (excluding high income) Romania Russian Federation Latvia Poland Lithuania Slovenia Czech Republic 0 5 10 15 20 25 30 35 Source: World Bank WDI for all countries except Georgia LFS (2017) and Armenia LFS (2017). (Attanasio and Vera-Hernandez 2004); in urban Argentina (Berlinski and Galiani 2007); in Japan (Asai, Kambayashi, and Yamaguchi 2015); in Canada (Lefebvre and Merrigan 2008). Closer to the region, in the European Union (Del Boca and Locatelli 2006); in Romania (Fong and Lokshin 2000); in the Russian Federation (Lokshin 2000); and in Turkey (World Bank 2015). 4 NEET rates in the 15-24 age group are lower than in the 15-29 age group as most youth in the former group are still attending school. 5 Notes: Data should be interpreted with caution because of possible differences in age coverage: countries vary somewhat in their operational definitions, some can go beyond age 24 considering that not all people complete their education by this age. No data available for Azerbaijan. Data for Kosovo, ECA, and Russian Federation refer to year 2016. Low rates of overall labor force participation and high proportions of NEETs suggest that Georgia’s and Armenia’s youth -especially women- are more likely to experience labor market exclusion. Since the NEETs are neither investing in skills nor gaining experience through employment, they have limited chances of finding productive employment in the future (World Bank, 2017) and run a high risk of remaining outside the labor market (Bynner and Parsons, 2002; Furlong, 2006; ETF, 2015), resulting in a waste for the economy, society and the individuals themselves. The longer a young person is both out of work and education, the more likely it is that he/she becomes ‘unemployable’ (UNFPA, 2017). It is important to note that the NEET indicator refers to a very heterogeneous population. It embraces a range of young people in difficult yet different situations. One sub-group are the unemployed (or job seekers), those who are without work, but currently available for work, and seeking work during a reference period (not currently receiving any education/training). The other sub-group are the (economically) inactive5 youth (not currently receiving any education/training), which at the same time contain different categories based on the reasons for the inactivity i.e. Re-entrants, Discouraged, Family caregivers/homemakers, Other inactive (sick/disabled), and Unavailable (ETF 2015; Eurofound 2016). Members of these different sub-groups have very different experiences, characteristics and needs. For instance, the different experience of the NEET condition between men and women find its roots mainly in gender roles about the role of women in the household. In many countries, childcare, eldercare and household responsibilities are primarily the domain of women, a task which prevents them from participating in economic life (World Bank 2016, 2019). Insufficient social services to reconcile family and work tend to keep this gender gap wide (ETF 2015, Carcillo et al. 2015). Georgia and Armenia cannot afford to underutilize a large share of youth whose lifetime productivity is being harmed by the experience of exclusion, from either education and training or employment. This is particularly relevant in the context of demographic forces driving population shrinkage and aging labor resources, which are characteristic of countries in the ECA region. The effect of demographics is further exacerbated by the fact that a significant share of women does not participate in the labor market, despite having higher educational attainment than men. (World Bank 2018, 2017a) As of today, most ECA countries have not yet focused on the NEET phenomenon and therefore no overall policy measures specifically targeting NEETs exist. Rather, general youth (and employment) policies have been developed. The most common target group in these policies is ‘unemployed youth’, but other important vulnerable sub-groups such as inactive young girls, family caregivers, discouraged workers and other inactive are left ‘off the radar’. (World Bank, 2017 and 2016a; ETF, 2015). Moreover, the lack of targeted policy interventions is partly explained because there is little information available on the different vulnerable subgroups. The purpose of this study is to characterize the profiles of female and male NEET according to individual and household characteristics, the different pathways they follow when transitioning from school to work, and the risk factors than matter the most in explaining the likelihood of becoming NEET. The study 5 The term ‘inactive’ in this note refers to the economically inactive and it excludes people in education or training. 6 expands previous work on NEETs produced in the South Caucasus6 by adding a gender dimension to the analysis, and to show the experience of becoming/being a NEET is very different between men and women. The note is structured as follows: Section 2 introduces the methodology and datasets that are the basis for this report. Section 3 provides a general picture of Georgia’s and Armenia’s youth, and then presents in more detail the gender profiles of NEETs of both countries. Section 4 discusses the gender differences in the pathways to becoming NEET from a simple characterization of the school-to-work transition. Section 5 presents results from regression analyses of the individual and household correlates that matter the most in being a NEET and discusses differences between women and men. Section 6 concludes by summarizing the main findings, presenting key questions not yet answered in the literature and discussing the policy response so far. Data Sources and Methodology The analysis presented here is based on two major surveys: the Georgia Labour Force Survey (LFS, 2017) and the Armenia Labour Force Survey (LFS, 2017), which are both representative at the national level and produced by National Statistics Office of Georgia (GeoStat), and the National Statistical Service of the Republic of Armenia (NSS), respectively.7 They include information on socio-demographic characteristics and status in the labor force and are the most recent datasets on labor market.8 Supplementary input from other sources was also included: the Georgia Household Integrated Survey (HIS, 2016) and the Armenia Integrated Living Conditions Survey (ILCS, 2017) were used to gather information on the aggregate welfare/income of households where youth live and, in the case of Armenia, information on the demographic composition of their household and migration as well.9 The terms ‘youth’ and ‘young people’ are used interchangeably and refer to individuals between the ages of 15 and 29 years old, which allows to observe the status of youth who have finished school and better capture gender differences in school-to-work transitions (indicators for people 15-24 are presented in Table A1 in the Annex). With regards to ‘young people neither in employment nor education and training’ (NEET) the study follows Eurostat’s definition of the indicator: the numerator refers to persons meeting any of the following conditions: i) they are unemployed and have not received any formal or non-formal education or training in the four weeks preceding the survey, or ii) they are economically inactive and 6 This study builds up on previous work produced by the World Bank (2017) under Technical Assistance (TA) No. P161162. This TA produced quantitative analysis of the situation of NEET youth in Georgia and reviewed existing youth programs and services available in the country. The task was conducted upon the request of the Ministry of Sports and Youth Affairs (MSYA) of Georgia. There are no major studies on NEET focusing on Armenia, except for a broader analysis of EU neighbor countries conducted by the European Training Foundation regarding the situation of NEET and the policy responses. More detailed research on NEET focuses in EU and OECD countries. 7 The Armenia LFS was downloaded from the website of the Statistical Committee of Armenia on 03/19/19. 8 The Armenia LFS records young men who are carrying out compulsory military service as inactive population. Following Eurostat’s guidelines, these observations have been excluded from the calculations of labor indicators presented in this repo rt as these are individuals that are not eligible to participate in the labor force. This adjustment reflects on the following shares of dropped observations per age group: 37 percent of men age 18; 63 percent of men age 19; and 26 percent of men age 20. When considering these individuals in the sample, figures indicate that men in the 18-20 age group will be either in compulsory military service (42.1 percent) or enrolled in school (28 percent). See https://ec.europa.eu/eurostat/web/microdata/european-union-labour-force-survey. 9 The Armenia LFS (2017) dataset only records wage of individuals who are in employment whereas the ICLS (2017) collects all sources of income perceived by members of the household. 7 have not received any formal or non-formal education or training in the four weeks preceding the survey (Figure 2). Participation in education or vocational training includes both full and part-time formats since surveys do not allow to differentiate duration.10 The denominator is the total population in the 15-29 age group and it excludes respondents without information on labor status. 11 The analysis follows two steps: in a first step, the study focuses on constructing the gender profile of NEETs in both Georgia and Armenia, and the pathways of men and women in becoming a NEET based on their school-to-work transition dynamic. As a second step, the study explores the characteristics that are correlated to becoming a NEET by estimating the marginal effects at the means of a set of candidate variables (and controlling by region), using probit models separately for men and women, and comparing whether the estimated correlates vary by sex (in magnitude and/or direction). To complement the analysis, linear probability models of being a NEET are also estimated. Figure 1: The NEET Population Source: ETS (2015) pp. 10. The Gender Profile of NEETs in Georgia and Armenia General Picture of Young People 10 Only the Georgia LFS allows to identify whether (non-formal) training is full or part-time. The Georgia HIS does not allow to identify individuals in training and therefore, when using data from this survey, NEETs are defined as young people not in employment nor in formal education only. 11 See: https://ec.europa.eu/eurostat/statistics- explained/index.php/Glossary:Young_people_neither_in_employment_nor_in_education_and_training_(NEET) 8 In Georgia, young people (15-29) account for one fifth of the total population and they concentrate towards the upper end of the age distribution with 70 percent in the 20-29 age bracket (Table 1). There are more youth men (53 percent) than youth women (47 percent). Georgian youth are mostly urban, with 61 percent of them living in this area; by sex, among young women, a larger proportion live in the cities relative to young men. Youth live in five regions mostly, which together concentrate 70 percent of the youth.12 In terms of marriage and family structure, while most young people are single, the proportion of ever married or in union women in the 15-29 age group is 45 percent, which is double the share among men (23 percent).13 As of their position in the household, only a minority declare being the head with higher shares among men than women (7 vs. 2 percent), and most of them are daughters or sons, grandchildren, daughter or son-in-law and, in the case of women, spouses of the head of the household. The percentages of youth reporting a disability or declaring IDP status are low (2 and 4 percent, respectively).14 In Armenia, young people represent one fourth of the total population. Even when they concentrate towards the 20-29 age group, Armenian youth are more evenly distributed between the 20-24 and the 25-29 brackets. Furthermore, the female-to-male ratio is higher in the 20-24 age group, for which there are 1.0 young women for every man, whereas in the 15-19 bracket the ratio is 0.8. These ratios are likely to reflect the strong male emigration pattern of the country. Five percent of male youth declare being returned migrants (according to ILCS, 2017 which collects migrants who returned after 2013). Youth are mostly urban, but the shares are not as high as in Georgia, meaning there are more youth living in rural areas in Armenia compared to Georgia. The shares of ever married/in union youth are lower than in Georgia, but the proportion of married young women is double that of married young men. The shares of those reporting household headship are low and similar to Georgia, with most youth reporting being children or other relatives of the head of household. By administrative region, youth are concentrated in seven areas which together represent 85 percent of this population.15 Table 1: Characteristics of Youth (15-29) Georgia Armenia Women Men Total Women Men Total % % % % % % # People 15-29 307,354 383,917 691,271 270,522 275,301 545,823 Share among 15-29 47 53 100 50 50 100 % total population 19 27 Age group 15_19 29 31 30 25 30 28 20_24 31 29 30 35 32 34 25_29 40 40 40 39 38 38 Total 100 100 100 100 100 100 Location Urban 62 59 61 60 53 56 Rural 38 41 39 40 47 44 Total 100 100 100 100 100 100 Status in labor force 12 Tbilisi (35 percent), Imereti (13 percent), Kvemo Kartli (12 percent) and Adjara (10 percent). 13 According to UNFPA (2017), the mean age at birth in Georgia is 26.3 years and girls between 15 and 18 years of age in rural areas are more likely to enter a marriage than those in urban areas. 14 Nevertheless, according to UNFPA (2017) youth constitute 21 percent of all displaced people. 15 Yerevan (27 percent), Armavir, Ararat, Lori and Shirak (10 percent each), Kotayk and Gegharkunik (9 percent each). 9 Georgia Armenia Women Men Total Women Men Total % % % % % % Education/training (not in employment) * 27 26 26 32 32 32 Employed 36 49 43 27 45 35 Unemployed (NEET) 9 13 11 11 14 12 Inactive (NEET) 27 12 19 31 10 21 Total 100 100 100 100 100 100 Educational attainment (formal education only) Lower Secondary or less 12 16 14 (Incomplete Sec or less, for Armenia) 24 27 26 Upper Secondary 46 57 52 (Secondary, for Armenia) 39 46 43 Vocational 12 9 10 15 9 12 Tertiary 25 18 21 28 18 22 Total 100 100 100 100 100 100 % early leavers from education and training (among people 18-24)** 8.3 9.5 8.9 2.1 3.3 2.7 Nationality (Georgia) / Place of Origin (Armenia) Georgian 88 88 88 1 0 0 Armenian 97 97 97 Azeri 0 0 0 Other*** 12 12 12 2 2 2 Total 100 100 100 100 100 100 Other Ever married/in union 45 23 37 41 18 29 Has a disability 1 2 2 1 2 1 IDP 4 5 4 Returned migrant**** 2 5 3 (after 2013) Head of Household 2 7 5 2 6 4 Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). Note: Empty cells indicate variable is not available in dataset. */ Refers to non-working people receiving any type of formal education and/or non-formal education or vocational training and includes full or part-time formats. **/ As per Eurostat definition: percentage of people aged 18 to 24 who have completed at most lower secondary education and is not involved in further education or training. ***/ For Georgia refers to all other nationalities; for Armenia, refers to the following places of origin: Russia, Other CIS (Ukraine, Belarus, Moldova, Kazakhstan, Uzbekistan, Kyrgyzstan, Tajikistan, Turkmenistan), Syria, and all other. The Georgia LFS (2017) does not include variable of ‘place of origin’ but ‘nationality’. ****/ Source is Armenia ILCS (2017); no information on returned migrants available in the Armenia LFS (2017). More than a quarter of youth in both countries attend/are enrolled in school or receive any training/vocational education, and the shares are higher in Armenia.16 When it comes to participation in education or training, it is important to distinguish between the youngest cohort, ages 15-22 for example, and the oldest cohort ages 23-29. Those in younger age brackets are still predominantly in their human 16 Neither LFS allows to identify educational level currently attended by those enrolled in school. 10 capital accumulation stage whereas those aged 23+ had join/tried to join the labor market. In Georgia for example, as many as 51 percent of women and 46 percent of men age 15-22 declare being in school, the shares among those 23_29 go down to 1.8 percent among women and 1.5 percent among men; the pattern is similar in Armenia (see Table A2 in the Annex). Educational attainment figures suggest that access to education is not a binding constraint in neither country. More than half youth in Georgia completed upper secondary or a higher non-tertiary level (including vocational education), and a non-negligible share of youth (21 percent) have a bachelor’s degree and/or a postgraduate degree. In Armenia, most of the youth (64 percent) have completed either Secondary or a higher non-tertiary level, and over one fifth of the young population hold a tertiary degree. In both countries more females (Georgia 25/Armenia 28 percent) than males (18 percent, both countries) have attained tertiary education, and these shares are very much in line to the EU28 averages of tertiary attainment (24 percent among women and 18 percent among men, according to Eurostat, 2018). The shares of ‘early leavers’ (percentage of 18-24 olds who have completed at most a lower secondary education and are not in further education or training) are below those of the EU28 (10.6 percent) and are lower among young women compared to young men. As a reference, the share in Georgia is the same level as in France, and Armenia, where the share is even lower, stands as the lowest if compared to EU countries (Eurostat, 2018). Labor force participation (share of employed and unemployed) is considerably lower for young women than for young men, with gender gaps of 17 and 22 percentage points in Georgia and Armenia, respectively. In addition, young women face greater challenges than men to secure employment: in Armenia the share of women who are employed is 18 percentage points less than the share among men, and in Georgia the gap is 13 percentage points (see more on employment in Box 1). The share of unemployment is higher among men than women in both Georgia and Armenia, and the opposite is true for the inactive with many more inactive women than men. In fact, there are three times as many economically inactive young women than men in Armenia, and in Georgia the ratio is 2.3 economically inactive women per one man. (See more on the pathways to unemployment/inactivity in the next section). Box 1. Youth and Employment: Where do youth work, what types of jobs do they have, and what do they think about working? The Georgian and Armenian labor markets are characterized by the persistence of a low productivity informal segment working in agriculture, and mostly performing subsistence jobs. This, despite the fact that both of these economies are transitioning towards services and away from agriculture (World Bank 2017a, 2018). In addition, as it is the case in many ECA countries, there is sex-segregation by industries and occupations, with most women employed in education and health, reflecting as well sex-segregation patterns by fields of education. When focusing on employed youth (15-29), results show that these patterns in the labor market - misallocation in informal low-productivity jobs in agriculture and sex-segregation by industry and occupation- start at early ages in productive life, right after school, when the school-to-work transition begins. In other words, employed youth in both countries follow the same pattern of employed population in general. In Georgia, about one third of employed young men and women are employed in agriculture. In Armenia the share is lower, with 23 percent of employed youth working in this sector (Table 2). The concentration in agriculture persists among the employed in the 30-40 age bracket, although the share of women working in the sector increases (it even duplicates between females 15-29 and females 30-40, in 11 Armenia). In Georgia, most female youth working in agriculture do so as unpaid family workers (89 percent) and among men, 74 percent are unpaid family workers and 19 percent own-account workers. In Armenia the quality of agriculture-related jobs among youth is also low, over 90 percent of youth (men and women) employed in the sector work as either own-account or unpaid family workers. In both countries, most (over 90 percent) rural youth -men and women- in the agriculture sector perform high-skilled blue-collar occupations (skilled agricultural workers). Among urban workers in the sector, there is sizable share performing elementary occupations (18 percent of urban men in Georgia as well as 17 percent of urban women). Besides agriculture, young women concentrate in health and education, and wholesale and retail, whereas men are more evenly distributed across sectors and participate more in those infrastructure- related (construction, transport, manufacturing). In addition, between 60 and 70 percent of young women in both countries (vs. 35 percent of men in Georgia and 48 percent in Armenia) are employed in white-collar occupations (meaning professionals, technicians, administrative support). Few young people are employed in high productivity sectors such as financial intermediation and telecommunications which indicates a missed opportunity since structural transformation in both countries is expected to increase the demand for IT specialists: in Georgia, for example, up to 90 percent of all jobs will require some knowledge related to ICT in the near future (World Bank 2017a, 2018). In general, there is sizable share of employed youth working as unpaid family workers in Georgia (28 percent, with no major differences by sex), and a sizable share working as own-account in Armenia (24 percent, with many more men than women). More men than women have temporary, seasonal or casual jobs (23 percent of men in Georgia, 15 percent of men in Armenia). The share of youth who declare entrepreneurship (i.e. employer with hired employees) is minimal, around 1 percent in Georgia and 0.3 percent in Armenia (no gender differences in neither country). Misallocation of employment among youth relates to the skills-mismatch phenomenon, which is proper of transition economies. In Armenia/Georgia, only 66 percent of workers’ jobs require education that match their own and some report being over-educated (See more on skills mismatch in Box 2). Qualitative findings from a recent study for Armenia, indicate that professional work presents itself as an ‘unprofitable’ occupation, reason why only a small portion of young people in the country value profession when entering the job market. (Friedrich-Ebert-Sung, 2016 and 2017). Table 2: Distribution of employed youth (15-29) across economic activity, by sex Georgia Armenia Women Men Total Women Men Total Agriculture, forestry, 31.9 35.9 34.3 agriculture, forestry and hunting fishing 17.1 27.7 23.5 Wholesale and retail trade; 14.9 13.2 13.9 wholesale and retail repair of motor vehicles trade; repair of motor Education, health and 14.8 3.7 8.1 vehicles 20.0 11.7 15.0 social work public administration Manufacturing, mining and 4.0 9.5 7.3 and defense; compulsory quarrying social security 3.6 19.0 13.0 Public administration 4.9 8.7 7.2 education, health and social work 24.2 3.3 11.5 Finance, real estate and 9.6 3.5 5.9 manufacturing, mining business and quarrying 6.4 9.7 8.4 Other 9.3 3.5 5.8 transportation and Construction 0.5 8.2 5.1 communications 5.7 7.0 6.5 Transport and 2.6 6.4 4.9 construction 0.2 7.0 4.4 communication accommodation and Hotels and restaurants 4.3 3.4 3.8 food services activities 3.2 4.1 3.8 Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). 12 In terms of the expectations of youth on finding employment, a recent representative survey among people aged 14-29 in Georgia and Armenia find that youth believe that the most important factor for �nding a job is the help from friends or acquaintances, and connections are more important that the role of work experience. At the same time, youth in Armenia for example perceive lack of employment as a ‘very urgent’ problem for Armenian society (86 percent); this perception increases with age and level of education. In Georgia, most youngsters (85 percent) believe that the employment situation in the country is alarming, and four out of ten believe that this issue is the most problematic in the country. (Friedrich- Ebert-Sung, 2016 and 2017). Nevertheless, most youth 18-29 (between 74 and 92 percent) in both countries value work as ‘very important’ in one’s life (see Figure 18 below). The Profile of Male and Female NEETs The shares of youth who are unemployed or inactive in Georgia and Armenia are considerably high.17 Applying the definition of NEET introduced in the Data and Methodology section, the proportion of young people (15-29 age group) not in employment, education, or training amounts to 31 percent in Georgia and 33 percent in Armenia (Table 3). As a reference, among people in the same age group, NEET rates in the Czech Republic, Slovenia, and Lithuania are less than 10 percent, in Romania 18 percent, in Turkey 28 percent, and the proportion in the EU28 is about 13 percent (Eurostat, 2017). High NEET rates indicate that young people in these two countries are not investing in their human capital and are disproportionately at risk of labor market exclusion. The NEET phenomenon is as problematic in urban as it is in rural locations. By age group, the share of NEET in the 15-19 age group is the lowest as this is the time when most youth are likely to be in school. Conversely, NEETs are over-represented in the 20-29 age group: in Georgia, the proportion of youth in this age bracket is 70 percent whereas among NEET, the share increases to 85 percent; in Armenia, these shares are 73 percent and 90 percent, respectively. NEET youth in both countries follow the same regional distribution as all youth do and concentrate in the capital cities. Table 3: NEET Rates among Men and Women, 15-29 age group Georgia Armenia Women Men Total Women Men Total Total 36.6 25.2 30.6 41.7 23.5 33.0 Urban 36.9 30.0 33.4 39.6 27.1 33.8 Rural 36.2 18.1 26.3 44.9 18.6 31.9 Share 56.8 43.2 100 66.0 34.0 100 NEET Rates by age group: 15_19 15.9 14.7 15.2 14.3 11.3 12.8 20_24 41.8 28.7 35.2 45.3 29.6 38.0 25_29 47.5 30.9 38.8 56.6 27.3 42.7 Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). 17Most unemployed youth are youth who are not receiving any type of education or training. The shares of those unemployed who are also in some type of education/training are rather low, between 2 to 3 percent in both countries as Figure A1 in the Annex shows. 13 Women are more likely to be NEET than men, and this disadvantage is more pronounced in rural areas Table 3 and Figure 3 show that NEET rates are consistently higher for females than for males in both countries, and that the gender gaps in NEET rates between women and men are at least twice as high in rural areas than in urban areas (three times as high in Georgia). In addition, the gender gaps (difference between women and men) in NEET rates in Armenia are significantly higher than in Georgia: in urban settings, for instance, the gender gap is almost twice the size in Georgia. Figure 3: Gender Gaps in NEET rates (Difference Female rate – Male rate) Georgia Armenia 30 30 25 25 26.3 Percentage points Percentage points 20 20 18.1 18.3 15 15 10 11.5 12.5 10 6.9 5 5 0 0 All Urban Rural All Urban Rural Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). The gender gap in NEET rates starts at age 20 and increases with age Figure 4 displays the age-specific NEET rate by sex. At age 15 the NEET rate is still below 10 percent in both countries, as most youngsters are still in school. Between ages 16 and 20 there is a sharp increase in the percentage of both males and females who are neither in employment or education.18 Ages 20 and 21 seem to be the ‘inflexion points’ at which significant gender differences in NEET rates appear, with women more likely to be NEET than men. This period coincides with the beginning of the peak productive and reproductive ages. At age 20 the NEET rate is 27 and 38 percent in Georgia and Armenia, respectively. After that age, there is a clear separation between the sexes: while the percentage of males who are neither in employment or education in Georgia drops to 23 percent at age 28, the percentage for females continues to rise to 43 and then 49 percent. The same occurs in Armenia except that the gender differences become even wider as youth become older: after age 20, the female NEET rate goes up to 56 percent at age 29 compared to 24 percent among men. Figure 4: Gender Differences in NEET: Rates by Age and Sex 18In the case of Armenia, the NEET ‘peak’ observed among men age 20 is mostly driven by a significant share that is coming back from compulsory military service that either become unemployed/inactive or go back to school. 14 Georgia 70% 60% 50% 40% 30% 20% 10% 0% 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Women Men Total Armenia 70% 60% 50% 40% 30% 20% 10% 0% 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Women Men Total Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). In Georgia the NEET condition is more urban, while in Armenia it is both urban and rural; yet, gender differences in the likelihood of being NEET persist in both areas The NEET pattern by age in rural areas in Georgia is quite different from the one observed urban areas (Figure 5). Before age 22, the NEET rate is similar in both areas (except between ages 15-16 where is slightly higher in urban areas); after age 22, there is a clear separation in the urban and rural NEET rates that remains consistently higher in urban areas and reaching a peak at age 25 (51 percent in urban vs. 34 percent in rural). This is mostly because rural youth find it easier to become employed by getting involved in agricultural work (see Box 1). In Armenia, on the contrary, the trends in NEET rates by location are relatively similar: NEET rates are practically the same between ages 15-17, then the rate in rural areas becomes higher than in urban areas between ages 17-21 (due to inequalities in access to education), to then go back to similar levels as in urban; after age 27, the NEET phenomenon becomes more urban. Figure 5: NEET Rate by Age and Location 15 Georgia Armenia 60 60 50 50 40 40 30 30 20 20 10 10 0 0 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Urban Rural Urban Rural Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). More information is revealed when breaking down NEET rates by gender and by location (Figure 6). In Georgia, after age 18 the NEET rate among rural men experiences a significant drop (as most young men are likely to join the agricultural sector) while NEET rates among urban men increase (mostly, due to lack of available jobs and an increase in unemployment). Because of this drop, the gender gap in NEET rates in rural areas starts widening at an earlier age (age 18) than in urban areas (age 23) and continues increasing even more (compared to urban) with age. In Armenia, urban men experience a higher likelihood to become NEET than rural men particularly after age 22, but the difference is smaller compared to Georgia; in addition, contrary to Georgia, gender gaps in rural and urban areas are both wide. Figure 6: Gender Difference in NEET Rates by Age and Location Georgia Armenia 60 80 70 50 60 40 50 30 40 20 30 10 20 10 0 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 0 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Women-Urban Men-Urban Women-Urban Men-Urban Women-Rural Men-Rural Women-Rural Men-Rural Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). NEETs are heterogeneous: Inactivity is more prevalent among NEET women whereas unemployment is more the condition among NEET men When breaking down female and male NEET populations by the reason for their non-employment (Figure 7), the distribution shows that most female NEET are inactive (74 percent in both countries) than are unemployed (i.e. job seekers). Moreover, inactivity among female NEETs is mostly driven by women who report being in charge of household and caregiving responsibilities. Men, on the other hand, are mainly 16 unemployed (54 percent of male NEET in Georgia and 58 percent in Armenia) and less than half of them are inactive. Male NEETs are also more likely than their female counterparts to be ‘unavailable’ (in most cases because they are back from compulsory military service and because of migration patterns i.e. returning or about to leave the country). Finally, ‘discouragement’ constitutes a non-negligible share of NEET s in both countries (between 4 and 12 percent) and it is more prevalent among NEET men in Georgia whereas in Armenia it is more prevalent among urban men and among rural women. In any case, figures indicate that NEETs comprise populations in very different statuses and with pronounced gender differences. Furthermore, among those NEET who are ‘discouraged’, the perception that that ‘there are not enough jobs’ is the most prevalent reason among both men and women in Georgia (88 and 81 percent of discouraged women and men, respectively) along with the perception of ‘lack of qualifications and experience’ in the case of discouraged men (12 percent men, 5 percent women); the later becomes more prevalent among the discouraged men and women in rural areas (18 and 10 percent, respectively). In Armenia, most discouraged women claim they are ‘lack of jobs in their area of expertise’ (63 percent) followed by ‘past failure to find a job’ (27 percent), whereas among men the shares are less divided and half of them perceive there is lack of suitable jobs and 36 percent have experienced a failure in the past finding a job. In urban areas, the shares are more evenly distributed between these two reasons among men and women. The gender differences in status in the labor market -and more specifically, in the status of those who are NEET - may be explained by the gender differences observed in the pathways in school-to-work transition (see next section) and the typical gender roles steering women towards taking care of a household, children or other relatives. Moreover, as figures of the condition of female and male NEET show, women are more likely to be responsible for the provision of care than men (the shares of male NEET reporting homemaking roles are in fact negligible) and may therefore opt for either not entering the labor market or postponing entry until a later stage in their lives (which is also the case in most ECA countries). (ETF, 2015) Figure 7: Heterogeneity of NEETs: Distribution of NEET across Sub-Group, by Sex and Location (%) Georgia Armenia 100 100 16 17 14 2 3 6 1 80 12 4 25 4 26 3 80 13 21 23 14 8 16 60 11 9 60 11 52 1 1 15 50 46 54 18 58 3 40 40 56 1 58 66 54 57 20 45 20 35 42 26 30 26 19 14 0 0 Women Men Women Men Women Men Women Men Women Men Women Men All Urban Rural All Urban Rural Job seeker Homemaker/caregiver Job seeker Homemaker/caregiver Discouraged Unavailable Discouraged Unavailable Disabled Other Illness Other Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). Note: For Georgia, ‘Unavailable’ refers to those unwilling to work and those about to go or have just returned from compulsory military service, and to those unwilling to work, who are about to go abroad in the case of Armenia. ‘Other’ 17 refers to pensioners and others in Georgia, and to those waiting for the work season to resume, those who have just returned from abroad, those about to go or have just returned from compulsory military service, and others in Armenia. Educational attainment does not necessarily protect men and women against being NEET Georgia and Armenia are among the cases where graduates of upper secondary and TVET (non-tertiary) perform worse in the labor market (i.e. succeed less in finding employment) than those with lower and higher levels of education, suggesting that more education is not a guarantee of fewer NEETs (Figure 8). In both countries there is high upper secondary and tertiary enrollment that is in line with the EU28 averages (Figure 9, below) but, at the same time, NEET rates are high. NEET rates are the highest among those with vocational, upper secondary and even education and this is particularly true among women. Figure 8: NEET Rate by Educational attainment, and Sex of youth. Georgia Armenia 70 70 60 60 50 50 <= Lower Sec <= Inc. Sec 40 40 UpperSec Secondary 30 30 Vocational Vocational 20 Tertiary 20 Tertiary 10 10 0 0 Women Men Women Men Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). Note: Refers to the share of NEET among people with that education level, that sex and that age (15-29). In Armenia there are too few cases of ‘Incomplete secondary or less’ to be reliable. The fact that education does not necessarily lowers the numbers of NEETs suggests shortcomings with respect to the quality and relevance of the education programs provided vis-à-vis the needs of the labor market. Another explanation might be a secondary education system that is dominated by general and academic programs and with a limited vocational provision (in Georgia, 2.1 percent of students are engaged in vocational programs) (ETF, 2015). A third explanation is the key role that the skills mismatch plays for job placement in both countries, which goes beyond educational attainment and diploma ownership (see more on skills mismatch in Box 2 and the Regression Analysis section). Out-migration trends of highly educated and qualified youth and the subsequent ‘brain drain’ in both countries, can also play a role in explaining the puzzle. In addition, part of the reason why NEET rates are so low for those with lower (or incomplete) secondary is that these individuals are most likely still attending school and are therefore not NEET (see Table 4). On the contrary, for those with tertiary (or secondary, vocational), higher NEET rates do reflect lack of employment opportunities and persistent unemployment and economic inactivity (particularly for women). Table 4: Status of Youth (15-29) by Educational Attainment and Sex Georgia Women Men 18 Lower Sec Upper Vocational Tertiary Lower Sec Upper Vocational Tertiary or less Sec or less Sec Employed 18.7 29.7 39.5 61.2 24.3 52.8 66.8 69.4 Unemployed 4.2 13.3 16.1 15.7 6.3 18.3 19.0 22.0 Inactive (not in edu) 15.9 33.0 43.9 21.5 9.9 14.1 13.7 7.1 In education/training 61.3 24.0 0.5 1.6 59.5 14.8 0.5 1.4 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Armenia Women Men Incomplete Sec Vocational Tertiary Incomplete Sec. Vocational Tertiary Sec or less Sec or less Employed 7.2 14.6 34.8 51.0 13.1 44.3 57.4 68.7 Unemployed 2.4 10.7 18.7 21.9 5.7 15.5 27.7 22.9 Inactive (not in edu) 8.1 38.6 40.9 22.6 9.0 11.8 13.8 3.8 In education/training 82.3 36.1 5.6 4.4 72.3 28.4 1.1 4.6 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). Note: In Armenia there are too few cases of ‘Incomplete secondary or less’ to be reliable. As noted before, in both countries more females than males attain tertiary education (a phenomenon that is very common in ECA countries) and the shares are in line to the EU28 averages of tertiary attainment (Figure 9). Yet, even in the presence of this advantage towards women, the gender gaps in NEET rates are still among the highest (after Turkey) relative to other ECA (and EU) comparators for whom tertiary attainment is also in line with the EU28 average. As a reference, countries such as Estonia, Slovenia, and Lithuania the gender gap in NEET rates ranges between 5 and 1 pp and the gap in the EU28 is low as well (only 4pp). In contrast, the gap in Georgia is 12pp and even higher in Armenia 18pp (Eurostat, 2018). Figure 9: Gender Gap in NEET Rates and Tertiary Attainment among Youth 15-29. Armenia, Georgia and selected ECA (and EU) comparators. 30 28 25 24 25 20 18 18 18 15 10 % 5 0 % Women 15-29 with tertiary attainment % Men 15-29 with tertiary attainment Gender Gap in NEET rates (Female - Male), 15-29 age group Source: Authors’ calculations based on Georgia LFS (2017), Armenia LFS (2017) and Eurostat (2018) for all other countries. Note: Countries are arranged in descending order according to the size of the gender gap in NEET rates. 19 Nevertheless, when considering all youth (NEET and non-NEET) and their educational attainment by status in the labor force, tertiary education does seem to be a protective factor against inactivity (Figure 10). According to a recent study led by UN Women19 in the South Caucasus, the factor that has the largest, positive impact on women’s labor force participation in Armenia, Azerbaijan, and Georgia is education. In these three countries, compared to women who only have secondary education, women with tertiary are 18, 30, and 20 percentage points, respectively, more likely to participate in the labor force. (UN Women, 2018). Still, tertiary is not itself a guarantee of employment; in Armenia for instance, 43 percent of unemployed young women have tertiary and so do 20 percent of the inactive ones. Figure 10: Educational attainment (distribution) by Status in the Labor Market, and Sex of Youth Georgia Armenia 100 100 80 80 60 60 40 40 20 20 0 0 Inactive Inactive Employed Employed Unemployed Unemployed Employed Inactive Employed Inactive Unemployed Unemployed Women Men Women Men LowerSec or less UpperSec Vocational Higher Inc. Sec or less Secondary Vocational Tertiary Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). Note: ‘Inactive’ excludes those in education and/or training. Both male and female NEETs belong to poorer households The NEET condition is clearly associated with household welfare. This finding comes as no surprise because youth from poorer families have lower chances of completing their education and tend to lack the networks needed to find decent jobs. At the same time, because of their NEET condition, these are household members who are out of employment, do not perceive any labor income, and therefore cannot contribute income-wise to their household (compared to their working peers). The average NEET -man or woman- lives in a household of a lower income quintile compared to their peers who are in school, employed, or active in the labor force and studying at the same time (Figure 11). Moreover, this is true for both types of NEETs i.e. the unemployed and the economically inactive. Similarly, the share of youth living in households in the bottom 40 percent of consumption aggregate is the highest among the NEET populations (job seekers or economically inactive). In Georgia for instance, 63 percent of inactive NEET women live in bottom 40 percent households (whereas the share among their employed 19This study was commissioned by UN Women, carried out by the Caucasus Research Resource Center for UN Women’s project “Women’s Economic Empowerment in the South Caucasus�, and funded by the Swiss Agency for Development and Cooperation for the South Caucasus. 20 peers is 43 percent) and so do 63 percent of unemployed NEET men. In Armenia, the shares are slightly lower, but still more than half of female and male NEETs live in a bottom 40 household. Figure 11: Association between the NEET Condition and Welfare of the Household A. Average Income Quintile of Household where B. Share of Youth (15-29) Living in a Household in the Youth (15-29) Live, by Sex and Status in the Labor Bottom 40% of Consumption Aggregate, according to Force their Status in the Labor Force Georgia 5 100 4 80 3 60 2 40 1 20 0 0 All Women Men Women Men In school only Work only Work+school In school only Work only Work+school Unemploy+school Unemployed Inactive Unemploy+school Unemployed (NEET) Inactive (NEET) Armenia 5 100.0 4 80.0 3 60.0 2 40.0 1 0 20.0 All Women Men 0.0 Women Men In school only Work only Work+school Unemploy+school In school only Work only Unemployed Inactive Work+school Unemploy+school Source: Authors’ calculations based on Georgia HIS (2016) and Armenia ILCS (2017). Previous work experience among NEET is scarce, particularly for women In Georgia about 42 percent of the NEET population report having ever performed a paid activity or owned a business, and the shares of those with previous work experience are higher among men (49 percent) than women (37 percent). In Armenia, the share of NEET with previous history of work is much lower, only 25 percent, with 33 percent men and only 21 percent women, but the gender gap is similar than in Georgia. Furthermore, as seen in Georgia, these experiences were characterized for being very short-term posts lasting a year or less. Data from Armenia shows that, among those who report previous work, the end of the post was rather recent, within the past 12 months or less. More noticeable, there are important gender differences in the reasons for leaving that last job: among female NEET, the most mentioned reasons are ‘family responsibilities’ and ‘caregiving’, whereas for men the most common reason relates to seasonality as these were jobs ‘of limited duration’ or to ‘staff reduction or dismissal’. (Figures 12 and 13) 21 Figure 12: History of previous employment among NEETs in Georgia, by Sex Duration of previous work Reason for leaving the last job 80 Other reasons Did not want working Men Women 60 Working conditions were not satisfying Compulsory military service 40 Education or training 20 Own illness or disability Other personal or family responsibilities 0 Looking after children or other relative Women Men Total A job of limited duration has ended One year or less Dismissed One year to up to 3 years More than 3 years 0 10 20 30 40 50 Source: Authors’ calculations based on Georgia LFS (2017). Figure 13: History of previous employment among NEETs in Armenia, by Sex Length of time since previous employment Reason for leaving the last job ended Staff reduction / dismissed A job of limited duration has ended Over 3 years Illness/Injury 2 to 3 years Care of sick family member 1 to 2 years Household chores / family ccircumstances Low wage 3 to 12 months Quitted job on my own Up to 3 months Going abroad Other reason 0 10 20 30 40 0 10 20 30 40 50 60 Women Men Women Men Source: Authors’ calculations based on Armenia LFS (2017). NEET rates are higher for those men and women from a minority group or with migration background In Georgia, the highest NEET rates are found among those with Armenian background (38 percent) followed by Georgians (30 percent). In Armenia, youth with Georgian background have the highest rates (44 percent), followed by Armenians (33 percent) and ‘Other’ (29 percent). When it comes to disabled youth, 64 percent of them are NEET in Georgia, with similar rates for men and women (62 percent and 65 percent, respectively) and 61 percent in Armenia with higher rates among disabled women (75 percent) than men (54 percent). Finally, in Armenia, half of youth who report being returned migrants (after 2013) are NEET and the rates are higher among men than among women (53 vs. 37 percent). 22 Box 2. New labor market entrants are not well equipped in terms of their skills… Skills deficiencies have been documented to be a structural issue in the South Caucasus as they are typical of transition economies such as Armenia and Georgia that undergo economic restructuring and structural transformation. The transition has induced a growing demand for new skills (not only measured by educational attainment, but also by other proxies that capture cognitive and non-cognitive skills), but the supply of skills has not kept up, despite the fast expansion in coverage of tertiary education. (Posadas et al., 2018; World Bank 2015, 2017a, 2018) Results from the 2015 round of the Program for International Student Assessment (PISA), indicate that poor quality of education remains a major issue. Results for Georgia (Armenia does not participate from this assessment) reveal that more than half of Georgian 15-year-old students scored below basic proficiency in core skills (Figure 14), and scores are well below OECD and ECA averages. Moreover, results show sharp differences in learning outcomes by socioeconomic status, urban/rural location and gender (World Bank, 2018). Based on the STEP Employment Survey, Handel, Valerio and Sanchez Puerta (2016) conclude that education mismatch is high in ECA’s transition economies. In Armenia/Georgia, only 66 percent of workers’ jobs require education that match their own and between 28-29 percent report being over-educated (Figure 15). The rates of over-education are higher than in other comparator countries. Over-education is also reported by 22 percent of young (15-29) workers in the Georgian LFS (no differences by sex) and by 16 and 20 percent of young women and men, respectively, in the Armenia LFS, and the figures are very similar among those in the 30-40 age group. Employers, on the other hand, report that lack of skills as a major obstacle in their business activities. In Georgia, occupation-specific technical skills are considered critical for employability, and at the same time a high proportion of firms say that young workers frequently lack them. Young workers also often lack cognitive skills (problem solving) and some key socio-behavioral skills (ability to work independently and in teams) (Rutkowski, 2013). In Armenia, the skills that are reported as frequently missing among younger workers are occupation-specific technical skills, problem solving, teamwork, time management, communication and ability to work independently (Rutkowski, 2013a). Furthermore, due to structural transformation in both countries, it is estimated that demand for IT specialists will continue to increase. In Georgia, for example, up to 90 percent of all jobs will require some knowledge related to ICT in the near future. However, one in five firms sees the lack of workers with solid digital skills as a major or severe problem. (World Bank 2017a, 2018) Figure 14. Performance of Georgian Youth (15-year- Figure 15. Education mismatch old) in PISA 2015 PISA (2015) Scores reveal quality concerns Individual match rates between worker education and job required education Reading FYR Macedonia 72.6 22.3 Math Ukraine 72.1 24 Science Armenia 66.2 28 Georgia 66.4 29.4 0% 50% 100% 0 20 40 60 80 100 Below basic proficiency (=Level 5) Source: OECD. Note: Functional literacy is defined as a score Source: Handel, Valerio and Sanchez Puerta (2016). above Level 2 in Reading. Estimates based on STEPS (Skills Toward Employment and Productivity) household surveys of working-age adults (ages 15-64) residing in urban areas. 23 Gender Differences in the Pathways to Becoming NEET While most young men transition from school to work, most young women remain economically inactive and out of the labor force for longer periods Figure 16 shows a simple characterization of the average school-to-work transition pathways for 15-year- olds (and up to age 40) in Georgia and Armenia. As graphs show, youth can be in school, working, doing both, unemployed but studying, or doing neither (i.e. being NEETs). Adults 30-40 are either economically active or inactive. The key findings from these pathways is that in Georgia and Armenia, the transition is very different for males and females: while most young men transition from school to work, most young women remain inactive and out of the labor force for longer periods as they stay home. A salient feature in the two countries is that boys who finish (or leave) high school -at age 18 - stay inactive for a relatively shorter period and within 1-2 years begin working. For those girls who complete high school and then leave school, the story is different because they will stay inactive for longer. The peak of inactivity among women takes place between ages 20 and 35 years old, which coincide with the peak family formation and childrearing age, then, in their late 30s and 40s, women re-enter the labor force (to be either employed or unemployed) but with an important share still inactive. Inactivity is lower among men and stabilizes after age 25, when most of them are either employed or unemployed. These results confirm that young women either do not to enter the labor force or leave the labor force after finishing education to take over household duties such as caring for children or elderly family members. Moreover, because women tend to leave the labor force, unemployment mainly affects young men and men close to retirement age (Posadas et al., 2018). In Armenia, the transition patterns confirm the share of women that transition into inactivity is higher than in Georgia. Young men tend to leave education earlier, typically between ages 17 and 18, when they are more likely to join the labor market -mostly as employed. In Armenia, due to compulsory military service, which applies to men once they reach age 18 (it covers men from 18 to 27 years old), an important share of male youth between the ages 18 and 20 will abandon the labor force or make a pause in their studies20 and start coming back at age 21 when they will join school (about a fifth of male youth age 21) or go into employment (about half of male youth age 21). 20The Armenia LFS records young men who are carrying out compulsory military service as ‘inactive’ population. Following Eurostat’s guidelines, these observations have been excluded from the calculations of labor force indicators presented in thi s analysis as these are individuals that are not eligible to participate in the labor force. This adjustment reflects on the following shares of dropped observations per age group: 37 percent of men age 18; 63 percent of men age 19; and 26 percent of men age 20. If considering this category as part of the sample, men in the 18-20 age group will be either in compulsory military service (42.1 percent) or enrolled in school (28 percent). See https://ec.europa.eu/eurostat/web/microdata/european-union-labour- force-survey. 24 Women, on the contrary, remain slightly longer in education typically until the age of 21, when they complete tertiary (this is consistent with young women having higher educational attainment than young men). After that, some of these women enter the labor market, with the majority finding employment (specially at age 22 and older), and a high share will remain economically inactive (reaching a highest share of 40 percent among women age 29 in Georgia, and 53 percent at age 27 in Armenia) until about age 35 when inactivity starts to slightly fall. Figure 16: Pathways into becoming NEET: School to Work Transition among Men and Women Georgia Men Women 100 100 80 80 60 60 40 40 20 20 0 0 15 17 19 21 23 25 27 29 31 33 35 37 39 15 17 19 21 23 25 27 29 31 33 35 37 39 In education (or training) Work only In education (or training) Work only Work+edu/train. Unemp+edu/train. Work+edu/train. Unemp+edu/train. Unemployed (NEET) Inactive (NEET) Unemployed (NEET) Inactive (NEET) Armenia Men Women 100 100 80 80 60 60 40 40 20 20 0 0 15 17 19 21 23 25 27 29 31 33 35 37 39 15 17 19 21 23 25 27 29 31 33 35 37 39 In education (or training) Work only In education (or training) Work only Work+edu/train Unemp+edu/train Work+edu/train Unemp+edu/train Unemployed (NEET) Inactive (NEET) Unemployed (NEET) Inactive (NEET) Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). Note: ‘In education or training’ refers to inactive people who are receiving any type of formal education or non- formal education or vocational training, including full or part-time formats. By location (see Figure A2 Annex), the gender differences in the transition patterns hold but rural women leave school earlier at age 19 (vs. the average age of 21) and inactivity is more pronounced among those who are between 20 and 30 years old (after which more rural women will be likely join employment 25 through the agriculture sector); as for men, by age 19 most rural men will have abandoned school, be inactive for one year to then find a job (few will remain inactive). Around 20 percent of rural youth (this is particularly true in Georgia, and for rural men in Armenia) who are between 15 and 18 years old will have a job while they finish secondary school. Family formation plays a role: compared to the average of young women (15-29), female NEET marry at higher rates Family formation helps explaining why many more young women than men fall into inactivity after exiting school. While neither survey allows to identify the age at first child, they do allow to calculate marital status by age. The transition from singleness to being married (or in union) gives a clue of when family formation (and therefore the role of caregivers) starts which can in turn influence inactivity. Figure 17 shows that NEET women are married in higher shares than young women (15-29) in general: while 30 percent of NEET women age 18 in Georgia are married or in union, the share goes down to 13 percent among all young women 15-29. In Armenia, while only 3 percent of all women age 17 are married, when considering only female NEET that same age, the share increases to 28 percent. The difference in marriage rates between NEET women and young women in general in both countries continues up until age 27 when the gap starts to narrow. Among men, marriage rates among NEET and among male youth in general are rather similar. In addition, when comparing NEET women by urban/rural location, it appears that rural NEET women start marrying at earlier ages and at higher shares, than their urban counterparts; also, urban NEET women seem to postpone marriage for more years than their rural counterparts. (See Figures A3 in Annex) A low mean age of women at the birth of their first child, which is the also the case in these countries, suggests that women are more likely to be responsible for the provision of care (for children but also for elderly) than men, and may therefore opt for either not entering the labor market or postponing entry until a later stage in their lives (ETF, 2015). Women’s early marriage, an issue in all three South Caucasian countries, is associated with lower levels of women’s economic engagement as well (UN Women, 2018). Figure 17: Family Formation and the NEET Condition: Share of Youth and NEET Youth who are Married, by Age Georgia Armenia 100% 100% 80% 80% 60% 60% 40% 40% 20% 20% 0% 0% 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Women NEET Men NEET Women NEET Men NEET Women Men Women Men Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). Note: Married category includes married, ever married (divorced, separated, widowed) or in union. In the case of Armenia there are too few observations of female NEETs ages 15 and 16 years reporting on their marital status, and the same is true for Georgia among female NEETs ages 15 to 18. Age figures for these data points are therefore omitted. 26 Social norms on care, work and motherhood, and the typical gendered roles that steer women towards household work, help shaping the gender differences in the transition pathways Research on female economic inactivity (covering women aged 18+) conducted by UN Women in the South Caucasus finds that, along with having children and marriage, the experience of having ever left a job out (or not seeking one) for family-related reasons, are strong predictors of labor force participation. In Georgia for example, women who have ever left a job (or not sought one) due to family-related reasons, are 18 percentage points less likely to participate in the labor force. This is particularly relevant for female NEET because, as these findings suggest, once they are out of the labor force (for the first time) it will be harder for them to ever join.21 (UN Women, 2018) While it is clear that family duties are the reason why many (young) women are out of employment (as data from Figure 7 shows), it is less clear how much it is due to social norms and typical gendered roles, how much of that it due to lack of quality/affordable childcare and/or potential legal barriers women face, and how much to (relative) lower wages (Posadas et al., 2018). Social norms on caring for family members, work, and motherhood are particularly relevant for young female NEETs; these norms, and the typical gendered roles attached to them, play a role in shaping the observed gender differences in the school-to-work transition pathways. Data from the most recent wave of the World Value Survey show that, while over 73 percent of male and female youth in both countries consider that work is ‘very important’ in life (Figure 18), there is also a significant share who agree with social constructs steering women towards household work and men towards the labor market (Table 5). For instance, in Armenia 78 percent of men and 44 percent of women in the 24-29 age group think that when jobs are scarce men should have more right to jobs than women; for Georgia, the shares are 46 and 32 percent, for men and women in the same age group respectively. More than half of Armenian men and women age 24-29 agree that when a mother works for pay, the children suffer, and the shares are even higher for Georgia (over 76 percent of both men and women). According to LiTS (2016), most youth in both countries agree that it is better for the household if the man earns money and the woman takes care of the home and the children. Except for a few cases, agreement with these perceptions increases among the oldest youth (24-29) compared to the youngest ones (18-23). On the other hand, in Armenia, barely half of youth consider that a job is the best way for a woman to be an independent person; the same occurs among male youth in Georgia and females in the 18-23 age group (but the share increases among females 24-29). Figure 18: Perception of the Importance of Work in one’s Life (% Youth in Age Group) 21According to the study, in Armenia, married women are 16 percentage points less likely to participate in the labor force. In Azerbaijan, a woman’s chances of being in the labor force is 12 percentage points lower when married. In Georgia, if a woman has ever left a job or not sought one out for reasons related to family, she is 18 percentage points less likely to be in the labor force. 27 How important is work in your life? Men 74.4 18_23 Armenia Women 74.4 Men 80.6 24_29 Women 73.3 Men 82.9 18_23 Georgia Women 74.2 Men 91.5 24_29 Women 80.2 0 10 20 30 40 50 60 70 80 90 100 Very important Rather important Not very important Not at all important Source: World Values Survey, Wave 6 (2010-2014). Note: Samples are representative of all people age 18+ living in private households in each country, regardless of their nationality, citizenship or language. Table 5: Percentage of People in Age Group who 'Strongly agree' or 'Agree' with the Statement Armenia Georgia 18_23 24_29 18_23 24_29 Men Women Men Women Men Women Men Women When jobs are scarce, men should have more right to a job than 63.2 42.0 77.8 43.7 43.9 25.8 45.8 31.9 women* Having a job is the best way for a woman to be an independent 40.1 48.5 48.3 52.8 56.1 56.1 50.9 85.7 person* When a mother works for pay, the 45.7 40.0 58.2 53.2 45.6 62.1 79.7 75.8 children suffer** It is better for everyone involved if the man earns the money and the 85.0 63.6 92.0 70.6 56.6 50.7 81.1 54.4 woman takes care of the home and children** Source: World Values Survey, Wave 6 (2010-2014) except for the bottom one which comes from LiTS III (2016). Note: Samples are representative of all people age 18+ living in private households in each country, regardless of their nationality, citizenship or language. (*) Other categories include 'Neither' and 'Disagree'; (**) Other categories include 'Disagree' and 'Strongly Disagree'. Potential legal barriers that women face can also reinforce the role of social norms in deterring young women (and women in general) from joining the labor force. While both Georgia and Armenia have provisions for paid maternity leave, there is no paid parental leave.22 Both countries allow women to work 22 In Georgia, the law mandates 730 days of maternity leave (approx. 24 months) during pregnancy and childcare. Of these, 183 days (6 months) are paid leave which is fully covered by the government. There are no legal mandates in terms of paid or unpaid paternity leave. In Armenia, the length of paid maternity leave is 140 days (7 months approx.), fully covered by the government. The law grants 60 days (3 months) of unpaid paternity leave and 1,025 days of unpaid parental leave. (WBL, 2018) 28 in same sectors/tasks and hours as men and the law mandate nondiscrimination based on gender in employment, however there are no legal provisions for equal pay for work of equal value. In Georgia, flex work arrangements are not available, and mothers are not guaranteed by law an equivalent position after maternity leave. Payments for childcare are not tax deductible in either of the countries, something that could be an incentive to free up women’s time for other activities. (WBL, 2018). Regarding childcare, while both countries offer a public system of free child care services, capacity and quality concerns 23 seem to move families into choosing more expensive, paid, private child care arrangements. According to a recent assessment of childcare services in Georgia (World Bank, 2019), the focus of the supply is primarily on preschool education and school readiness services that involve children ages 3+. This implies a gap in capacity and provision of childcare services for the 0–2 age group and a constraint for women who make use of the paid maternity benefit (6 months) and need support during the early months after birth. Finally, the fact that women earn significantly less than men seems to be another strong deterrent for them not working (UN Women, 2018).24 Yet, given that working and non-working women do on average the same amount of unpaid care work, reducing the gender pay gap would not necessarily lead women into the labor market as they still face the barrier of persistent traditional care roles. (Posadas et al., 2018) What Characteristics are correlated to Being NEET? A Regression Analysis Conceptual Framework The analysis presented in this section is based in the conceptual framework proposed by Eurofound (2012) and ETF (2015) to understand the individual characteristics, family circumstances and country-wide situations that affect young people’s situation in the EU and neighbor countries.25 (Figure 19) Empirical studies carried out for EU and OECD countries find that young people’s situation is affected by individual characteristics such as gender, education and health, as well as by their family circumstances (e.g. their socio-economic or migration background). Other influential elements linked to their countries’ education systems (e.g. school systems not aligned to the labor market demand), the labor market conditions where young people live, and/or their perception of lack of suitable opportunities or qualification deficits also relate to being NEET (Eurofound, 2012). With regards to gender, the higher incidence of the NEET phenomenon among women relates to ‘established’ roles towards family responsibilities. In many countries, childcare and household maintenance are almost exclusively the 23 In 2013, the Georgian Parliament, in the context of a broader reform promoted by the government to increase state social expenditures, abolished parental fees to public early and preschool education establishments. As a result, an increased demand for children’s enrollment into preschools could be observed, which, in turn, resulted in excessive number of children in limited physical space and infrastructure of public kindergartens (UNESCO 2015; World Bank, 2019). 24 Among the reasons for not working cited by women 18+ who participated in the UN Women study, are: available jobs do not pay enough; that they have left a job because it did not pay enough; or that they never sought out a job in the first place because the available jobs did not pay well. 25 EUROFOUND and the ETF are EU agencies supporting countries in and around the EU region to reform their education, training and labor market systems. They have produced most of the diagnostic studies on the NEET phenomenon in the region. 29 domain of women, which prevents them from participating in economic life and blocks their educational career (Furlong 2007, ETF 2015, Carcillo et al. 2015). Figure 19. Conceptual Framework - Risk factors associated to the probability of being NEET Source: Adapted from Eurofound (2012) and ETF (2015). This section explores the degree to which these risk factors (whenever available in the datasets) are correlated with the probability of being a NEET in Georgia and Armenia, and whether the correlation (size and direction) differs by gender. Two probability models were estimated, a probit (marginal effects at the means of the explanatory variables) and a linear probability model (LPM). Regressions were estimated separately for men and women. The dependent variable is a dummy indicating whether someone in the 15-29 age group is a NEET. Independent variables include individual characteristics such as age, education, marital status, area of residence, history of migration (Armenia only), other vulnerabilities (IPD status for Georgia only, and disability), and household characteristics such as household income and household composition, controlling by region. Results focus on the probit correlates (Figures 20 and 21, and Tables A3 and A4 in the Annex), LPM results confirm various of the probit results and are presented in the Annex (Figures A4 and A5). Household composition and marital status stand out as important correlates of being NEET for women but not men; for all other correlates, the direction of the effect is the same but the size is not Regression analysis confirms that the direction in which most of the correlates are associated with the probability of being NEET -for men and for women- is the same, except for marital status and presence of children in the household which are strong correlates in the case of female NEET, but not for male ones. Yet, the size of the effect of the individual and household characteristics is very different between men and women. In both countries, age is positively associated to the probability of being NEET, but the positive effect decreases as youth become older (as LPM estimates indicate). The size of the effect is small: in Georgia, 30 and additional year increases the likelihood that men are NEET in 2 percentage points (pp) and in 1pp among women. In Armenia, the increase in the probability is 1.1pp and 1.8pp for men and women, respectively. In Georgia, ethnicity is a strong correlate to the probability of being NEET, particularly for men. Armenian men and Azeri men are 21pp and 7 pp more likely to be NEET than male Georgians. Among women, Armenian background is the only one that seems to matter with Armenian women being 14pp more likely to be NEET than Georgian ones. In Armenia, only men from ‘other’ ethnicities (different than CIS26) are 11pp more likely to be NEET than their Armenian counterparts. No significant effect from ethnicity was found among women in Armenia. Holding everything else constant, women in rural areas/men in rural areas in Georgia are 5pp/7pp less likely to be a NEET compared to those women/men living in urban areas which confirms the that the NEET phenomenon in Georgia is particularly urban and that women are more at a disadvantage (inactivity-wise) than men in rural locations. Location is not statically significant for Armenia. Being a male IDP (only available for Georgia) is associated with a lower likelihood of being NEET in 13pp compared to those men who are not; among women, being and IDP is not significant. In the case of Armenia, migration is an important correlate of the probability of being NEET for young men only which is expected as they are the ones more prone to migrate. A male returned migrant is 27pp more likely to fall under the NEET condition than a male who did not leave the country. Having a disability is strongly correlated to being NEET in both countries, and the size of the correlation is bigger among women. In Armenia, young women who have a disability are 63pp more likely to fall under the NEET category than women without any disability, in Georgia the size of the change is 37pp. Among men, the likelihood is 31pp and 39pp in Georgia and Armenia, respectively. As noted in a previous section (see Figure 8), Georgia and Armenia are cases where educational attainment does not guarantee fewer NEETs: NEET rates are higher among those with upper secondary, TVET and tertiary relative to those with lowest levels of education. The regression analysis results reflect this as well: education is not necessarily associated with a lower likelihood of becoming a NEET. Probit models indicate a positive correlation between education and the probability of being NEET in both countries. Furthermore, while the correlations are statistically significant for both sexes, the magnitudes are consistently higher among women (confirming that educational attainment is less rewarded in terms of labor market attachment among women than men). 27 LPM models, on the other hand, are not conclusive: tertiary education is significantly correlated to a lower likelihood of being NEET among women but in Armenia only. It is important to note that there are additional factors not captured in the education variable that might explain this ‘apparent’ disconnection between educational attainment and the NEET condition. Educational attainment and diplomas do not necessarily reflect the skills dimension which is equally (if not more) important for youth to find employment. As discussed in Box 2, Georgia and Armenia exhibit skill deficiencies that are proper of transition economies and employers report that the lack of skills 26 The category CIS, employed in the Armenia ILCS (2017) refers people from to Ukraine, Belarus, Moldova, Kazakhstan, Uzbekistan, Kyrgyzstan, Tajikistan, Azerbaijan, Turkmenistan. 27 Tertiary education (compared to those with less than upper secondary in Georgia, and less than secondary in Armenia), for example, is correlated with an increased likelihood of being NEET of 8pp among women and 5pp among men in Georgia. Vocational education is even more correlated: 24pp for women in Georgia and 27 pp for women in Armenia, and among men, 11pp for Georgia and 17pp for Armenia. 31 (occupation-specific technical skills but also cognitive and behavioral) is a major obstacle for their business. In addition, the school-to-work transition is different for rural and urban youth. While youth in rural areas have lower educational attainment, their chances to be employed (i.e. non- NEET) are higher because these are areas intensive in labor in the agriculture sector; on the contrary, urban youth who are more educated than their rural peers, find it harder (or can afford to wait longer) to get a job (because there are not enough jobs in cities or because these do not match their qualifications). As mentioned before, out-migration and brain drain (specially in Armenia) can also play a role. Marriage (including ever married or in union) is a strong correlate of the probability that a young woman falls under the NEET condition, so does household composition (i.e. presence of children). Married women in Georgia are 32pp more likely to be NEETs, and the increase is 37pp for married women in Armenia. Among men, being married (compared to being single) is associated with a decrease in the likelihood of being a NEET of 7pp in Georgia and 9pp in Armenia. Furthermore, LPM estimations indicate that, in both countries, being married is positively associated with young women becoming NEET regardless of the age group these women belong to (reference category are women in the 15-19 age group who are single).28 In Georgia, young women who live in households with presence of children 7-years-old or less are 6pp more likely to become NEET than those who live in households without children (in Armenia the effect is significant under the LPM model only), confirming that women in general are more likely to stay home to perform the caregiving role. The presence of seniors 65+ in the household was not found significant in neither country. Household size increases the likelihood of being NEET among both men and women in Georgia, and the effect is not significant in Armenia. Regression results confirm that household income is one other important correlate of the NEET condition. The association is a two-sided one: for a non-NEET, his/her household is more likely to have more earners and therefore higher aggregate welfare; at the same time, youth in wealthier households have a higher likelihood of being non-NEET because they can afford to attend schools and/or have better networks in the labor market. In Georgia for instance, young women living in households with at least one source of income29 are 6pp less likely to be NEET and the size of the correlation is bigger (12pp less likely to be NEET) among men living in the same type of households (compared to those who do not). In Armenia, the higher the household income (welfare quintile) the lower the likelihood that both young men and women are NEET (compared to men/women living in households of welfare quintile one). The relationship is higher among women at the 4th and 5th quintiles. In addition, the presence of a non-working member in the 30- 65 age group, is associated to a higher likelihood to being NEET among young men in Armenia (7pp) but the effect is not significant among women (perhaps because gender norms assigning women the role of homemakers prevail over them looking for a job and become earners). For Georgia, most results are in line with estimates from a previous study of the World Bank on the NEET phenomenon in Georgia (World Bank, Technical Assistance 2017). According to that study, ownership of productive inputs such as land is one other important protective factor of becoming a NEET. 28 Marriage and family formation are associated with lower participation in the labor force among female population in general (not only youth). For instance, probit estimates of the determinants of participating in the labor force in Armenia show that marriage and motherhood are strongly associated with lower labor force participation; married women or women living with a partner and mothers of young children show a lower probability of engaging in the labor market. Similarly, there is a negative correlation between the proportion of children aged 6-14 and the participation of women in the labor market. These factors do not appear to affect men’s probability of being in the labor market. (World Bank, 2017a) 29 Georgia LFS (2017) does not collect information on household income but it does include information on what the sources of household income are. 32 Note that the regression analysis investigates the association of age and being NEET but does not break in sub-samples by age group because there are too few observations of NEETs in the 15-18 cohort. Yet, it is important to recognize that there may be important differences across age groups in the way risk factors influence the probability of being NEET: between the ages of 15-18 most of the youth are still in school, but after age 19, or, more notably, around age 21, the paths begin to diverge by both age and gender. In addition, the models presented here do not control by the heterogeneity (Figure 7) of the NEET condition i.e. long-term unemployed, job seeker, discouraged, unavailable, homemaker, etc., which constitutes a relevant area for further qualitative research. (See policy implications and pending questions in the Concluding Remarks section). On a final note, while regression analyses did not test for overlapping disadvantages, understanding these can also help to better identify the specific population groups that should be targeted through different types of interventions. For instance, in Armenia, a woman aged 21, married, living in the rural area has a higher chance of being NEET (NEET rate is 67.8 percent among women with these characteristics) than a man the same age, single and living in a city (21.3 percent). In Georgia, an female immigrant of Armenian background with upper secondary living in rural areas has a higher likelihood of being NEET (48.3 percent) than a Georgian male with tertiary living in urban areas (28.7 percent). 33 Figure 20: Correlates of Being NEET for Youth Aged 15-29, by Sex: Marginal effects (at the means) on the probability of being a NEET GEORGIA Women Men Age *** *** Ethnicity = 2, Azeri ** Ethnicity = 3, Armenian ** *** Ethnicity = 4, Other Highest level achieved = 2, UpperSec *** *** Highest level achieved = 3, Vocational *** *** Highest level achieved = 4, Higher * *** Rural *** *** Ever married or union *** *** IDP *** Has a disability or chronic patient *** *** Any children 0-7 in HH *** ** Any Senior 65+ in HH The total number of members in the household *** *** HH has a source of income *** ** At least one paid-employed 15-64 in HH *** *** Region code = 15, Adjara A/R ** Region code = 23, Guria *** Region code = 26, Imereti ** *** Region code = 29, Kakheti *** Region code = 32, Mtskheta-Mtianeti *** *** Region code = 38, Samegrelo and Zemo Svaneti *** *** Region code = 41, Samtskhe-Javakheti Region code = 44, Kvemo Kartli ** Region code = 47, Shida Kartli *** *** -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 Source: Authors’ calculations based on Georgia LFS (2017). Note: Reference categories are Georgian (ethnicity), lower secondary education or less, Tbilisi. All predictors at their mean value. *** p<0.01, ** p<0.05, * p<0.1 34 Figure 21: Correlates of Being NEET for Youth Aged 15-29, by Sex: Marginal effects (at the means) on the probability of being a NEET ARMENIA Women Age *** *** Men Nationality = 2, Russia Nationality = 3, Other CIS Nationality = 4, Other ** Highest level achieved = 2, Secondary *** ** Highest level achieved = 3, Vocational *** *** Highest level achieved = 4, Tertiary * * Rural Ever married or union *** *** Returned Migrant (after 2013) *** Has a disability *** *** Any children 0-7 in HH Any Senior 65+ in HH Household size Agg Welfare quint = 2 ** ** Agg Welfare quint = 3 ****** Agg Welfare quint = 4 ****** Agg Welfare quint = 5 *** *** Presence non_working 30-64 in HH *** Yerevan (omitted) Aragatsotn Ararat ** Armavir * Gegharkunik *** Lori * * Kotayk *** *** Shirak *** Sjunik *** *** Vayots Dzor Tavush -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Source: Authors’ calculations based on Armenia ILCS (2017). Note: Reference categories are Armenia (nationality), incomplete secondary education or less, aggregate welfare quintile 1, Yerevan. ‘Other CIS’ include refer to Ukraine, Belarus, Moldova, Kazakhstan, Uzbekistan, Kyrgyzstan, Tajikistan, Azerbaijan, Turkmenistan. All predictors at their mean value. *** p<0.01, ** p<0.05, * p<0.1 35 Concluding Remarks The analysis presented in this study points to persistent and high numbers in Georgia and Armenia of youth who are jobless and/or economically inactive and with limited access to learning and/or training opportunities, and these figures are among the highest in the ECA region. As many as 31 percent youth aged 15-29 in Georgia and 33 percent in Armenia are NEET. In Georgia, the phenomenon is mostly urban (due to the limited availability of jobs in cities) whereas in Armenia is as rural as urban. NEET rates are consistently higher among female than male youth, pointing to the fact that being female is a risk factor itself for labor market exclusion, and gender disparities in the likelihood of being NEET between women and men are even more pronounced in rural areas. In addition to gender gaps in the likelihood of being NEET, there are important gender differences in the condition of those who are NEET: economic inactivity – in the form of homemaking and/or caregiving work- is the most prevalent condition among NEET women whereas unemployment is the most common status among NEET men. The differences in the profile of male and female NEET emerge, and widen, during the transition from school to work. This study shows that, while young men will transition from school to work, most young women will remain inactive and out of the labor force after finishing school and this condition will last for a long period as they stay home. Furthermore, the peak of inactivity among women coincides with the family formation and childrearing years (20-35 age group), then, once in their late 30s, women re-enter the labor force (but they never do so at the same rate as men). The transition into inactivity among young women persists even though they remain slightly longer in education relative to young men. Women tend to reduce their labor supply on either the extensive or intensive margin when market, normative, and institutional forces push them toward fulfilling their caregiving mandate in the household (World Bank, 2019). Labor market detachment and career interruptions can have a permanent negative impact on women’s lifetime income, affecting their households’ living standards and human capital investments. Georgia and Armenia need to increase labor participation among young men and women alike and capitalize the investments in education of a large group of young women by implementing policies that address care needs and policies that help balance care and work responsibilities in the household. (World Bank, 2019) Besides gender, regression analysis of the probability of being NEET indicates that other individual characteristics that are strongly and positively correlated to the NEET condition are age, belonging to a vulnerable minority group (i.e. being IDP or having a disability) and having an immigration background. Household composition (i.e. presence of young children) and marital status (i.e. being married/in union/ever married) stand out as important correlates of being NEET for women but not men, even when women’s educational attainment is higher than that of men. Household aggregate welfare stands as other important correlate of the NEET condition with most NEETs living in poorer households. Finally, with regards to education, Georgia and Armenia stand out as cases where higher educational attainment does not necessarily prevent young people from becoming NEETs, and the NEET rate among higher education graduates is even bigger (particularly among women) than for those with lower education. This is a strong sign of mismatch between the demand and supply side of labor market: the economies of these countries are not able to create an adequate supply of skilled jobs for graduates, and there is persistent subsistence low-productivity employment the agriculture sector. At the same time, the quality and relevance of TVET and higher education might also be low, and there seem to be a -cognitive, socio-behavioral and technical- skills deficiency. 36 NEET youth represent a misallocation of human resource potential. These are youth who experience a difficult school-to-work transition and face some kind of ‘exclusion’ because they lack access to learning opportunities and are jobless and/or economically inactive. However, the reasons for their exclusion might be completely different and so should be the policy approach: those cyclically unemployed (i.e. job seekers) can benefit from policies that allow them to stay somewhat engaged with the labor market or in training during downturns; on the other hand, those structurally disengaged from the labor market (i.e. the different types of inactive) are more likely to become long-term unemployed or drop out of the labor force altogether and deserve a deeper policy agenda (World Bank, 2016a). According to a comprehensive review of the various policy responses to the phenomenon of NEETs conducted by the European Training Foundation, 30 EU-neighbor countries -Georgia, Armenia and Azerbaijan included- have not yet focused on the NEET phenomenon and no policy measures specifically targeting NEETs exist. Rather, general youth (and employment) policies have been developed. The most common target group in these policies is ‘unemployed youth’, but other important vulnerable sub-groups such as young women, family caregivers, discouraged workers, minorities and the inactive are left ‘off the radar’. (ETF, 2015). A recent review conducted in Georgia (World Bank, Technical Assistance, 2017) confirms this assessment and concludes that there are no targeted State programs that directly address NEET youth in the country. Moreover, NEETs are not on the list of the young people with special needs indicated in the National Youth Policy Document 31 which prioritizes youth development. Most State programs and services are either not tailored to the specific groups of youth or do not clearly state which categories of young people they are aiming to engage. The lack of targeted policy interventions is partly explained because there is little information on the vulnerable sub-groups of NEETs. Future qualitative research will shed more light about the different needs each disadvantaged group of youth faces as well as on the pending questions regarding NEET that have not been answered yet (see Box 3 below). Box 3. What pending questions regarding NEET have not been answered yet? o The absence of an articulated and comprehensive youth agenda to tackle NEETs partly reflects the limited research conducted on disadvantaged groups of youth, and their pathways to becoming NEET. The reasons why young women fall under inactivity are different than those faced by the discouraged or re-entrants. In addition, youth may fail to transition either because they have a disability, they are members of an ethnic minority, refugees or displaced persons, or returned migrants (ETF, 2015; World Bank, 2017). o Investigating the root of youth discouragement from participating in the labor market education or training is also key as it is likely to be different than the root of youth unemployment. Moreover, data from most labor force surveys (the EU included) do not allow for an evaluation of whether becoming NEET is voluntary or not (ETF, 2015). Understanding young women’s and men’s decision- 30 The ETF is an EU agency that provides support to countries bordering the EU to improve their vocational education and training systems, analyze their skills needs, and develop their labor markets. Partner countries include countries preparing for EU accession, North Africa, the Middle East, the Balkans and the former Soviet Union. 31 The Policy recognizes, and commits to meeting, young people’s needs for education, sexual and reproductive health services, employment, and participation in decision-making. It was followed with the Youth Policy Action Plan 2014-2020 which focuses on four areas: participation, education, health and protection issues. 37 making processes to identify “risky points� and potential ways of influencing the processes (e.g. behavioral interventions). o Data to increase understanding of labor migration and the extent to which it impacts youth’s transition from school to work is still limited. Labor migration in the South Caucasus (Armenia, particularly) remains a rich area of research. Large gaps remain in the literature, including information on young migrants’ motives and characteristics. o There is also limited research on more ‘unobservable’ obstacles to transition such as norms and values that are upheld by both employers and potential workers, which may also narrow employment opportunities for youth (World Bank, 2016a). o Exploring attitudes and social norms that shape individual and family decisions and that relate directly and indirectly to the labor market (World Bank, 2016a). Moreover, norms and values in the South Caucasus tend to be most detrimental for young women, who are often confronted with ‘traditional’ roles around caregiving and household responsibilities. Understanding what works to expand men’s role as caregivers, which could also inform an extension of parental benefits. National policies need to take a tailored and integrated approach to mitigate the failed school-to-work that triggers the NEET condition. The policy response should consider all the subgroups within the NEET population. Necessary steps to develop policy measures are the identification of excluded groups of young people (See Box 3), making these groups more visible in the political agenda, and recognizing what stage of their specific exclusion they are at, to then intervene with the most effective type of support. Policies should also employ a life-cycle approach as age seems to be an important correlator of being NEET: between the ages of 15-18 most of the youth are still in school, but after age 19, or, more notably, around age 21, the paths begin to diverge by both age and gender. Furthermore, given the strong gendered roles aspects, interventions should include specific considerations to tackle constraining norms and could include behavioral experiments to better understand what works before scaling up. 38 References Alfieri, S., E. Sironi, E. Marta, A. Rosina, D. Marzana. 2015. “Young Italian NEETs (Not in Employment, Education, or Training) and the Influence of Their Family Background.� Europe's Journal of Psychology, 11(2), 311-22. Bardak, U, M. Rubal, F. Rosso. 2015. “Young People not in Employment, Education or Training (NEET). An overview in ETF partner countries.� ETF, Turin, Italy. https://www.etf.europa.eu/sites/default/files/m/BFEEBA10DD412271C1257EED0035457E_NEE Ts.pdf Carcillo, S., R. Fernández, S. Konigs and A. Minea. 2015. "NEET Youth in the Aftermath of the Crisis: Challenges and Policies", OECD Social, Employment and Migration Working Papers, No. 164, OECD Publishing, Paris. https://doi.org/10.1787/5js6363503f6-en Cuberes, D., and M. Teignier. 2016a. “Aggregate Effects of Gender Gaps in the Labor Market: A Quantitative Estimate.� Journal of Human Capital 10 (1): 1–32 –––––––. 2016b. “How Important Are Labor Market Gender Gaps in the South Caucasus?� Paper prepared for the “FY16 South Caucasus Gender Assessment Program.� de Hoyos, R, H. Rogers, M. Székely. 2016. Out of School and Out of Work: Risk and Opportunities for Latin America’s Ninis. World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/22349 Erdoğan, E., N. Yentürk, A. A. Akyüz, Y. Kurtaran, L. Yurttagüler, K. Dursun, B. Oy. 2017. “Being a NEET in Turkey: Determinants and Consequences.� Working Paper No. 30 – April 2017. Istanbul Bilgi University. http://www.iai.it/sites/default/files/p2y_30.pdf ETF. 2014. Young people not in employment, education or training in the EU neighborhood countries. https://www.etf.europa.eu/sites/default/files/m/13B02C4480535C87C1257C980035C440_NEET s%20in%20EU%20neighbourhood%20-%20report.pdf Eurofound. 2016. Exploring the diversity of NEETs. Publications Office of the European Union, Luxembourg. https://www.eurofound.europa.eu/sites/default/files/ef_publication/field_ef_document/ef160 2en.pdf _______. 2012. NEETs – Young people not in employment, education or training: Characteristics, costs and policy responses in Europe, Publications Office of the European Union, Luxembourg. https://www.eurofound.europa.eu/sites/default/files/ef_publication/field_ef_document/ef125 4en.pdf Friedrich-Ebert-Sung. 2017. Generation in Transition. Youth Study 2016 – Georgia. FES, South Caucasus Regional Office. Tbilisi, Georgia. http://library.fes.de/pdf-files/bueros/georgien/13150.pdf _______. 2016. Independence Generation. Youth Study 2016 – Armenia. FES, Yerevan, Armenia. http://library.fes.de/pdf-files/bueros/georgien/13149.pdf Handel, Michael J., Alexandria Valerio, and Maria Laura Sanchez Puerta. 2016. Accounting for Mismatch in Low- and Middle-Income Countries: Measurement, Magnitudes, and Explanations. Directions in Development. Washington, DC: World Bank. https://openknowledge.worldbank.org/bitstream/handle/10986/24906/9781464809088.pdf?se quence=4&isAllowed=y Posadas, P, M. Makovec, R.F. Jaef, C. Gruen, and M.I. Ajwad (2018) “Georgia at Work: Assessing the Jobs Landscape�. World Bank, Washington, DC. http://documents.worldbank.org/curated/en/995521527068940160/pdf/126461-WP-P165644- PUBLIC-GeorgiaJDPrinting.pdf 39 Quintano, C., P. Mazzocchi, A. Rocca. 2018. “The determinants of Italian NEETs and the effects of the economic crisis�. Genus (2018) 74: 5. https://link.springer.com/content/pdf/10.1186%2Fs41118-018-0031-0.pdf Rutkowski, Jan Jerzy. 2013. “Workforce skills in the eyes of the employers: results of the Georgia STEP employer skills survey (English).� Washington, D.C.: World Bank Group. http://documents.worldbank.org/curated/en/191321468178757151/Workforce-skills-in-the- eyes-of-the-employers-results-of-the-Georgia-STEP-employer-skills-survey _______. 2013a. “Skills Employers Seek: Results of the Armenia STEP Employer Skills Survey.� World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/22530 Serrière, Nicolas. 2014. “Labour market transitions of young women and men in Armenia�. International Labour Office, Youth Employment Programme. Geneva: ILO. https://www.un.am/up/library/Labor_Market_Armenia_eng.pdf UNESCO. 2015. “Education for All. National Review 2015. Georgia.� UNFPA and Geostat. 2017. Young People in Georgia. An overview based on the 2014 General population Census Data. Tbilisi, Georgia. https://georgia.unfpa.org/sites/default/files/pub-pdf/5.%20YOUTH-Print_ENGL_F_0.pdf UN WOMEN. 2018. Women’s economic inactivity and engagement in the informal sector in Georgia. Causes and Consequences. Tbilisi, Georgia. http://georgia.unwomen.org/en/digital-library/publications/2018/12/womens-economic-inactivity-and- engagement-in-the-informal-sector-in-georgia World Bank. 2019. Why should we care about care? Supply and Demand Assessment of Care Services in Georgia: A Mixed Methods Study (English). Washington, D.C.: World Bank Group. http://documents.worldbank.org/curated/en/887801548269893452/Why-should-we-care- about-care-Supply-and-Demand-Assessment-of-Care-Services-in-Georgia-A-Mixed-Methods- Study –––––––. 2018. Georgia: From Reformer to Transformer. A Systematic Country Diagnostic (English). Washington, D.C.: World Bank Group. http://documents.worldbank.org/curated/en/496731525097717444/pdf/GEO-SCD-04-24- 04272018.pdf –––––––. 2017. Technical Assistance # P161162: Strengthening the Capacity of the Ministry of Sports and Youth of Georgia to Address the Needs of Youth Who Are Not in Employment, Education, or Training (NEETs). Washington, D.C.: World Bank Group. –––––––. 2017a. Future Armenia: connect, compete, prosper. A Systematic Country Diagnostic (English). Washington, D.C.: World Bank Group. http://documents.worldbank.org/curated/en/716961524493794871/pdf/Armenia-SCD-in-Eng- final-04192018.pdf –––––––. 2016. Armenia Gender Assessment 2016 (English). Washington, D.C.: World Bank Group. http://documents.worldbank.org/curated/en/250481491242159715/Armenia-gender- assessment-2016 –––––––. 2016a. Ten messages about youth employment in South East Europe. South East Europe Regular Economic Report No. 9S, Special Topic, World Bank, Washington, DC. http://documents.worldbank.org/curated/en/940151479220585911/pdf/110291-REVISED-NWP- PUBLIC.pdf 40 Annex: Additional Figures and Tables Table A1: Status in the Labor Force and NEET Rates Among Men and Women, 15-24 age group Georgia Armenia Women Men Total Women Men Total In school/training only 40.3 37.9 39.0 44.9 45.9 45.4 Work only 21.0 31.0 26.2 14.2 22.0 18.0 Work + school/train. 4.7 6.0 5.4 4.0 7.4 5.6 Unemployed + school/train. 4.3 3.5 3.9 4.8 3.6 4.2 Unemployed (NEET) 8.2 9.7 9.0 10.0 10.9 10.4 Inactive (NEET) 21.6 12.0 16.5 22.2 10.2 16.4 Total 100.0 100.0 100.0 100.0 100.0 100.0 Georgia Armenia Women Men Total Women Men Total NEET Rate 29.5 21.4 25.2 32.2 21.1 26.9 NEET, Urban 28.4 24.9 26.6 29.3 24.4 27.0 NETT, Rural 31.0 16.3 23.1 36.1 16.9 26.8 Share of NEET 55.4 44.6 100 62.2 37.8 100 Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). Note: ‘In education or training’ refers to inactive people who are receiving any type of formal education or non-formal education or vocational training, including full or part-time formats. Figure A1: Status of Youth 15-29 in the Labor Force, by Sex and Location Georgia Armenia 100 100 80 80 60 60 40 40 20 20 0 0 Women Total Women Total Women Total Men Men Men Women Total Women Total Women Total Men Men Men Total Urban Rural Total Urban Rural In education (or training) Work only In education (or training) Work only Work + edu/train. Unemp + edu/train. Work + edu/train. Unemp + edu/train. Unemployed (NEET) Inactive (NEET) Unemployed (NEET) Inactive (NEET) Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). Note: ‘In education or training’ refers to inactive people who are receiving any type of formal education or non - formal education or vocational training, including full or part-time formats. 41 Table A2: Status of Youth, by Age Group and Sex Georgia Armenia Women Men Women Men 15_22 23_29 15_22 23_29 15_22 23_29 15_22 23_29 Employed 20.1 49.8 30.8 66.8 12.8 38.7 22.2 65.5 Unemployed 11.2 12.7 11.2 20.2 13.7 14.2 11.8 20.5 Inactive (not in edu) 17.8 35.7 12.0 11.4 16.2 43.6 10.1 9.7 In education or train. 50.8 1.8 46.1 1.5 57.3 3.5 55.9 4.4 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). Figure A2: School to Work Transition among Men and Women, by Urban/Rural location GEORGIA - URBAN Men Women 100 100 80 80 60 60 40 40 20 20 0 0 15 17 19 21 23 25 27 29 31 33 35 37 39 15 17 19 21 23 25 27 29 31 33 35 37 39 In education (or training) Work only In education (or training) Work only Work+edu/train. Unemp+edu/train. Work+edu/train. Unemp+edu/train. Unemployed (NEET) Inactive (NEET) Unemployed (NEET) Inactive (NEET) GEORGIA - RURAL Men Women 100 100 80 80 60 60 40 40 20 20 0 0 15 17 19 21 23 25 27 29 31 33 35 37 39 15 17 19 21 23 25 27 29 31 33 35 37 39 In education (or training) Work only In education (or training) Work only Work+edu/train. Unemp+edu/train. Work+edu/train. Unemp+edu/train. Unemployed (NEET) Inactive (NEET) Unemployed (NEET) Inactive (NEET) ARMENIA - URBAN 42 Men Women 100 100 80 80 60 60 40 40 20 20 0 0 15 17 19 21 23 25 27 29 31 33 35 37 39 15 17 19 21 23 25 27 29 31 33 35 37 39 In education (or training) Work only In education (or training) Work only Work+edu/train Unemp+edu/train Work+edu/train Unemp+edu/train Unemployed (NEET) Inactive (NEET) Unemployed (NEET) Inactive (NEET) ARMENIA - RURAL Men Women 100 100 80 80 60 60 40 40 20 20 0 0 15 17 19 21 23 25 27 29 31 33 35 37 39 15 17 19 21 23 25 27 29 31 33 35 37 39 In education (or training) Work only In education (or training) Work only Work+edu/train Unemp+edu/train Work+edu/train Unemp+edu/train Unemployed (NEET) Inactive (NEET) Unemployed (NEET) Inactive (NEET) Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). Note: ‘In education or training’ refers to inactive people who are receiving any type of formal education or non - formal education or vocational training, including full or part-time formats. Figure A3: Family Formation and the NEET Condition: Share of Married/In union/Ever married NEET, by Age and Urban/Rural Location Georgia 100 Urban 100 Rural 80 80 60 60 40 40 20 20 0 0 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 NEET-Women NEET-Men NEET-Women NEET-Men 43 Armenia Urban Rural 100 100 80 80 60 60 40 40 20 20 0 0 16 17 18 19 20 21 22 23 24 25 26 27 28 29 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 NEET - Women NEET - Men NEET - Women NEET - Men Source: Authors’ calculations based on Georgia LFS (2017) and Armenia LFS (2017). Note: Married category includes married, ever married (divorced, separated, widowed) or in union. In the case of Armenia there are too few observations of female NEETs ages 15 and 16 years reporting on their marital status, and the same is true for Georgia among female NEETs ages 15 to 18. Age figures for these data points are therefore omitted. 44 Table A3. Probit Estimates of Being NEET for Youth Aged 15-29, by Sex. Georgia. VARIABLES Women Men Age 0.008*** 0.017*** (0.003) (0.002) Ethnicity = 2, Azeri -0.006 0.074** (0.040) (0.038) Ethnicity = 3, Armenian 0.137** 0.207*** (0.057) (0.056) Ethnicity = 4, Other 0.018 0.075 (0.054) (0.069) Highest level achieved = 2, UpperSec 0.152*** 0.085*** (0.021) (0.020) Highest level achieved = 3, Vocational 0.241*** 0.111*** (0.033) (0.031) Highest level achieved = 4, Higher 0.078*** 0.049* (0.029) (0.027) Rural -0.049*** -0.066*** (0.018) (0.016) Ever married or union 0.317*** -0.065*** (0.026) (0.023) IDP -0.036 -0.134*** (0.041) (0.036) Has a disability or chronic patient 0.369*** 0.306*** (0.074) (0.054) Any children 0-7 in HH 0.059** -0.064*** (0.025) (0.024) Any Senior 65+ in HH -0.027 -0.007 (0.019) (0.017) The total number of members in the household 0.029*** 0.015*** (0.006) (0.005) HH has a source of income -0.058** -0.122*** (0.028) (0.022) At least one paid-employed 15-64 in HH -0.199*** -0.141*** (0.036) (0.027) At least one unemployed 15-64 in HH 0.256*** 0.281*** (0.020) (0.016) Region code = 15, Adjara A/R -0.015 0.061** (0.027) (0.025) Region code = 23, Guria 0.061 0.167*** (0.038) (0.035) Region code = 26, Imereti 0.131*** 0.055** (0.030) (0.025) Region code = 29, Kakheti 0.119*** 0.041 (0.034) (0.025) Region code = 32, Mtskheta-Mtianeti 0.181*** 0.121*** (0.039) (0.035) Region code = 38, Samegrelo and Zemo Svaneti 0.133*** 0.130*** (0.034) (0.029) Region code = 41, Samtskhe-Javakheti 0.019 -0.044 (0.041) (0.028) 45 Region code = 44, Kvemo Kartli 0.063** 0.006 (0.031) (0.024) Region code = 47, Shida Kartli 0.181*** 0.069*** (0.030) (0.024) Observations 5,917 6,138 Notes: Reference categories are Georgian (ethnicity), lower secondary education or less, Tbilisi. Standard errors in parentheses. All predictors at their mean value *** p<0.01, ** p<0.05, * p<0.1 Table A4. Probit Estimates of Being NEET for Youth Aged 15-29, by Sex. Armenia. VARIABLES Women Men Age 0.018*** 0.011*** (0.004) (0.002) Nationality = 2, Russia 0.019 -0.090 (0.099) (0.059) Nationality = 3, Other CIS -0.136 0.068 (0.150) (0.118) Nationality = 4, Other -0.074 -0.109** (0.130) (0.052) Highest level achieved = 2, Secondary 0.202** 0.094*** (0.100) (0.033) Highest level achieved = 3, Vocational 0.296*** 0.173*** (0.106) (0.051) Highest level achieved = 4, Tertiary 0.186* 0.075* (0.105) (0.041) Rural -0.022 -0.006 (0.030) (0.024) Ever married or union 0.359*** -0.095*** (0.036) (0.030) Returned Migrant (after 2013) 0.021 0.269*** (0.088) (0.032) Has a disability 0.638*** 0.395*** (0.153) (0.063) Any children 0-7 in HH 0.020 -0.020 (0.034) (0.030) Any Senior 65+ in HH -0.034 -0.006 (0.029) (0.022) Household size 0.007 0.005 (0.009) (0.008) Agg Welfare quint = 2 -0.072** -0.072** (0.037) (0.034) Agg Welfare quint = 3 -0.107*** -0.152*** (0.037) (0.032) Agg Welfare quint = 4 -0.192*** -0.159*** (0.035) (0.033) Agg Welfare quint = 5 -0.216*** -0.181*** (0.040) (0.034) 46 Presence non_working 30-64 in HH 0.038 0.069*** (0.027) (0.020) Admin. Region = 2, Aragatsotn 0.046 -0.013 (0.073) (0.061) Admin. Region = 3, Ararat 0.040 -0.071** (0.051) (0.031) Admin. Region = 4, Armavir -0.021 0.064* (0.043) (0.038) Admin. Region = 5, Gegharkunik 0.036 0.124*** (0.053) (0.043) Admin. Region = 6, Lori 0.081* 0.073* (0.046) (0.038) Admin. Region = 7, Kotayk 0.237*** 0.134*** (0.044) (0.040) Admin. Region = 8, Shirak 0.151*** 0.050 (0.048) (0.038) Admin. Region = 9, Sjunik -0.187*** -0.110*** (0.047) (0.026) Admin. Region = 10, Vayots Dzor 0.087 0.081 (0.059) (0.049) Admin. Region = 11, Tavush 0.056 -0.016 (0.058) (0.038) Observations 2,659 2,425 Note: Reference categories are Armenia (nationality), incomplete secondary education or less, aggregate welfare quintile 1, Yerevan. ‘Other CIS’ include refer to Ukraine, Belarus, Moldova, Kazakhstan, Uzbekistan, Kyrgyzstan, Tajikistan, Azerbaijan, Turkmenistan. All predictors at their mean value *** p<0.01, ** p<0.05, * p<0.1 47 Figure A4: Correlates of Being NEET for Youth Aged 15-29, by Sex: Linear Probability Estimates on being a NEET GEORGIA Age *** *** Age2 *** *** Women Ethnicity = 2, Azeri Men ** Ethnicity = 3, Armenian ** *** Ethnicity = 4, Other Highest level achieved = 2, UpperSec Highest level achieved = 3, Vocational *** Highest level achieved = 4, Higher Rural *** *** Marital_st#age_group Single#20_24 ** Single#25_29 Ever married/Union#15_19 *** Ever married/Union#20_24 ** *** Ever married/Union#25_29 * *** IDP *** Has a disability or chronic patient *** *** Any children 0-7 in HH ** *** Any Senior 65+ in HH The total number of members in the household *** ** HH has a source of income *** ** *** At least one paid-employed 15-64 in HH*** Adjara A/R Guria ** Imereti *** ** Kakheti ** *** Mtskheta-Mtianeti *** Samegrelo and Zemo Svaneti *** *** Samtskhe-Javakheti *** *** Kvemo Kartli Shida Kartli ** -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 Source: Authors’ calculations based on Georgia LFS (2017). Note: Reference categories are Georgian (ethnicity), lower secondary education or less, Tbilisi. *** p<0.01, ** p<0.05, * p<0.1 48 Figure A5: Correlates of Being NEET for Youth Aged 15-29, by Sex: Linear Probability Estimates on being a NEET ARMENIA Age *** *** Age2 *** *** Nationality = 2, Russia Women Nationality = 3, Other CIS Men Nationality = 4, Other Highest level achieved = 2, Secondary Highest level achieved = 3, Vocational Highest level achieved = 4, Tertiary * Rural * Marital_st#age_group Single#20_24 Single#25_29 * Ever married/Union#15_19 *** *** Ever married/Union#20_24 *** *** Ever married/Union#25_29 ** ** Returned Migrant (after 2013) *** Has a disability *** *** Any children 0-7 in HH ** Any Senior 65+ in HH Household size Agg Welfare quint = 2 ** ** Agg Welfare quint = 3 ****** Agg Welfare quint = 4 *** *** Agg Welfare quint = 5 *** *** Level of Education of the HH = 2, Primary *** Level of Education of the HH = 3, Secondary *** Level of Education of the HH = 4, Post-Secondary *** Presence non_working 30-64 in HH *** Yerevan (omitted) Aragatsotn Ararat * Armavir Gegharkunik *** Lori * * Kotayk *** *** Shirak *** Sjunik *** *** Vayots Dzor * Tavush -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 Source: Authors’ calculations based on Armenia ILCS (2017). Note: Reference categories are Armenia (nationality), incomplete secondary education or less, aggregate welfare quintile 1, Yerevan. ‘Other CIS’ include refer to Ukraine, Belarus, Moldova, Kazakhstan, Uzbekistan, Kyrgyzstan, Tajikistan, Azerbaijan, Turkmenistan. *** p<0.01, ** p<0.05, * p<0.1 49