Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Closing the Gender Gaps among Marginalized Roma in the Western Balkans A Summary of Findings and Policy Recommendations Monica Robayo-Abril, Carmen de Paz Nieves, and Trinidad Saavedra Facusse Poverty and Equity Global Practice Europe and Central Asia Region The World Bank May 2019 i CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS © 2019 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. ii CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Contents List of Figures ................................................................................................................................ iv List of Tables .................................................................................................................................. v List of Boxes ................................................................................................................................... v Acknowledgements ........................................................................................................................ vi Acronyms and Abbreviations ....................................................................................................... vii Country Codes .............................................................................................................................. vii 1 Introduction ............................................................................................................................. 8 2 Literature Review: What do previous data tell us about the status of Roma women in Western Balkans ........................................................................................................................... 10 3 Roma Gender Profile and Determinants of Gender Gaps among the Roma Population ....... 13 3.1 Gaps In And Constraints to Roma women´s Agency and Voice ................................... 14 3.2 Access and barriers to Health and Education ................................................................. 21 3.3 Access and constraints to Economic Opportunities ....................................................... 32 4 Policies and Interventions ...................................................................................................... 41 4.1 Institutional Framework and Policies on Roma Gender Equality .................................. 41 4.2 The Road Forward: What Works for Roma Women Inclusion ..................................... 44 Bibliography ................................................................................................................................. 48 Appendix. Regional Overview Tables by Gender ........................................................................ 61 iii CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS List of Figures Figure 1. WDR Framework to Assess Gender Inequalities .......................................................... 14 Figure 2. Child Marriage is Common Among Roma Women Ages 20-49 .................................. 16 Figure 3. Roma Women Are Less Likely to Report They Have Money of Their Own, Even If One Controls for Employment ...................................................................................................... 17 Figure 4. Most Roma Women Report They Decide How to Spend Some or All Their Own Money ........................................................................................................................................... 17 Figure 5. Child marriage is considered socially acceptable among a large share of the Roma population; marriage among Roma girls is more likely to be acceptable than among Roma boys ....................................................................................................................................................... 18 Figure 6. High tolerance of gender-based violence among Roma communities .......................... 18 Figure 7. Time Use among Roma Wives and Husbands .............................................................. 19 Figure 8. Self-Reported Discrimination by Roma Men and Women ........................................... 21 Figure 9. High incidence of unmet needs for medical care; in some countries, a large share of the gap remains unexplained , Year 2017 ........................................................................................... 22 Figure 10. Gender Disparities in Pre-primary Education Are Not as Large as in Other Levels, Since Enrollment is Low Among Both Roma Girls And Boys .................................................... 25 Figure 11. The Gender Gap is Wide in Upper Secondary Completion Only for Serbia and Kosovo .......................................................................................................................................... 26 Figure 12. The Tertiary Education Completion Rate is Extremely Low Among Roma Males and Females ......................................................................................................................................... 26 Figure 13. Large Gender Gaps in Compulsory Education Completion (ISCED Levels 1 & 2) ... 26 Figure 14. Blinder-Oaxaca Decomposition of The Gender Gap in Compulsory Education Completion Rate Among Roma Population, ages 18-21, year 2017 ............................................ 29 Figure 15. Child Marriage is Correlated with Lower Educational Attainment ............................ 30 Figure 16. The Educational Aspirations of Roma parents for Their Boys and Girls are Similar . 31 Figure 17. Large Gender Gaps in Labor Force Participation among Roma ................................. 32 Figure 18. Labor Force Participation Is Low among Less Well-Educated Roma Females .......... 33 Figure 19. Large Gender gaps in Unemployment and Employment-to-Population Ratio among Roma, Ages 15–64 ........................................................................................................................ 34 Figure 20. High Incidence of NEETS and Low Employment-to-Population Ratio among Roma, Ages 15–64 ................................................................................................................................... 35 Figure 21. Framework for Contraints to Employment for Roma Females ................................... 35 Figure 22. Most of the Gender Gap in Labor Force Participation and Employment among Roma Cannot Be Explained by Differences in Characteristics between Roma Men and Roma Women37 Figure 23. High Young Dependency Ratios and Limited Use of Formal Child care in the Western Balkans .......................................................................................................................................... 39 Figure 24. Working-age Roma Females Attain Less Education on Average than Men, Already at Low Levels; Self-reported Literacy .............................................................................................. 40 Figure 25. The Tax Burden is High for Low-Wage and Part-time Workers, Disproportionately Affecting Roma Males and Females Who Are Relatively More Likely to Take These Type of Jobs ............................................................................................................................................... 41 iv CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS List of Tables Table 1. Reasons for Not sending 3-to-5 years Old Children to Childcare, Preschool or Kindergarten, By Gender .............................................................................................................. 27 Table 2. Factors of making respondents reconsidering sending child to pre-primary: No school fees (attendance for free) .............................................................................................................. 27 Table 3. Many Roma Cite Cost as the Main Reason for Not Attending School; among Roma Women, Child Marriage Is Also a Factor ..................................................................................... 28 Table 4. Legal, Institutional and Policy Framework for Gender Equality in Western Balkan Countries ....................................................................................................................................... 42 Table 5. Legal, Institutional and Policy Framework for Roma Inclusion in Western Balkan Countries ....................................................................................................................................... 43 Table 6. Regional Overview at a Glance: 2011 Roma Levels by Sex .......................................... 61 Table 7. Regional Overview at a Glance: 2011 Non-Roma Levels by Sex .................................. 62 Table 8. Regional Overview at a Glance: 2017 Roma Levels by Sex .......................................... 63 Table 9. Regional Overview at a Glance: 2017 Non-Roma Neighbor Levels by Sex .................. 64 Table 10. Regional Overview at a Glance: Roma 2011-2017 Percentage Point Changes by Sex 65 Table 11. Regional Overview at a Glance: Non-Roma neighbor 2011-2017 Percentage Point Changes by Sex ............................................................................................................................. 66 List of Boxes Box 1. Who Are The Roma?........................................................................................................... 8 Box 2. Relevant Studies and Data on Roma and Gender in the Western Balkans ....................... 11 Box 3. Behavioral Economics and Roma Exclusion .................................................................... 45 v CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Acknowledgements This work was prepared by a team from the Poverty and Equity Global Practice (GP). The lead authors are Monica Robayo-Abril (Economist, Poverty and Equity GP), Carmen de Paz Nieves (Consultant, Poverty and Equity GP), and Trinidad Saavedra Facusse (Consultant, Poverty and Equity GP). This report was prepared under the guidance of Linda Van Gelder (Country Director, ECCWB) and Carlos Silva-Jauregui (ECA Practice Manager, Poverty and Equity GP). The team is grateful for valuable inputs received from the peer reviewers Andrea Woodhouse (Senior Social Development Specialist, GSU03) and Ferdous Jahan (Senior Social Development Specialist, GSU06). vi CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Acronyms and Abbreviations ALMP active labor market policy CCT conditional cash transfer CoE Council of Europe EC European Commission ECA Europe and Central Asia ECAPOV Europe and Central Asia poverty database EU European Union GDP Gross Domestic Product HBS Househould Budget Survey ISCED International Standard Classification of Education MICS Multiple Indicator Cluster Survey OECD Organization for Economic Co-operation and Development PISA Program for International Student Assessment RRS Regional Roma Survey SILC Survey on Income and Living Conditions WB World Bank UNDP United Nations Development Programme NEET Youth who is not in employment, education or training Country Codes ALB Albania BiH Bosnia and Herzegovina KSV Kosovo MNE Montenegro MKD Republic of North Macedonia SRB Serbia vii CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS 1 Introduction Roma women are one of the most deprived groups in Europe, as they suffer a double layer of exclusion: as women, and as members of Europe´s largest ethnic minority. Although there are no reliable data on the Roma population in the Western Balkans, available estimates suggest that the share of national populations represented by Roma ranges between 1.7 percent in Bosnia and Herzegovina and 9.6 percent in North Macedonia. Over half of these are women. As evidenced in the recently published World Bank report on Roma (Robayo-Abril and Millan 2019), the Roma population in Europe has very limited access to basic services and economic opportunities, including education, health care, housing, employment and documentation, even relative to their non-Roma neighbors who live in close physical proximity. As will be shown in this report, the status and access to services and opportunities for Roma women is even poorer not only vis-à-vis non-Roma female neighbors but also when compared to Roma men. Box 1. Who Are The Roma? Roma is used to refer to a number of groups (for example, Roma, Sinti, Kale, Gypsies, Romanichels, Boyash, Ashkali, Egyptians, Yenish, Dom, Lom, Rom, Abdal), including travelers, without denying the specificities of these groups. These groups are all considered under the wider Roma umbrella in the European Union (EU) Framework for National Roma Integration Strategies (European Commission 2011). This report draws extensively from the 2011 and 2017 rounds of the Regional Roma Survey (RRS), the most comprehensive survey to date on living conditions and human development outcomes among marginalized Roma households in the Western Balkans, as well as non- Roma households in the vicinity of Roma. In the survey, Roma populations are defined as those who self-identified as Ashkali, or Egyptians, or Roma. Therefore, the Roma sample was constructed based on the implicit endorsement of external identification (or implicit self-identification). In both waves of the RRS, the sampling framework included communities in which the share of the Roma population was larger than the share of Roma in the national population, referred to as marginalized Roma. The survey relies on non-Roma neighbors as a comparator group. In theory, because the two populations live in close physical proximity, varying circumstances that are expected to affect living conditions and outcomes, such as those related to geography, should be controlled for (Robayo-Abril and Millan 2019). Ensuring the social and economic inclusion of Roma women is not only a moral imperative but also smart economics. On the one hand, gender inequality in human capital investments and in access to economic opportunities leads to untapped talent and potential that limits economic growth prospects, and to lower development outcomes related to poor agency. On the other, the educational and labor market gaps affecting the Roma also translate into foregone fiscal revenues, higher social protection spending and large opportunity costs. Investing in developing the human capital of the Roma, and particularly of Roma women, and working towards their economic inclusion become even more pressing in the context of ageing; the Roma represent a particularly young population group, and therefore a large labor reservoir (Robayo- Abril and Millan 2019); female Roma remain largely – and more so than men – excluded from economic opportunity. 8 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Both gender gaps and the exclusion of Roma entail large wealth losses. Previous evidence (Cuberes and Teigner 2015) shows that the constraints that female face may result in misallocated talent, which leads to sizable income per capita losses. For instance, GDP could be 20 percent higher, on average, if women were to participate in the labor market at the same levels as men do; 5 percent of this GDP loss is due to the gap in participation of women as entrepreneurs. No similar estimates exist to quantify the losses due to gender gaps among Roma communities. Female’s employment not only promotes development through their impact on productivity, employment and economic growth, but also by increasing women’s voice and agency, leading to positive spillovers at the individual, family and societal level (Klugman et al. 2014). In addition, if the Roma population of working age were to exhibit the same employment rates and labor earnings as the general population in Serbia, the total gains from increased productivity alone could range from €314 million to €1.28 billion a year, or from 0.9 percent to 3.5 percent of 2017 Serbian gross domestic product (GDP) (World Bank 2015b). Poor access to economic opportunities among vulnerable communities is not surprising considering the poor labor market outcomes registered in the region. Labor markets in the Western Balkan countries are characterized by low employment rates and high unemployment compared to European standards. Almost the entire region is facing an ageing and shrinking population due to low birth rates and high migration. Unemployment fell between the second quarter of 2016 and that of 2017 by an estimated 169,000 people, from 18.6 to 16.2 percent, reaching historical lows in some countries. Despite the overall rise in employment and drop in unemployment, low activity rates particularly among women and young people, a large share of long-term unemployment (close to 80 percent in some countries) and high informality (i.e., those self-employed in unregistered businesses, wage workers without written contract and unpaid family workers) remain key challenges. Youth unemployment continued to be high compared to other EU countries: almost one quarter of the youth population was not in employment, education or training (NEET), which despite representing a decline from the previous year is still high by international standards (World Bank 2018b). Ethnic minorities like Roma communities, who are already in a hardship securing their basic rights, disproportionately suffer the consequences of unemployment and poor access to the labor market, with Roma females being particularly affected. According to the 2017 UNDP-WB-EC Regional Roma Survey, on average 68 percent of marginalized Roma aged 15- 24 in the Western Balkans are not in employment, education or training (NEET) compared with only 34 percent among their non-Roma counterparts. Among women aged 15-24, 78 percent are NEET, vis-à-vis 59 percent of their male counterparts. Furthermore, Roma face higher unemployment rates than non-Roma populations living in their close proximity, with females being particularly affected. In the region, the unemployment rate among Roma females is 54 percent, compared to 41 percent among Roma males. The labor force participation rate among Roma women is extremely low: in 2017, just 21.3 of Roma women participated in the labor market, compared to 49 percent of Roma males. 1 Limited access to jobs and socio-economic rights let Roma people remain trapped in a vicious cycle of poverty and social exclusion, where Roma women suffer the most, as they experience ‘double discrimination’(O’Higgins 2013). 1 Population weighted regional statistics constructed with the 2017 Roma Regional Survey. 9 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS This report builds on the analysis presented in the recently published Roma report (Robayo-Abril and Millan 2019), exploring in more detail the gender dimensions of exclusion. It is intended to be a concise and timely summary that highlights the key aspects of gender equality among marginalized Roma communities in the Western Balkans. The goal of the report is to strengthen the knowledge base and evidence to understand the key determinants of gender gaps among the Roma population. For that the report offers a summary diagnostic of the most important barriers that female Roma face, in particular, accessing education and employment; further, it explores the ways in which Roma women’s employment and educational outcomes are constrained above and beyond the constraints faced by Roma males. The report is organized as follows. Section 2 provides a brief synthesis of available studies on gender among Roma populations. Section 3 presents a gender profile and reviews gender inequality in its multiple dimensions, involving access to human capital, economic opportunities, and agency. Section 4 presents a brief summary of the the current policy environment with regards to Roma gender equality, including the institutional and legal framework, as well as policies and interventions that may work to promote gender equality among marginalized communities. 2 Literature Review: What do previous data tell us about the status of Roma women in Western Balkans Previous evidence of the gender gaps among Roma populations in the Western Balkans and their driving factors remains limited, mostly due to data constraints. Ethnicity is generally not included in national household surveys, and sources of data are limited. The World Bank Country Gender Assessments for Bosnia and Herzegovina (World Bank 2015d), North Macedonia (World Bank 2013a), and Serbia (World Bank 2016c) and the respective policy notes on access to economic opportunities for women in the three countries provide some descriptive analyses in this regard (mostly based on the 2011 UNDP-WB-EC Regional Roma Survey and the more recent Multiple Indicator Cluster Survey or MICS on Roma2). Qualitative studies in Serbia (Majumdar and Woodhouse 2019) and Bosnia and Herzegovina (World Bank 2017a) conducted by the World Bank offer further information. Other non-World Bank relevant sources of information in this area are highlighted in the box below. There seems to be more data and analysis on the situation of Roma women available in the case of Bosnia and Herzegovina and Serbia, while these seem to be particularly scarce for Kosovo.3 See Box 2 for details. 2 UNICEF MICS 2013 Montenegro, MICS 2011 North Macedonia, MICS 2014 Serbia, MICS 2013-2014 Kosovo, MICS 2011-2012 Bosnia and Herzegovina. 3 As an example, the World Bank Country Gender Assessment (World Bank 2017c) barely pays any attention to the ethnicity dimension, possibly due to the fact that the 2011 Regional Roma survey data was not collected in Kosovo. 10 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Box 2. Relevant Studies and Data on Roma and Gender in the Western Balkans World Bank studies reviewed include: • World Bank Country Gender Assessments for all countries, when available: Bosnia and Herzegovina (World Bank 2015d), North Macedonia (World Bank 2013a), Serbia (World Bank 2016c). • Policy notes on Roma female access to economic opportunities: Bosnia and Herzegovina (World Bank 2017a), North Macedonia (World Bank 2017b), and Serbia (World Bank 2017d). • Qualitative studies on Roma women in Bosnia and Herzegovina (World Bank 2016d) and Serbia (Majumdar and Woodhouse 2019). Non-World Bank relevant studies include: • Browne (2017), which analyzes gender gaps in general in the Western Balkans, with a specific focus on Roma based on data from the 2011 UNDP-WB-EC Survey. • Cukrowska and Kóczé (2013) explore the gender angle of the 2011 UNDP-WB-EC Regional Roma Survey results for the Western Balkans. • O´Higgings (2013) explores the persistence and driving factors of the existing gender wage gaps among the Roma population in Central and South East Europe. • UNWomen (2015) explores the reasons that prevent Roma women from joining the labor market in Serbia through qualitative methods. • UNICEF (2014) provides insights on the rights of Roma girls in Bosnia and Herzegovina, North Macedonia, and Serbia. • UNICEF (2017) concentrates on the phenomenon of child marriage among the Roma in Serbia. • Hughson (2014) studies gender gaps among Roma populations in Bosnia and Herzegovina. • Kóczé and Raluca (2009) explore the gaps in research and policy that takes into account the intersectionality between ethnicity and gender in Europe. • Milenković (2018) presents the results of the 2017 unweighted UNDP-WB-EC survey on gender- based violence in Montenegro, Albania and North Macedonia together with those of field visits. Sources of data include: • 2011 UNDP-WB-EC Regional Roma Survey in Eastern European countries – including all Western Balkan countries, except Kosovo. • UNICEF MICS in all countries on Roma populations: MICS 2013 Montenegro, MICS 2011 North Macedonia, MICS 2014 Serbia, MICS 2013-2014 Kosovo, MICS 2011-2012 Bosnia and Herzegovina. • CARE (2011) survey – conducted by CARE in Serbia to explore the educational outcomes and social inclusion of Roma girls. All these studies and data already highlight the persistence of important vulnerabilities among Roma women in the region across the three major dimensions of agency, endowments and economic opportunity. Gender gaps are evident within the Roma population living in these countries even in areas where, on average, a certain degree of equality has been achieved among the non-Roma population, such as education and health. A recent report that explores gender gaps in the region indicates that all structural disadvantages experienced by women are experienced doubly by Roma women, and there appear to be no indications of change over the last decade (Browne 2017). 11 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS According to this literature, particularly large gaps can be observed with regards to agency. As will be seen in further detail in the section on agency below, Roma women persist to hold and be bounded by traditional and patriarchal social norms on their roles vis-à-vis those of men. The public and social participation of Roma men is already compared to those of non- Roma men. It can thus be expected that Roma women, who are most often expected to stay in the private sphere, taking care of the household and the family, would be even further excluded from public life. Clear indications of the very limited agency of Roma women across countries are provided by the higher prevalence and tolerance of gender-based violence, even by women themselves, and the high incidence of child marriage, which remains a widespread practice. This previous evidence suggests that gender gaps are also noticeable with regards to endowments, particularly in education. Roma males in Western Balkan countries spent on average 6.7 years in education and Roma females 5.6 years, compared to 10.9 among non-Roma males and 10.7 among non-Roma females (Cukrowska and Kóczé 2013). Around 28 percent of Roma women aged 16 to 64 have no formal level of education compared to 18 percent of Roma men and 2 percent of non- Roma women. In North Macedonia, Bosnia and Herzegovina and Serbia, Roma girls start to drop out of school after age 12, and fewer than 80 per cent of Roma women are literate (UNICEF 2014). In Bosnia and Herzegovina, for instance, the literacy rate for young Roma women aged 15-24 is estimated at around 69 percent compared to almost 99 percent for other non-Roma women. Only 47 percent of Roma girls are enrolled in primary school and almost 80 percent of those do not finish it (World Bank 2015). Differences do not seem to be explained by performance or aspirations, but by persistent barriers mostly associated with the prevalence of patriarchal norms. Roma girls tend to do better in school than Roma boys in Serbia (CARE Serbia 2011), which is attributed to their learned obedience in a patriarchal society (Cukrowska and Kóczé 2013, Browne 2017). In addition, a review of aspirations shows that Roma women do not have significantly different aspirations and do not assign significantly different values to education than Roma men (Browne 2017). According to a survey in Serbia (CARE 2011) specific barriers for girls in the Roma community to access education appear to exist: increased risk of peer violence against them, violence by school personnel, and most critically “traditional” gender norms that promote early marriage and responsibility over housework and family rather than women’s education (CARE Serbia 2011). Indeed, Serbian teachers noted that early marriage was a big problem for Roma girls (Cukrowska and Kóczé 2013, Browne 2017). Differences in health outcomes are also noteworthy. The infant mortality rate in Roma settlements in Serbia is estimated to be more than twice as high as among non-Roma, at 13 to 1,000 live births in 2012 (World Bank 2016c). Over half of Roma women aged 16-50 living in settlements near Belgrade, Serbia, were found to be undernourished. Almost all women in Roma settlements around Belgrade smoke tobacco, many beginning at the age of 11 or 12, and 71 percent of Romani women taking part in the 2011 survey have declared experiences of ethnic discrimination from medical staff (European Union 2013). Roma, Ashkali, Egyptian and Gorani women and men had among the most geographic, financial and cultural barriers to accessing healthcare in Kosovo (Browne 2017). UNICEF (2014) reports that Roma women have skilled birth attendance in Bosnia and Herzegovina, North Macedonia, and Serbia at the same rate as non-Roma (99 per cent). However, Cukrowska and Kóczé (2013) state that Roma women are 12 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS more likely to give birth outside a hospital, and that some countries have a very low rate of births attended by a professional: 23 percent in North Macedonia, 18 percent in Bosnia and Herzegovina, 16 percent in Serbia, and 16 percent in Montenegro (Browne 2017). This literature also emphasizes that Roma women remain largely excluded from economic opportunities. Almost 65 percent of Roma women are not participating in the labor market at all, and only 15 percent are formally employed. As a result, most Roma women are financially dependent on men (Cukrowska and Kóczé 2013, Browne 2017). More than 80 percent of Roma women are unemployed in Bosnia and Herzegovina (World Bank 2015d). Unemployment is estimated to be twice as high among the Roma and significantly higher among Roma women than among the non-Roma population in North Macedonia and Serbia (World Bank 2015d). Serbian Roma women most often do not work outside of the house, and when they work, they do so in low quality jobs (i.e., informally as cleaning ladies in non-Roma houses in Vranje). In most Central and South East European countries the gender wage gap is significantly larger among Roma compared to non-Roma (O´Higgins 2013). Although Serbia’s civil code protects women’s right to property, discriminatory traditions and social norms undermine this and other basic rights, especially among the Roma population (World Bank 2017d). The following sections will confirm and further elaborate on the findings presented in this section, drawing extensively from the more recent 2017 Regional Roma Survey and a complementary qualitative survey conducted in Serbia. The purpose is three-folded: to assess whether and how much the status of Roma women is more vulnerable than that of mainstream Roma men, to evaluate whether and to what extent the situation of Roma women is even more vulnerable than that of mainstream women, and to identify recommendations on how to properly address the challenges that affect these populations. It must be noted in this regard that comparisons between the 2011 RRS and the 2017 RSS must be make with caution given changes in survey and sampling design between the two survey rounds. 3 Roma Gender Profile and Determinants of Gender Gaps among the Roma Population This gender profile reviews gender inequality based on the framework developed in the World Development Report 2012, Gender, Equality and Development. The framework revolves around the notion that gender equality is essential to achieving sustainable poverty reduction and shared prosperity (see figure below). The economic costs of the misallocations entailed by gender gaps are large – around US$160 trillion according to recent estimations (Wodon and de la Briere 2018). On the contrary, gender equality leads to multiple positive economic outcomes including increased productivity and more inclusive institutions and policies (World Bank 2012d). The framework assumes that gender gaps are the result of the interplay between households, markets and institutions, across three main areas: agency, endowments and economic opportunity. In the case of Roma women, the interaction between markets, institutions, and households presents particular characteristics when compared to mainstream populations, which combine with general conditions to make the situation of Roma women even more vulnerable. 13 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Figure 1. WDR Framework to Assess Gender Inequalities Source: (World Bank 2012d) Many factors have contributed to the key challenges that marginalized Roma females in the Western Balkans face today across the main areas of agency, endowments and economic opportunity. Agency refers to the capacity of Roma women vis-à-vis men to make decisions on their own lives, and act on these decisions. It is not only important in and of itself, but also because it helps to shape outcomes across all other areas. Endowments focuses on the accumulated investments in education and health for Roma women compared to Roma men. Differences in this area will lead to gaps in the potential of both groups to participate in economic and social life over time. Lastly, economic opportunity makes references to Roma women´s access to employment, productive assets and entrepreneurship on an equal footing with Roma men. The situation of mainstream women across all areas is already particularly vulnerable and important gaps can be identified. 3.1 Gaps In And Constraints to Roma women´s Agency and Voice The agency of women in the Western Balkans appears to be limited overall;4 it can thus be expected that this would also be the case of Roma populations. The lack of capacity of women in general, and of Roma in particular, to decide on aspects that affect their lives and act on those decisions, is reflected in their lower levels of political representation, the high prevalence of gender-based violence, or the spread nature of practices such as girls´ marriage. These are underpinned by traditional and patriarchal norms that constrain women, and especially Roma women, to the role of housewives, mothers and wives. The gaps in agency are manifested in the high levels of domestic violence and child marriage registered, their minimal representation in democratic institutions and their weak decision-making capacity within the household, and, ultimately, in the widespread nature of patriarchal and discriminatory social norms on their role vis-à-vis that of men leading for instance to gendered patters of time use. 4See Browne (2017) based on Petričević (2012), UNDP (2016), KGSC (2017), Hughson (2014), Drezgi ć (2010), Djikanovic et al. (2013), Božanić (2016), Avlijaš et al. (2013). 14 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Domestic violence is perceived as a common problem in the region (Browne 2017) and thus can also be expected to affect Roma women to a larger extent. Intimate partner violence is socially accepted as a legal exercise of men’s power in many Roma communities in EU countries (European Union 2013). Different studies in Bosnia and Herzegovina suggest that Roma women continue to suffer discrimination and domestic violence, with little improvement in their situation since the end of the conflict (Hughson 2014, World Bank 2015d, Browne 2017). Moreover, Roma women are more reluctant to report the occurrence of episodes of violence due to fear, the existing community dynamics, and their distrust of institutions. In addition, Roma women may not have health insurance enabling them to access support (Hughson 2014). Without birth certificates, they are also more vulnerable to trafficking (Hughson 2014, World Bank 2015). Indeed, Roma are estimated to represent between 50 and 80 percent of trafficking victims in Eastern European countries (European Union 2013). A recent study on this issue confirms the heightened vulnerability to violence of female Roma across these countries. Despite the lack of up-to-date official data on the incidence of gender-based violence in these communities, evidence from the 2017 RRS and qualitative work in Montenegro, North Macedonia and Albania (Milenković 2018) suggest that most Roma women who are victims of violence do not report it and tend to try to solve these situations with the mediation of the family and in-laws. The rates of conviction of perpetrators in any case are very low. Legislation is poorly implemented and institutions for the protection and assistance of these women do not function properly and are subject to disruptions with the political cycle. Roma girls are especially susceptible to early marriage in Western Balkan countries. Arranged and child marriage are common practice across European countries with Roma populations, and accepted by Roma women (European Union 2013). A UNICEF study of Roma in Bosnia and Herzegovina, North Macedonia, and Serbia (UNICEF 2014) shows that 15 percent of Roma women are married before age 15, and around 50 percent before they turn 18 years old. Many girls who are married young also have children early: 40 percent of 15 to 19-year old Roma girls in Serbia had already had a live birth or were pregnant with a first child (Browne 2017). A qualitative study in Serbia (Majumdar and Woodhouse 2019) conducted by the World Bank argues that the high prevalence of early marriage is partly related to the customarily required virginity of girls.5 The lack of motivation for further education, primarily among girls, makes marriage seem desirable because there is no alternative model (UNICEF 2017). Early marriage is understood as harmful by Roma families in Serbia, but is still supported by the community as a way to control adolescent sexuality (Cukrowska and Kóczé 2013, and CARE Serbia 2011 in Browne 2017). 5 In some parts of Serbia (Novi Bečej), for instance, if a bride is found not to be a virgin, common-law court is still summoned to determine that the bride’s family should return the money invested in the wedding to the family of the groom. 15 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Evidence from the 2017 UNDP-WB-EC Figure 2. Child Marriage is Common Among Roma Women Ages 20-49 Regional Roma Survey confirms that child marriage is more common among Percent of women aged 20–49 years who first married Roma than among non-Roma before the age of 18 populations living in close vicinity. The share of Roma girls aged 20-49 who are married before they turn 18 years old 31 29 29 amounts to as much as 31 percent in Albania, 29 percent in Montenegro and Percent 21 19 Serbia, and 21 percent in North 18 Macedonia (Figure 2). Even in the two 9 countries that register the lowest shares, 4 6 4 5 4 almost a fifth of all 20-49 girls were married before they turned into adults. ALB BIH MNE MKD SRB KSV Roma Non-Roma These percentages are much higher than those observed within non-Roma populations, reflecting the dimensions of Source: World Bank estimates based on weighted 2017 UNPD-WB-EC this phenomenon among the Roma Regional Roma Survey data. Note: Child marriage refers to women who were first married before the community in the region. age of 18. The representation and participation of women in public life and decision making appears to be minimal. The very low level of Roma civic and political participation suggests that the interests of Roma communities – both men and women – are not represented in local decision- making processes and policy-making mechanisms (Cukrowska and Kóczé 2013). It can therefore be expected that the public profile of Roma women is almost inexistent. Indeed, the number of female Roma that make it to the Parliament or local political structures is almost negligible across countries, as for instance indicated by survey data from 2004 (Cukrowska and Kóczé 2013). There are reports that Roma women only become visible in some countries (e.g., North Macedonia) over the electoral periods.6 Roma female activists in civil society also face particular challenges. According to Oprea (2015), when female Roma activists in Romania criticize internal practices, ¨they are portrayed as traitors to their communities¨ (gadjikanime or Westernized), and therefore risk ostracism, being forced to choose their race over their gender, as if both were mutually excluding. Previous evidence from the 2017 Regional Roma survey also suggests that Roma women have less bargaining power and voice in financial decision making within the household. Regardless of their employment status, Roma women are less likely to report that they have money of their own than their neighboring non-Roma counterparts (Figure 3 Panel a and b). More importantly, this holds true when controlling for employment, marital status, and age, suggesting that among Roma households, women have less agency and that intrahousehold dynamics tend to favor men more so than in neighboring non-Roma households. When women who report having money of their own are asked if they decide how to spend some or all of it, most Roma women do report that they do. The proportions, however, are somewhat lower for married women, especially in North Macedonia (Figure 4). As such, between 13 percent (in 6 See http://romalitico.org/new/index.php/content/item/30-roma-women-between-voice-and-voicelesness-in-politics-in-the- republic-of-macedonia and Cukrowska and Kóczé (2013) 16 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Serbia) and 46 percent (in North Macedonia) of married Roma women have no control over money that they consider to be their own. However, the data do not show a gap with respect to their neighboring non-Roma counterparts in this regard except for in Montenegro. Qualitative evidence from Serbia shows a more nuanced picture in terms of intrahousehold dynamics in which both men and women report taking decisions over household finances; however, men tend to have a strong say regarding decisions around child marriage and preserving women’s virginity prior to marriage (Robayo-Abril and Millan 2019). Figure 3. Roma Women Are Less Likely to Report They Have Money of Their Own, Even If One Controls for Employment a. Non-working women b. Working women 80 100 60 80 Percent Percent 60 40 40 20 20 0 0 ALB BIH MKD MNE SRB KSV ALB BIH MKD MNE SRB KSV Roma Non-Roma Roma Non-Roma Source: World Bank estimates based on weighted 2017 UNPD-WB-EC Regional Roma Survey data. Figure 4. Most Roma Women Report They Decide How to Spend Some or All Their Own Money a. Married women b. Unmarried women 100 100 80 80 Percent Percent 60 60 40 40 20 20 0 0 ALB BIH MKD MNE SRB KSV ALB BIH MKD MNE SRB KSV Roma Roma Non-Roma Source: World Bank estimates based on weighted 2017 UNPD-WB-EC Regional Roma Survey data. Traditional and patriarchal values and norms prevail among Roma communities in the Western Balkans. Indeed, Roma women are generally subordinated to men within the Roma patriarchal family system (European Union 2013). As an example, Roma women in these countries continue to hold traditional values about sexual health and maternity: 31 percent of Roma men and 33 percent of Roma women prefer to get their daughter married before she completes basic education to make sure she does not start her sex life before marriage 17 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS (Cukrowska and Kóczé 2013), while 51 percent of Roma men and 54 percent of Roma women find it unacceptable for a woman to lose her virginity before getting married (Browne 2017). Based on the 2017 RRS, between around one third of Roma respondents in Bosnia and Herzegovina, Serbia and North Macedonia and almost three fourths in Kosovo consider the arranged marriage of boys acceptable. These figures are much lower among non-Roma respondents. The share of Roma who consider the arranged marriage of girls acceptable is even higher across countries with the exception of North Macedonia. The most extreme are the cases of Kosovo (where 75 percent of respondents believe that girl marriage is acceptable) and Albania (at 64 percent). Again, the share of non-Roma respondents that report this belief is much lower, between 41 percentage points in Montenegro and 8.75 in Kosovo (see Figure 5). Figure 5. Child marriage is considered socially acceptable among a large share of the Roma population; marriage among Roma girls is more likely to be acceptable than among Roma boys a. Boys b. Girls 80 80 75.0 70 70 64.5 60 60 55.0 50 50 Percent 40 Percent 40 34.1 31.7 38.3 30 30 20 20 10 10 0 0 ALB BIH MNE MKD SRB KSV ALB BIH MNE MKD SRB KSV Roma Non-Roma Roma Non-Roma Source: World Bank estimates based on weighted 2017 UNPD-WB-EC Regional Roma Survey data. Figure 6. High tolerance of gender-based violence among Tolerance of gender-based violence is Roma communities comparatively high among Roma populations. As an example, 43.5 Justify Violence by Husband Under Certain Circumstances percent of Roma women aged 15-49 in 70 65 Bosnia and Herzegovina believe that it is 60 50 41.1 43.5 acceptable for a man to beat his wife or 37 Percent 40 partner compared with only 4.8 percent 30 25.4 of the general population; the share of 20 14.5 Roma women who justify violence by 10 2.7 3.8 4.8 their husbands is twice that of men 0 (World Bank 2015d, MICS 2012-2013). Montenegro Macedonia Serbia Kosovo BiH (MICS (MICS (MICS (MICS (MICS 2011-2012) This is also the case in Kosovo (65 2013) 2011) 2014) 2013-2014) percent of Roma women compared to Roma women Non-Roma women 38.7 percent of men report accepting this form of violence). Many more Roma Source: UNICEF MICS than non-Roma women also report holding such belief in North Macedonia, Montenegro, Kosovo and Serbia, based on the MICS (see Figure 6). The 2017 RRS allows to analyze perceptions of violence against men or women among the whole marginalized Roma population (not only Roma females). According to the 18 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS RRS, between 8 percent of Roma respondents in North Macedonia to close to 41 percent in Kosovo accept a ¨slapping husband¨. These shares were systematically lower among non-Roma respondents. Interestingly, gender violence against husbands is less tolerated among the Roma community. The gender role of a woman as a housewife, mother and wife is imposed on girls from early childhood, denying them the opportunity of pursuing different life choices. Within Serbian households, for instance, there is a clear gendered division of labor, with women bearing the brunt of household and care related chores (Majumdar and Woodhouse 2019). According to a UNICEF study on child marriage in Serbia, in Roma communities the traditional division of gender roles is almost completely preserved: men deal with the public and women with the private segment of life (UNICEF 2017). A study by the World Bank on gender norms and agency in twenty countries confirms this trend (World Bank 2013b). Our analysis of the 2017 Roma Regional Survey data on time use confirms the clear gender division in tasks within the Roma community across all countries. Conditional on labor market status, wives generally dedicate more time to home production than husbands. Albania, where women dedicate twice the amount of time used by men to home production, presents the most extreme case. Moreover, husbands tend to participate more actively in the labor market, and the proportion of husbands with jobs is higher than the proportion of wives. In Montenegro and Kosovo, indeed, only 1 percent of female respondents had a job. Albania is the country where a higher share of wives has a job - about 13 percent (see Figure 7). Figure 7. Time Use among Roma Wives and Husbands Albania Husbands Wives 100% 100% 50% 50% 0% NP UE PT FT 0% NP UE PT FT Leisure Work Home production Leisure Work Home production Bosnia and Herzegovina Husbands 100% Wives 100% 50% 50% 0% 0% NP UE PT FT NP UE PT FT Leisure Work Home production Leisure Work Home production Montenegro 19 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Husbands Wives 100% 100% 50% 50% 0% NP UE PT FT 0% NP UE PT FT Leisure Work Home production Leisure Work Home production North Macedonia Husbands 100% Wives 100% 50% 50% 0% NP UE PT FT 0% NP UE PT FT Leisure Work Home production Leisure Work Home production Serbia Husbands Wives 100% 100% 50% 50% 0% NP UE PT FT 0% NP UE PT FT Leisure Work Home production Leisure Work Home production Kosovo Husbands Wives 100% 100% 50% 50% 0% 0% NP UE PT FT NP UE PT FT Leisure Work Home production Leisure Work Home production Source: World Bank estimates based on weighted 2017 UNPD-WB-EC Regional Roma Survey data. Note: NP= Not participating; UE= Unemployed, PT=Employed Part-time, FT=Employed Full-Time 20 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Discrimination against Roma Figure 8. Self-Reported Discrimination by Roma Men and women also appears to be Women comparatively high based on self- Self reported discrimination in the past 12 months due to reported perceptions. Around 11 ethnic origin (% Population) percent of Roma women reported 45 having experienced gender-based 40 discrimination in society, as opposed 35 Percent to 5 percent of non-Roma women in 30 25 Western Balkan countries (Cukrowska 20 and Kóczé 2013 in Browne 2017). 15 Self-reported discrimination data from 10 the last round of the RRS survey also 5 0 indicates that female Roma tend to ALB BIH MNE MKD SRB KSV perceive higher levels of Male Female discrimination across countries. The difference in self-reported Source: World Bank estimates based on weighted 2017 UNPD-WB-EC discrimination is as large as 10 Regional Roma Survey data. percentage points in Bosnia and Note: Child early and forced marriage refers to women who were first married before the age of 18. Herzegovina, followed by 8 percentage points in Montenegro (Figure 8). The only country where, interestingly, the levels of self-reported discrimination are higher among men is North Macedonia. In the rest the level of self-reported discrimination is only slightly higher among Roma women than men. The poor agency of Roma women across Western Balkan countries is a major factor behind the existing gaps in outcomes in education, health and economic opportunity. Roma women in the countries of study have limited agency, as refected by the higher incidence of violence and child marriage among them, their almost non-existent decision making capacity in public and private spheres, and the high prevalence of social norms that constrain them to the role of mothers and housewives. The weak ability of Roma women to decide on their own lives is likely to translate into large differences in access to education and health services, and ultimately in their labor market engagement. The next section will analyse the status of Roma women vis-à-vis Roma men and non-Roma women with regards to investments in basic endowments. Both their agency and accumulated human capital will in turn determine the capacity of women in the region to generate income, which is covered in section 3.3. 3.2 Access and barriers to Health and Education Access and barriers to health services In several Western Balkan countries, there is a high prevalence of unmet needs for medical care among Roma females. The gender gaps in self-reported unmet needs for medical care among the Roma population aged 16 and above are relatively large and statistically significant in almost all countries, except for Albania and Kosovo (Figure 9 Panel a). Indeed, in Bosnia and Herzegovina, 40 percent of Roma females declare having unmet healthcare needs compared to only 11 percent of Roma males – a 16 percentage point difference. In North Macedonia, Montenegro and Serbia, the gender gap in self-reported unmet healthcare needs reaches 11 percentage points, 9 percentage points, and 7 percentage points, respectively. On the other hand, 21 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Albania and Kosovo exhibit the smallest gender gaps: 3 percentage points and 2 percentages points, respectively, which are not statistically different from zero. Barriers to accessing health services include cost, distance, waiting times, lack of cultural sensitivity and discrimination. Figure 9. High incidence of unmet needs for medical care; in some countries, a large share of the gap remains unexplained , Year 2017 a. Self-reported unmet need for medical care, % of b. Blinder Oaxaca decomposition of the population ages 16+ percentage point gender gap in self-reported unmet need for medical care: explained versus unexplained 40 10 35 Percentage Points 30 0 -3 -2 Percent 25 -7 -10 -9 20 -11 15 -16 10 -20 5 -30 0 ALB BIH KSV MKD MNE SRB ALB BIH MKD MNE SRB KSV Roma Male Roma Females Explained Unexplained Total c. Decomposition of the explained component Percentage points 10 Percentage Points 5.1 0 -3.6 -6.3 -7.0 -6.8 -10 -10.3 -20 ALB BIH KSV MKD MNE SRB Sampling variables Socio-demographics Employed Documentation Health indicators Living conditions Assets Constant Total Source: World Bank estimates based on weighted 2017 UNDP-World Bank-EC Regional Roma Survey data. Gender gaps are only statistically significant in Bosnia, North Macedonia, Montenegro and Serbia. Note: Variables included in each category are as follows: Sampling variables: urban residence and Roma concentration in settlements (greater than 40 percent). Socio-demographic variables: household head, age, legal marriage, educational attainment, household size, number of children in the household, number of elders in the household. Employment status: employed during the reference period of the 2017 RRS. Documentation: possession of an identity card and status as an internally displaced person (IDP). Health indicators: self- perceived health status, use of preventive health care services, health insurance coverage, and disability. Living conditions: number of rooms in the dwelling, access to piped water, and access to toilet inside the dwelling. Assets: computer, internet, color tv, washing machine, and car. 22 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Most of the gender gap in self-reported unmet needs for medical care remains unexplained in Bosnia and Herzegovina and North Macedonia. Blinder-Oaxaca decompositions7 show that in the countries with the largest gender gaps in self-reported unmet healthcare needs - Bosnia and Herzegovina and North Macedonia - a large proportion of the gender gap remains unexplained after controlling for individual and household characteristics, suggesting that most of the gender gap can be attributed to differences in the returns to the individual and household characteristics or unobserved variables that are not included in the model due to the unavailability of information (e.g. access to health care services, lifestyle, or discrimination, among others). Indeed, and based on a recent assessment (World Bank 2019), health service provision in Western Balkans remains limited, and especially for the most vulnerable groups. The high out-of-pocket expenditures required leave many of the poorest households in the region without effective access to basic health services. The coverage of health insurance is generally low, which increases the risk of imporverishment due to catastrophic health costs. The lack of recent reforms to the health system has led to special difficulties and poor outcomes overall. This is particularly the case with regards to maternal and early child health. As an example, the number of women dying in birth in Albania is twice that observed on average in the EU, while in North Macedonia child mortality rates are three times those registered on average in the EU. Child stunting remains a challenge in Serbia, Kosovo and North Macedonia, and still many vulnerable children are not vaccinated against measles, for instance, across these countries. Roma, and Roma women and girls in particular, are likely to be especially affected by these challenges. On the contrary, in Montenegro and Serbia, all the gender gap can be explained by differences in observed household or individual characteristics - such as sex of head of household, age and children (see Figure 9 Panel b). In Montenegro, the explained component of the gender gap in unmet healthcare needs is mainly driven by socio-demographic variables, namely being the head of the household, age, and number of children in the household. Male household’s heads are less likely to report unmet healthcare needs than female household’s heads. Besides, age is positively correlated with the probability to report unmet medical needs for Roma males, but it is not statistically significant for Roma females. In addition, the number of children in the household increases the probability of reporting unmet healthcare need for Roma females, but it is not statistically significant for Roma males. Indeed, 16 percent of Roma females (against only 6 percent for males) declare not being able to consult a doctor or medical specialist since they could not go due to work or family matters, such as taking care of children. Employment status and health insurance coverage also play an important role in explaining the gender gap in Montenegro, while self-perceived help is particularly important in Serbia. Being employed is negatively correlated with the probability to report unmet medical needs; however, the employment rate among Roma females is extremely low in Montenegro (only 3 percent). Lower levels of health insurance increase the probability to report unmet needs for medical care and support the findings that the main reason for reporting unmet 7 The determinants of the gender gap are calculated using the Blinder-Oaxaca decomposition – a methodology developed by Blinder (1973) and Oaxaca (1973)- that allows to decompose the gap between an explained component that accounts for differences in observable characteristics (endowments), and an unexplained component that accounts for differences in the returns to the characteristics. 23 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS healthcare needs is affordability. On the other hand, in Serbia, a large proportion of the gender gap can be explained by differences in socio-demographics characteristics and self-perceived health status. In fact, better self-perceived health status reduces the probability to report unmet medical needs for Roma males, but it is not significant for Roma females (Figure 9 Panel c). Access and barriers to education services Access to education is critical for girls and women’s opportunities in the Western Balkans. The years during the early childhood period from the conception through birth to eight years of age are key to the complete and healthy cognitive, emotional and physical growth of children.8 Successful pre-primary integration has several economic and social paybacks in terms of reduced dropout and repetition rates, improved school achievements, greater adult productivity and improved social and emotional behavior.9 Roma children experience a cycle of inequalities starting from their birth due to language barriers, parents’ limited education, poor investments in education and household status. In addition, discrimination against Roma children starts from the early years within their locality and marks a life-long impact. Pre-primary enrolment is low for EU standards in Western Balkan countries except Albania. In Bosnia-Herzegovina, Montenegro, Kosovo and North Macedonia pre-primary enrolment rates are close to the average for low-income countries and much below that of countries of similar income levels. The quality of such services appears to be also poor; as an example, Albania, the country where coverage seems to be higher, no national quality standards exist regarding childcare for 0-3 years old children. This is particularly challenging for vulnerable children including Roma, for whom access to quality early child education appears to be especially important (World Bank 2019). Indeed, enrollment in pre-primary education is very low among Roma communities, affecting both Roma boys and girls; no significant gender gaps are observed in any of the countries except Bosnia and Herzegovina. As evidenced in the World Bank report on Roma exclusion (Robayo-Abril and Millan 2019), in nearly all countries, Roma children are significantly less likely to be enrolled in preprimary school than the national population and even their non-Roma neighbors, also vulnerable populations. Roma girls do not seem to be disproportionately affected by low enrollment rates when compared to Roma boys. Gender gaps in enrollment in early childhood education are not statistically significant in any of the countries, except in Bosnia and Herzegovina, where the preprimary enrollment rate is just 1 percent among Roma girls, versus 6 percent among Roma boys. In Kosovo, the gap is only 3 percentage points (and not statistically significant), while the gender gap among non-Roma neighbors is 22 percentage points (Figure 10). 8 UNICEF defining Early childhood development. Unicef. "Early Childhood Development: The key to a full and productive life." ed (2014). 9 World Bank, 2018. Roma Gender Qualitative Study – Literature Review. 24 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Providing universal access to education Figure 10. Gender Disparities in Pre-primary and ensuring the completion of Education Are Not as Large as in Other Levels, Since compulsory10 school for all Roma girls Enrollment is Low Among Both Roma Girls And Boys and boys is one of the key areas of Net preprimary enrollment (% children ages 3-5) concern: compulsory school is not universal among Roma children, with 60 both Roma boys and girls equally 50 affected by low enrollment. While primary 40 Percent and lower-secondary education are nearly 30 universal for the majority population in all 20 countries, they are not for Roma children. For instance, in Montenegro only 62 percent 10 of marginalized Roma children aged 7 to 15 0 were enrolled in school in 2017. Across the ALB BIH MKD MNE SRB KSV region enrollment levels are quite low for Roma Male Roma Females both boys and gilrs, ranging from 61 percent Sources: World Bank estimates based on weighted 2017 Regional Roma Survey data. National data: most recent available: BIH MICS 2011- among Roma girls in Montenegro, to 85 2012, KSV MICS 2012-2013, MNE MICS 2013, MKD MICS 2011, SRB percent in Serbia. Differences are not MICS 2014. Recent data for Albania is not available (latest MICS was collected in 2005). statistically significant. Note: Gender gaps are statistically significant at the 10 percent level only in Bosnia and Herzegovina. Among older cohorts, gender gaps in education are narrower since few Roma boys and girls complete upper-secondary; completion of tertiary education is extremely low regardless of gender. Completion rates in upper secondary among Roma females ages 22–25 show a wide range across countries, from only 2 percent in Montenegro to 32 percent in North Macedonia (Figure 11).11 Gender gaps in favor of boys in upper-secondary completion are significant in Kosovo and Serbia, at 12 and 16 percentage points, respectively. Completing secondary education is associated with higher labor force participation, higher earnings, higher levels of participation in civic life, and better health (Levin and Belfield 2007, Pleis 2010). Consequently, policies aimed at improving school completion rates and reducing the number of students dropping out of school are relevant, especially in the context of Roma, among whom upper-secondary completion rates are relatively low. Tertiary completion, increasingly important in the labor market, ranges from 0 to 4 percent among Roma females (ages 26–29.) There are no statistically significant differences with Roma males, among which tertiary completion ranges between 0 and 2 percent (Figure 12). 10 In this note, compulsory education is understood as primary and lower-secondary (ISCED levels 1 and 2); in all countries, compulsory education covers up to lower-secondary, except in North Macedonia, where it covers up to upper-secondary (ISCED level 3). In North Macedonia, upper-secondary education (ISCED 3) has been compulsory since 2008. General secondary education covers a period of four years, whereas vocational secondary education lasts from two to four years, depending on the specific program. 11 ISCED level 3. 25 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Figure 11. The Gender Gap is Wide in Upper Figure 12. The Tertiary Education Completion Secondary Completion Only for Serbia and Kosovo Rate is Extremely Low Among Roma Males and Females Upper secondary (ISCED level 3) completion rate (ages 22-25) Tertiary education completion rate (ages 26-29) 40 10 9 8 30 7 Percent 6 Percent 20 5 4 3 10 2 1 0 0 ALB BIH MKD MNE SRB KSV ALB BIH MKD MNE SRB KSV Roma Male Roma Females Roma Male Roma Females Source: World Bank estimates based on weighted 2017 UNPD-WB-EC Regional Roma Survey data. For upper secondary, gender gaps are statistically significant at the 10 percent level in Serbia and Kosovo. For tertiary, gender gaps are not statistically significant at the 10 percent level in any country. However, Roma females ages 18-21 face significantly large gender gaps in completion of compulsory education. Completion rates in Figure 13. Large Gender Gaps in Compulsory compulsory education among Roma youth aged Education Completion (ISCED Levels 1 & 2) 18-21 years old in the Western Balkans range from 34 percent to 70 percent, being Montenegro Compulsory education completion rate, % of the country with the lowest rate and North children ages 18–21 80 Macedonia the country with the highest rate. The 70 completion rates in compulsory education are 60 especially low among Roma girls, ranging from Percent 50 30 percent in Montenegro to 63 percent in North 40 Macedonia and Serbia. The gender gap is 30 relatively large and statistically significant in 20 three out of the six countries. Large and 10 statistically significant gender gaps are observed 0 ALB BIH MKD MNE SRB KSV in Bosnia and Herzegovina (14 percentage Roma Male Roma Females points), North Macedonia (14 percentage points), and Kosovo (12 percentage points). In Albania, Source: World Bank estimates based on weighted 2017 UNPD- Montenegro, and Serbia, the gender gaps range WB-EC Regional Roma Survey data. Gender gaps are from 7 to 9 percentage points, although they are statistically significant at the 10 percent level in Bosnia and Herzegovina, North Macedonia, and Kosovo. not statistically significant (Figure 13). While many factors can determine girls’ access to education, previous evidence shows that cost and social norms are significant barriers to preprimary attendance among both Roma boys and girls. Previous empirical evidence using the 2017 Regional Roma survey (Robayo- Abril and Millan 2019), shows that the majority of Roma caregivers report that they do not send their 3-5-year-old boy or girl to preschool or kindergarten either because they cannot afford to and/or because they do not see a need - either because someone in their household can care for 26 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS the child, or because they feel their child should stay with the family, or because their child is too young to attend school (Table 1). In Bosnia and Herzegovina, the country with the largest gender gap in enrollment, a higher share of parents is willing to send their girls to school if the affordability constraint is solved, compared to boys - 72 percent vs. 62 percent respectively. This is also the case in Serbia (Table 2). Eliminating tuition fees may thus allow more children, especially girls, to enroll. Table 1. Reasons for Not sending 3-to-5 years Old Children to Childcare, Preschool or Kindergarten, By Gender Bosnia and ALB KSV MKD MNE SRB Herze govina Roma Roma Roma Roma Roma Roma Roma Roma Roma Roma Roma Roma Male Female Male Female Male Female Male Female Male Female Male Female For children aged 3-5, reason of not being in pre- primary affordability 71 75 89 86 62 61 66 60 67 67 51 60 (option 1) availability 21 19 29 19 22 25 33 32 13 12 26 34 (option 2, 3, 10) don't see a need 72 62 54 55 72 71 72 72 77 74 68 73 (4, 8, 9) documents/repat riated child (11, 10 9 1 1 2 10 5 1 7 12 3 4 12) language/treatm 20 8 3 4 5 6 7 6 14 11 6 6 ent (5, 6, 7) Other reasons 4 2 0 0 0 0 0 0 4 5 6 8 (95) Source: World Bank estimates based on weighted 2017 UNDP- World Bank-EC Regional Roma Survey data. Table 2. Factors of making respondents reconsidering sending child to pre-primary: No school fees (attendance for free) ALB BIH KSV MKD MNE SRB Roma Roma Roma Roma Roma Roma Roma Roma Roma Roma Roma Roma Male Female Male Female Male Female Male Female Male Female Male Female Yes 77 74 62 72 58 55 67 46 71 58 57 63 Maybe 13 8 24 12 22 25 19 27 17 14 17 9 No 11 18 14 17 21 20 14 27 12 28 26 28 Source: World Bank estimates based on weighted 2017 UNDP- World Bank-EC Regional Roma Survey data. Additionally, structural constraints to access to education for Roma can play a critical role. Even when schooling is free of charge, going to school can be associated with several costs such as textbooks, clothes and transport; and in terms of foregone revenue that children may generate through work12. Further barriers, particularly to preschool education include: (i) insufficient and 12 World Bank, 2018. Roma Gender Qualitative Study – Literature Review. 27 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS unevenly distributed infrastructure, with fewer preschools spaces in underdeveloped, low- income (including cities’ peripheries) and rural areas and over-crowded preschools where they exist; (ii) a tendency for preschool institutions with long waiting lists to give priority to families with working parents and/or to have non-transparent admission criteria; (iii) parents’ lack of understanding of the benefits of ECEC, especially when they can take care of the child at home (e.g. if they are not working) or through relatives, and/or their dissatisfaction with the quality, sensitivity (e.g. to the specific needs of individual children) and practical aspects (e.g. hours of operation) of the service provided; and (iv) other barriers such as language and lack of personal documents and proof of citizenship, even if not officially required for enrollment (especially for Roma and internally displaced families) (World Bank 2017d). Based on the most recent RRS results, lack of affordability seems to be the most significant barrier that delays enrollment of Roma boys and girls in compulsory education and above; among girls, early marriage is also an important barrier, while among boys, the need to provide additional income for their families seems important. Across countries, close to one- half of individuals ages 6–24 who are not in school and who have, at most, completed compulsory education report that they did not attend school because of affordability reasons, namely, the cost of education or related expenses such as transport and books. A large share (especially among boys and men) also report that they needed to work for income or have found jobs; this is especially the case among boys and men in Albania (20 percent) (Table 3). A large share of girls is not attending school because of marriage, ranging from 11 percent in Bosnia to 32 percent in Serbia. Further qualitative evidence from Serbia shows that many Roma perceive that their neighborhoods are unsafe, and this affects their willingness to send children, especially girls, to school (Robayo-Abril and Millan 2019). Table 3. Many Roma Cite Cost as the Main Reason for Not Attending School; among Roma Women, Child Marriage Is Also a Factor % ages 6–24 not attending school who have only completed ISCED level 2 or less Roma, males ALB BIH KSV MKD MNE SRB Costs of education too high (for example, fees, transport, and books) 52 49 57 46 41 46 Need to work for income/found job 20 15 8 11 11 10 Did not pass entry exam or did poorly in last level 1 5 0 0 2 2 Feel sufficiently educated 1 11 6 7 9 14 Marriage 5 3 4 7 8 8 Pregnancy NA NA NA NA NA NA Other 21 18 24 29 29 20 Roma, females Costs of education too high (for example, fees, transport, and books) 46 49 49 35 40 30 Need to work for income/found job 10 11 1 7 5 3 Did not pass entry exam or did poorly in last level 0 2 2 2 1 1 Feel sufficiently educated 3 10 4 4 9 16 Marriage 21 11 19 28 18 32 Pregnancy 2 0 0 1 2 3 Other 18 16 25 23 24 16 Source: World Bank estimates based on weighted 2017 UNPD-WB-EC Regional Roma Survey data. Note: Data are for the 6–24 age-group who have only completed ISCED 2 or less. 28 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Based on a Blinder-Oaxaca decomposition analysis while in Bosnia and Herzegovina and North Macedonia most of the gender gap in compulsory education completion rate can be explained by differences in individual and household characteristics, in Kosovo a large proportion of the gender gap remains unexplained. In North Macedonia the gender gap is almost totally explained by differences in observed characteristics. In Bosnia and Herzegovina, 11 out of 14 percentage points are explained by differences in observed characteristics, which represents almost 80 percent of the total gap. On the contrary, in Kosovo, the unexplained component accounts for the highest proportion of the total gap, whereas the explained component accounts for 42 percent of the total gender gap (Figure 14 panel a). These findings show that, overall, unobservable characteristics, which may include social norms and discrimination, do not significantly affect the gender gap in compulsory education completion rate in Bosnia and Herzegovina and North Macedonia, while this might not be the case for Kosovo. Figure 14. Blinder-Oaxaca Decomposition of The Gender Gap in Compulsory Education Completion Rate Among Roma Population, ages 18-21, year 2017 a. Decomposition of the total gender gap: explained b. Decomposition of the explained component versus unexplained, Percentage points Percentage points 20 20 15 14 15 14 14 Percentage Pioints 12 10 11 10 9 8 7 5 5 3 0 0 0 -5 -10 -10 ALB BIH KSV MKD MNE SRB ALB BIH KSV MKD MNE SRB Assets Living conditions Health indicators Household head Explained Unexplained Demographics Sampling variables Total Source: World Bank estimates based on weighted 2017 UNDP-World Bank-EC Regional Roma Survey data. Gender gaps are only statistically significant in Bosnia, North Macedonia and Kosovo. Note: Variables included in each category are as follows: Sampling variables: urban residence and Roma concentration in settlements (greater than 40 percent). Socio-demographic variables: age, early marriage (before age of 18), household size, number of other employed members in the household, other youth in the household who have completed compulsory education; speaking Romani language at home: Household head characteristics: sex, age, marital status, educational attainment, Health indicators: self-reported health status and disability. Living conditions: number of rooms in the dwelling, access to public sewerage, and access to toilet inside the dwelling. Assets: computer, Internet, more than 30 books, color tv, and washing machine. The main factors driving the explained component of the gender gap in compulsory education completion rate are characteristics of the household head. Household head’s characteristics including sex, age, marital status and, particularly, educational attainment, play an important role to explain the gender gap in compulsory education completion rate in Bosnia and Herzegovina, North Macedonia, and Kosovo (Figure 14 panel b). The higher the level of education of the household head, the higher the probability that boys and girls complete compulsory education. However, in almost all countries, Roma girls are more likely to live in households where the household head has attained lower levels of education, and therefore are less likely to complete compulsory education compared to Roma boys. This could be related also 29 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS to the fact that Roma girls are more likely to live in settlements with a higher concentration of Roma population, where access to schools can be even more limited. On the other hand, differences in socio-demographic characteristics between boys and girls, especially early marriage and household size, also contribute to explain the gender gap in compulsory completion rate. Indeed, marrying before the age of 18 is statistically significant and negatively correlated with the probability of completing compulsory education. Besides, early marriage is much more prevalent among girls than boys, and consequently, young girls are more likely to drop out from education (either voluntarily or forcedly) to undertake housekeeping and/or child-rearing responsibilities. Finally, girls living in households of larger size are also less likely to complete compulsory education, since they might be expected to help with households’ chores or to take care of younger siblings or elder family members. Child marriage, which as seen before is Figure 15. Child Marriage is Correlated with Lower particularly widespread among the Educational Attainment Roma, is associated with female Distribution of educational attainment among Roma educational attainment. Indeed, and as women ages 20-49 reflected in table 3, between 32 percent in 100% 80% Serbia and 11 percent in Albania of 6-24 60% 40% years old Roma women report dropping out 20% of school due to marriage. Figure 15 shows 0% Non-early marriage Non-early marriage Non-early marriage Non-early marriage Non-early marriage Non-early marriage Early marriage Early marriage Early marriage Early marriage Early marriage Early marriage that educational attainment is higher across countries among women 20-49 years old who did not marry early as compared to those who did. Barely any women who ALB BIH KSV MKD MNE SRB married early completed ISCED 3 level, ISCED 3 complete/Tertiary complete while the share of those with incomplete ISCED 2 complete ISCED2 incomplete or below ISCED2 or below is higher among those who married before they turned 18 years Source: World Bank estimates based on weighted 2017 UNPD-WB-EC old. Therefore, delaying marriage may lead Regional Roma Survey data. to higher educational attainment among Note: Early marriage refers to women ages 20-49 who were first women. married before the age of 18. Qualitative evidence confirms that the reasons for Roma boys and girls to drop out from school early are different and related to social roles and expectations. Based on focus group discussions with Roma in Serbia, the most common reason for boys was to start working; among girls, these included child marriage, preserving sexuality, and responsibility over the household.13 These gaps are clearly underpinned by the prevalent traditional and patriarchal social norms attributing specific features and roles to masculinity and femininity in what is considered the natural and correct pathway. Among the Roma, building a masculine identity for boys is more related to working than studying, as compared to girls. 13 Quantitative evidence in the RRS also corroborates these results. Although the most commonly cited reason for dropping out or not attending school is the high cost of education, the next most commonly cited reason is marriage in the case of women and the need to work in the case of men (see Robayo-Abril and Millan 2019). 30 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS I was thinking in a different way. Ever since I was a child I wanted to work, to make something of my life. When I finished primary school, I was 15, and I wanted to work, to be my own man. And that is exactly what I did —Young Roma male (February 21, 2018, Male Focus Group, Kamendin, Belgrade. However, the educational aspirations Figure 16. The Educational Aspirations of Roma among parents for Roma boys and girls parents for Their Boys and Girls are Similar tend to be similar. Consistent with low Parental aspirations among Roma for boys and girls returns to schooling, Roma have significantly 100 lower educational aspirations for their 80 Percent 60 children in general than non-Roma, and this is 40 so across all countries (Robayo-Abril and 20 Millan 2019). However, there are no 0 Roma Boys Roma Boys Roma Boys Roma Boys Roma Boys Roma Boys Roma Girls Roma Girls Roma Girls Roma Girls Roma Girls Roma Girls significant gender differences in aspirations. Across the six countries, responses among Roma parents concerning the level of education that is considered adequate for a ALB BIH MKD MNE SRB KSV child do not differ much depending on ISCED 1 & ISCED 2 ISCED 3 Tertiary whether the child is a boy or a girl (Figure 16). Roma respondents were equally likely to Source: World Bank estimates based on weighted 2017 UNPD-WB- respond that a girl or a boy should aspire to EC Regional Roma Survey data. either upper-secondary or tertiary education, Note: Data are based on responses from a randomly selected suggesting that parents may not necessarily household member ages 16 and above. hope for their daughters to marry young and only fulfill domestic duties, while aspiring for more for their sons. The apparent inconsistency between reported parental aspirations – which are similar with regards to both girls and boys – and the behavior observed in practice is a particularly puzzling result requiring further research. It must finally be noted that the quality of educational services appears to be low across countries. Although access to affordable compulsory education at the primary and secondary level is fundamental to children´s opportunities, the quality of instructions is also a critical determinant of how much students learn. PISA results indicate that performance among students in these countries tends to be low, which may be a reflection of the quality of instruction. Learning outcomes are below the EU average in all countries but Serbia. About 70 percent of the students 15 years old and above in Kosovo and North Macedonia and 50 percent in Albania are functionally illiterate. Almost half of students in Montenegro do not attain basic proficiency in science and mathematics (World Bank 2019). Low performance is more pronounced among the most vulnerable populations such as the Roma, and among girls. As an example, the share of resilient students – those who come from the top bottom quarter of the PISA index of economic, social and cultural status for the country but perform in the top quarter of all countries – among disadvantaged students (2015) is extremely low – among the lowest in all the OECD sample in Kosovo, North Macedonia and Montenegro (2.5, 4 and 9.4 percent). Differences in performance also exist between girls and boys across countries. For mathematics, a statistically significant gap of 9 percentage points exists in Kosovo – to the advantage of boys – and Albania – to the advantage of girls. The gap 31 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS in reading is much larger and to the advantage of girls in North Macedonia (46 percent), Kosovo (36 percent) and Montenegro (34 percent) (PISA 2015). Females may be key entry points and positive agents in education interventions. Qualitative evidence conducted in Serbia suggests that while Roma girls enjoy going to school, several of the interviewees were pulled out of education early to fulfill their household duties or as a precautionary measure to preserve their virginity for marriage. Indeed, one of the stark differences between girls and boys who drop out of school is that, while the former drop out voluntarily, the latter do so involuntarily relatively more often. Hence, young mothers typically aspire to send their girl child to school longer because the mothers were more often forced to drop out, unlike the boys their age – and even if both groups dropped out at about the same time. Increasing job opportunities among young Roma, especially young Roma women, could be a powerful tool to increase educational attainment. Qualitative evidence conducted in Serbia suggests that in places where good job opportunities exist, attending school among females is considered important. The study points to the relevance of understanding the influence of social norms within wider structural constraints by arguing that, in many Roma communities, child marriage, for example, is a frequent outcome of widespread unemployment leading to discouragement to pursue further education or opportunities for better jobs. In this sense, if women see that they can have a future outside the home, they are more likely to continue in school and not marry early. 3.3 Access and constraints to Economic Opportunities Women’s access to economic Figure 17. Large Gender Gaps in Labor Force opportunities is a key challenge to attain Participation among Roma gender equality among Roma Labor force participation rates, % population ages 15- communities in the Western Balkans. 64 Roma are at a disadvantaged position in the labor market and there are wide gender gaps 100 in labor market outcomes; Roma females 80 Percent have significantly lower labor force 60 participation rates than Roma males, which combined with the high incidence of 40 unemployment have led to poor employment 20 prospects. 0 ALB BIH MKD MNE SRB KSV EU28 Roma female Roma male In all countries in the region the labor National female National male force participation rate among Roma females is extremely low, but especially so Source: World Bank estimates based on weighted 2017 Regional Roma Survey data, World Bank SEE Jobs Gateway and Eurostat. in Bosnia and Herzegovina, Montenegro, Note: Chart shows labor force participation rate among the working- and Kosovo. Roma females are much less age population (15–64). National level data are for 2016. likely to participate in the labor market than Roma males and significantly less likely to participate than the majority population as a whole (see Figure 17). Indeed, the labor force participation rates among Roma females range from 5 percent in Montenegro to 30 percent in North Macedonia. Further, the gender gaps are markedly 32 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS large and statistically significant in all countries, varying from 23 percentage points in Albania to 33 percentage points in Kosovo. Roma women with low educational attainment most often choose not to participate in the labor market. As will be seen later, this appears to be related to generally low returns. However, the few Roma women with higher educational attainment can perhaps obtain sufficiently attractive rewards; indeed, a larger share of them use this asset, although their labor force participation rates still fail to match those of Roma men with similar or even lower educational attainment levels. In Albania and Montenegro, labor force participation rates are lower among Roma women who have completed ISCED level 3 (upper-secondary) or above than among Roma men who have not completed compulsory education (Figure 18). Figure 18. Labor Force Participation Is Low among Less Well-Educated Roma Females a. Roma women b. Roma men 80% 80% 60% 60% 40% 40% 20% 20% 0% 0% ALB BIH MNE MKD SRB KSV ALB BIH MNE MKD SRB KSV ISCED2 Incomplete or below ISCED2 Incomplete or below ISCED 2 Complete ISCED 2 Complete ISCED 3 Complete/Tertiary Complete ISCED 3 Complete/Tertiary Complete Source: World Bank estimates based on weighted 2017 Regional Roma Survey data and Eurostat. Note: Panels shows labor force participation rates by educational attainment among the working-age population (15–64). Even when Roma females are active, they experience higher rates of unemployment across the region. Despite the narrowing gender gap in unemployment among Roma in recent years, the incidence of unemployment continues to be significantly higher among Roma females, and their employment prospects are still poor. Across the region, over 53 percent of economically active working-age Roma females were unemployed in 2017, compared to 41 percent of Roma males.14 Unemployment rates among Roma females ranged from only 33 percent in Montenegro to 73 percent in Bosnia, much lower than among Roma males and their non-Roma neighbors, which are already much lower than the average national employment rates and those registered in the EU28 (Figure 19 panel a). Despite recent improvements, the employment rates of Roma women remain significantly lower than those of men and non-Roma women. Gender disparities in employment among Roma have recently narrowed, but they are still wide. The employment rates among Roma females are lower than 5 percent in Bosnia and Herzegovina, Montenegro, and Kosovo; and around 11-13 percent in Serbia and North Macedonia. The gender gaps in employment rates, although smaller than the gender gaps in labor force participation rates, are still large and 14 Population-weighted regional statistics based on the 2017 Roma Regional Survey. 33 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS statistically significant, fluctuating between 13 to 23 percentage points. The largest gap is observed in Montenegro, where the employment rate of marginalized female Roma was 3 percent in 2017, compared to 23 percent among men (Figure 19 panel b). Discrepancies in employment and the largely poor outcomes of women may be related to their lower educational attainment, social norms and time use patterns, and discrimination. Figure 19. Large Gender gaps in Unemployment and Employment-to-Population Ratio among Roma, Ages 15–64 a. Unemployment rate b. Employment-to-population ratio 80 73 40 70 63 35 58 60 53 30 50 45 Percent 25 Percent 40 33 20 30 15 20 10 10 5 0 0 ALB BIH MKD MNE SRB KSV EU28 ALB BIH MKD MNE SRB KSV Roma Males Roma Females National Roma Male Roma Females Sources: World Bank estimates based on weighted 2017 Regional Roma Survey data, World Bank SEE Jobs Gateway and Eurostat. Significant gender disparities are observed among young adults not in education, employment or training (NEET), placing Roma females at higher risk of becoming socially excluded. NEET rates among young (15–24) marginalized Roma males remain high, ranging from 51 percent in Serbia to 77 percent in Bosnia. Young Roma women, with lower school attendance and employment rates, are also more likely to be NEET than their men counterparts (Figure 20 panel a). The gender gap is especially large in Albania, Montenegro, and Serbia. Roma women are more likely to be inactive, or out of the labor force, often engaged in domestic and caretaking activities, given the high incidence of child marriage and early family formation. On the other hand, men NEETs are more likely to be unemployed. When employed, Roma females are more likely to be engaged in formal employment. The informal sector is defined as a sector not covered by labor market regulations. In all countries except Montenegro, Roma females are much less likely to be informal. For instance, informal employment15 accounts for 66 percent of overall employment among males in Albania, compared to only 46 percent among Roma females. The gender gap favoring females ranges from 5 percent in Bosnia to 31 percent in Kosovo. Montenegro is the only country where females are more likely to hold informal jobs, with informality rates reaching 72 percent among females versus 56 percent among males (Figure 20 panel b). This is probably related to the fact that Roma women are more often than not confined to household work and care activities. 15 Here, we define employment jobs the proportion of employed population ages 15-64 who do not have pension or health insurance for their job, paid for either by themselves or their employer (Social Protection definition). 34 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Figure 20. High Incidence of NEETS and Low Employment-to-Population Ratio among Roma, Ages 15 – 64 a. Not in employment, education, or training (NEET) b. Informal employment (% of total employment) (ages 15-24) 100 80 90 70 80 70 60 60 50 Percent Percent 50 40 40 30 30 20 20 10 10 0 0 ALB BIH MKD MNE SRB KSV ALB BIH MKD MNE SRB KSV Roma Male Roma Females Roma Male Roma Females Sources: World Bank estimates based on weighted 2017 Regional Roma Survey data. Being Roma and being female both Figure 21. Framework for Contraints to Employment operate as mutually reinforcing for Roma Females disadvantages in the labor market. A Time and recent UN Women (2015) report that access to services: child analyzes the structural barriers to Roma and elder care labor force participation highlights their low levels of education and vocational Skills Discrimination training and discrimination by potential employers as the two main constrains that they face to access employment. Work However, when it comes to Roma Attitudes and social norms arrangements women, two additional factors are cited: patriarchal norms within the Roma Source:Own adaptation based on (Arias, et al., 2014) society, and the difficulty in accessing information and services/support. The more segregated the settlement, the more pronounced these structural barriers for women’s employment. Other potential institutional and market barriers include lack of access to childcare, limited access to information, networks and productive inputs such as credit, and limited flexible work arrangements and labor regulations – such as maternity leave provisions. Disincentives to work from taxes and social protection systems may also play a role (Figure 21). Access to effective social protection and employment services is generally constrained for the most vulnerable populations in Western Balkan countries. First, and although social assistance programs in the region are not properly targeted at the poorest or the most 35 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS vulnerable,16 the eligibility rules for some last-resort programs may be discouraging beneficiaries from seeking employment. In addition, and despite the large dimensions of the unemployment challenge across countries, spending on employment services and active labor market policies remains very limited and below the OECD average of 0.56 percent, with the exception of Serbia and Montenegro. Moreover, the existing services and programs are rather tailored to the needs of formal sector middle-class workers, while the most vulnerable – including the Roma populations and certainly Roma women– remain largely excluded from their coverage (World Bank 2019). In addition to these market factors, informal institutions including social norms and discrimination seem to be playing an important role to keep Roma women out of employment. Previous evidence (Robayo-Abril and Millan 2019) shows that in almost all countries, most of the gender gap in labor force participation is barely explained by differences in individual and household characteristics between Roma women and men. Blinder- Oaxaca decompositions carried out in this report also show that, in all countries, a significant share of the gender gaps in labor force participation and employment rates remains unexplained after controlling for individual and household characteristics. This indicates that gender gaps in labor force participation and employment rates are mostly attributable to differences in the returns to individual and household characteristics, and therefore likely to be associated with gender discrimination against Roma women or with social norms that tend to relegate women to the home. The explained component of the gender gap in labor force participation accounts for a small proportion of the total gap, especially in Albania, Montenegro, and Kosovo. In North Macedonia, Bosnia and Herzegovina, and Serbia, it represents 16 percent–28 percent of the total gender gap (Figure 22 Panel a). The explained component of the gender gap in employment rate is especially low in Bosnia and Herzegovina and Montenegro. In Albania, Kosovo, Serbia, and North Macedonia, it accounts for 12 percent-36 percent of the total gap (Figure 22 Panel b). Several studies point out that gender discrimination in the labor market in developing countries may be the result of the cultural bias of employers, but also of household decision making, especially in countries or among groups in which patriarchal social norms predominate (Vella and Farre 2013), (Dao 2014), (Sukanya 2014), (Contreras and Plaza, 2010), (Verick 2014), (Jain- Chandra et al, 2018). 16The levels of expenditure on social assistance are quite high in countries of the region. Bosnia and Herzegovina, Kosovo and Montenegro spend for instance the average observed in countries in Eastern Europe and Central Asia (where social assistance spending is the the highest). However, programs targeted at the poor account for a small share of such amount, as most social assistance is allocated to specific categories of people. The share of poor people receiving social assistance ranges from 44 percent in Bosnia and Herzegovina to 22 percent in Montenegro. The low and fragmented coverage of social assistance programs appears to constrain their effectiveness to reduce poverty in the region. 36 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Figure 22. Most of the Gender Gap in Labor Force Participation and Employment among Roma Cannot Be Explained by Differences in Characteristics between Roma Men and Roma Women a. Blinder-Oaxaca decomposition of the gender gap in b. Blinder-Oaxaca decomposition of the gender labor force participation rate among Roma gap in employment rate among Roma population, population, ages 15-64, year 2017: explained versus ages 15-64, year 2017: explained versus unexplained (Percentage points) unexplained (Percentage points) 40 33 30 28 30 Percentage Points Percentage Points 30 25 25 23 22 23 20 18 18 20 15 13 10 10 0 0 ALB BIH KSV MKD MNE SRB ALB BIH KSV MKD MNE SRB Explained Unexplained Total Explained Unexplained Total Source: World Bank estimates based on weighted 2017 UNDP-World Bank-EC Regional Roma Survey data. Note: Variables included in the Blinder-Oaxaca decomposition: urban residence, Roma concentration in settlements (greater than 40 percent), household head, age, age squared, early marriage (before age of 18), enrolled in education, highest educational attainment, number of children in de household, number of elders in the household, number of other employed members in the household, internally displaced person (IDP); speaking Romani language at home, whether the household received social assistance, whether the household received private transfers, self-reported health status, disability, number of rooms in the dwelling, access to piped water, access to toilet inside the dwelling, computer, Internet, color tv, washing machine, and car. Qualitative evidence from Bosnia and Herzegovina and Serbia confirms that Roma women face discrimination and adverse attitutes in the labor market. According to a recent qualitative study (World Bank 2016d) in Bosnia and Herzegovina many officials reported that Roma attitudes and lifestyles were the reason why they are disadvantaged in the workforce. Qualitative evidence from Serbia shows that while domestic and unpaid work is considered the domain of women, work is typically considered an activity for men. “My wife does the typically woman chores” and “why should I work if my husband is providing for me?” are common statements. Several women spoke of their limited mobility as a matter of pride and saw it as being linked to their husband’s accomplishment of the classic provider-and-protector masculinity. Women, even those working in the informal sector, seldom see themselves as working but as coping or making ends meet. Indeed anthropological studies on women’s work in the Balkans suggest that Roma women (much more than Roma men) are relying on innovative methods to make ends meet but identify these as coping or survival strategies rather than work or a good job. There is a strong desire among Roma women to have decent jobs, but they become caught up in coercive webs of appropriate gendered behavior in public spaces and within the household. I search through dumpsters, gather what’s valuable, prepare the goods and go to the marketplace on my own. Often police and communal inspectors come to the marketplace and take away my goods. I would rather do a job instead! Mop the stairs, anything! I would clean anything, just to have some source of income besides welfare, so I can use the welfare payments only for the bills. --Roma woman, 40-45 years (February 11, 2018, Women’s Focus Group, Kamendin, Belgrade) 37 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Further analysis indicates that age, education and number of children in the household matter to explain lower female labor market participation and employment rates.17 In almost all countries, age and age squared are statistically significant with an inverse quadratic structure, meaning the probability to participate (work) first increases with age and then decreases. Besides, higher levels of educational attainment are highly correlated with the probability to participate (work). When significant, being enrolled in education is negatively correlated with the probability to participate (work), especially among Roma males. Not surprisingly, the presence of young children in the household decreases the probability to participate (work) among Roma females in almost all countries, but has no significant effects among Roma males, reflecting the fact that women are the ones who are mainly responsible for child rearing.18 Family formation and care responsibilities disproportionately affect Roma women; Roma women tend to live in households with relatively more children, and as such, face additional challenges when combining employment and parenthood or care of other adult family members. Roma households are notably larger in size than non-Roma households (on average, 41 percent of households in the region have more than 4 family members compared to only 20 percent among Roma neighbors19), and youth dependency ratios are much higher (Figure 23 panel a). Notably, accessible and affordable formal childcare and eldercare services are limited across the region, as shown by recent child and elder care assessments conducted by the World Bank (World Bank 2015e, World Bank 2015f, World Bank 2015g, World Bank 2016e). The use of formal child care is consistently low, even when compared with countries of similar economic development (Figure 23 panel b). Also, despite of the existence of maternity leave in each country, there are additional barriers to employment that have not yet been addressed. For example, mothers are not guaranteed equivalent positions after maternity leave in North Macedonia and Serbia, and mothers in Kosovo are not entitled to nursing breaks (Iqbal 2015). Furthermore, even after the first several months of a child’s life, women continue to be considered primary care givers past the duration of maternity leave and therefore often face barriers to combining participation in the labor market with household duties. Parents are entitled to flexible or part-time schedules only in Bosnia and Herzegovina (Iqbal 2015). 17 We model both the propensity to participate in the labor market and the propensity to work, separately for men and women. We include as explanatory factors socio-demographic characteristics; health indicators, such as self- perceived health status and having disability; living condition variables, and household’s assets holdings. 18 The number of other employed members in the household is positively correlated with the propensity to participate (work), suggesting that both men and women in households with higher work intensity are more likely to seek job opportunities. In addition, being the head of the household increases the propensity to participate (work), especially among Roma females. On the other hand, having any disability is negatively correlated with probability to participate (work), while better self-perceived health status is positively correlated with the probability to participate (work), especially among Roma males. Finally, household’s living conditions and household’s assets holding, have mixed effects on the probability to participate (work), although most of them are not statistically significant. 19 Regional population weighted statistics based on the 2017 Roma Regional Survey. 38 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Figure 23. High Young Dependency Ratios and Limited Use of Formal Child care in the Western Balkans a. Young-age dependency ratio, % b. Population in households using childcare (% of total population) Percent of population in households using KSV 70 Russian Federation Latvia SRB 60 Belarus institutional childcare Slovenia Slovak Republic MNE 50 Mongolia Poland Cyprus Percent Kazakhstan 40 Ukraine Serbia Lithuania MKD Kyrgyz Republic Montenegro Germany Georgia Romania Italy Albania 30 Bulgaria BIH Bosnia and Uzbekistan North Herzegovina 20 Macedonia Croatia Armenia Estonia Greece Moldova Hungary ALB Tajikistan 10 Turkey Kosovo Azerbaijan Czech Republic 0 20 40 60 80 100 0 0 10000 20000 30000 40000 50000 National Non-Roma Roma GDP per capita, PPP (constant 2011 international USD) 2015 Sources: Panel a. Roma and non-Roma: World Bank estimates based on weighted 2011 and 2017 UNDP-World Bank-EC Regional Roma Survey data. National estimates: ECAPOV (circa 2017), ALB: HBS 2016; KSV: ECAPOV 2017; MNE: ECAPOV 2015; MKD: EU-SILC 2016; SRB: ECAPOV 2015. Panel b. World Bank estimates based on 2016 data of LiTS (Life in Transition Survey) (database), European Bank for Reconstruction and Development, London, http://www.ebrd.com/what-we-do/economic-research-and-data/data/lits.html; World Development Indicators. Note: The young-age dependency ratio = ratio of children ages 0–14 in the household to the working-age population (ages 15–64). In addition, a low level of skills among working-age Roma has been recognized as a major barrier to employment in the Western Balkans; skill barriers seem to affect significantly more Roma females than males. Roma are much constrained than non-Roma neighbors in terms of access to formal education. Across countries, educational attainment is even more deficient among working-age Roma women, except in Albania (Figure 24 panel a). For example, in Kosovo, most of working-age Roma women (64 percent) have not completed compulsory education; in contrast, this is the case among only 45 percent of Roma men, and 17 percent of Roma men have completed upper-secondary or above. Low educational attainment among Roma women translates into low labor force participation and low earnings. Literacy rates among working-age Roma female are also lower (Figure 24 panel b) putting them at an additional disadvantage. The failure of the education system to adequately prepare Roma males and females, and the resulting difficulty in accessing jobs, point to the need for interventions to improve the public provision of education and training among vulnerable communities, actively promoting their skill acquisition. 39 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Figure 24. Working-age Roma Females Attain Less Education on Average than Men, Already at Low Levels; Self-reported Literacy a. Distribution of educational attainment among b. Literacy rates by gender, ages 15–64 working-age Roma, by Gender 100 100 80 90 60 80 70 40 Percent 60 20 50 0 40 30 Roma female Roma female Roma female Roma female Roma female Roma female Roma male Roma male Roma male Roma male Roma male Roma male 20 10 0 Males Males Males Males Males Males Females Females Females Females Females Females ALB BIH MKD MNE SRB KSV ALB BIH MNE MKD SRB KSV ISCED 3 Complete/Tertiary Complete ISCED 2 Complete Roma Non-Roma ISCED2 Incomplete or below Source: World Bank estimates based on weighted 2017 UNPD-WB-EC Regional Roma Survey data.Note: Panel a) shows the distribution of educational attainment among the working-age population (15–64). It is also possible that some Roma women, and especially those with lower educational levels, choose not to join the labor market based on the expected returns. As suggested by qualitative work, participating in the labor market is less profitable than home production and not highly valued by women or their communities (Robayo-Abril and Millan 2019). This provides an indication that their lack of labor market inclusion may also be the result of a conscious decision in some instances. The low progressivity of labor taxes may be affecting the low-wage and part-time workers, groups in which Roma males and females are overrepresented, potentially affecting the hiring of this type of workers as well as their labor force participation. The tax burden is high in the Western Balkans, particularly among low-wage and part-time workers (Figure 25 Panel a). Arias et al. (2014) demonstrate the low progressivity of labor taxes in many Europe and Central Asia countries. Among both men and women, marginalized Roma are more likely than their non-Roma counterparts to take part-time jobs. In 2017, the share of part-time employment among employed Roma men ranged from 23 percent in North Macedonia (versus 8 percent among non-Roma men) to 51 percent in Serbia (versus 13 percent among neighboring non-Roma men)(Figure 25 Panel b). Among women, the share reached 62 percent in Montenegro (versus 15 percent among neighboring non-Roma women). Low progressivity may also discourage Roma females, whose potential earnings are at the bottom of the earnings distribution and part-time workers to search for formal-sector jobs. 40 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Figure 25. The Tax Burden is High for Low-Wage and Part-time Workers, Disproportionately Affecting Roma Males and Females Who Are Relatively More Likely to Take These Type of Jobs a. Labor taxes, levels and progressivity (selected b. Part-time workers (% of total workers) countries) 20 Roma Male Non-Roma Male Belgium Ireland Progressivity of tax wedge (percetnage 18 Roma female Non-Roma Female France 16 Switzerland Germany Austria Hungary 70 14 Italy Percent United Kingom Netherlands 60 12 points) 50 10 NorwaySpain Slovenia Portugal Finland 8 Czech Republic 40 Slovak Republic United States Greece Latvia 30 6 Estonia LithuaniaSweden PolandBosnia - 4 Macedonia Bosnia - Romania 20 Republika Federation 2 Japan Srpska Serbia 10 0 Bulgaria 0 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% ALB BIH MNE MKD SRB KSV Tax wedge for single at 33 percent of average wage Source: Panel a) Arias et al. (2014) based on OECD Tax and Benefit Model. Note: Tax wedge measures personal income tax and social security contributions paid by workers and employers as a share of total labor costs. The y-axis shows progressivity of labor taxes, measured as the tax wedge difference between average and low-wage workers (for a single person with no children at 100 percent or 33 percent of average wage, respectively). Progressivity refers to the increase in the tax wedge in percentage points. The x-axis shows the tax wedge for low-wage earners (earning 33 percent of the average wage). Panel b) World Bank estimates based on weighted 2017 UNDP-WB-EC Regional Roma Survey data. Note: Sample restricted to working-age individuals (15–64). In addition to the high tax wedge, the existing restrictive labor legislation in some countries may particularly affect Roma women. Countries are adopting more flexible labor legislation, but it is still generally restrictive. For example, in Montenegro, part-time work cannot be less than one-fourth of a full-time schedule (that is, 10 hours a week). Employees with minor children have additional legal rights to flexible or part-time work arrangements. However, a new labor law proposal currently under consideration restricts part-time work even further to a minimum of 20 hours (half of full-time schedule) potentially affecting significantly Roma females. In Serbia, the wage that determines social security contributions is not adjusted for hours worked; so, part-time workers disproportionately contribute more than their full-time counterparts; lack of access to part-time work may disincentive women from joining the labor market because they are more likely to have caregiving responsibilities during regular working hours. Labor legislation related to maternity leave may be operating as a constraint to employment for young women in Kosovo. Indeed, women face discrimination by employers in Kosovo as firms bear most of the financial costs of maternity provisions and maternity leave is also the main family leave available. 4 Policies and Interventions 4.1 Institutional Framework and Policies on Roma Gender Equality To tackle gender gaps, each country in the Western Balkans has national legislation and a National Strategy and a Gender Action Plan in place. All countries indeed have passed a Law on gender Equality (Law on Equality of Opportunities in North Macedonia) (see table 4 below). Most of the policy instruments outlined in the table refer to the more vulnerable position of 41 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Roma women in the respective countries and foresee specific objectives or actions with regards to, for instance, access to health (Albania, Montenegro and North Macedonia), political participation (Albania and Montenegro), employment (North Macedonia and Bosnia and Herzegovina), human trafficking (Montenegro) and education (Albania). In addition, all countries have mechanisms to support gender equality. Bosnia and Herzegovina and Kosovo both have a Gender Equality Agency. In Serbia, where the Gender Equality Directorate was dismantled in 2014, the Coordinating Body for Gender Equality has since been created. In North Macedonia, the Sector for Equal Opportunities within the Ministry of Labor and Social Policy contains the Unit for Gender Equality as well as the Unit for Prevention and Protection against all forms of discrimination. In addition to these regulatory structures in place in, each country has a Gender Action Plan that aims to tackle a range of gender gaps in endowments, economic opportunities, and agency (World Bank, 2016f) (see Table 4 below). Table 4. Legal, Institutional and Policy Framework for Gender Equality in Western Balkan Countries Country Legislation Policies Institution Montenegro Law on Gender Equality Action Plan for Achieving Gender Department for Gender (2007, amended in 2015). Equality (2017-2021). Equality in the Ministry for Human and Minority Rights. Albania Law on Gender Equality in National Strategy and Action Plan on Sector for Gender Equality Society (2008). Gender Equality (2016-2020). at the Ministry of Social Welfare and Youth. Bosnia and Law on Gender Equality Gender Action Plan of Bosnia and Agency for Gender Herzegovina (2003, amended in 2009). Herzegovina (2018 – 2022). Equality. Kosovo Law on gender equality Program of Kosovo for Gender Equality Agency for Gender (2004). (2008-2013). Equality, Office of the A new program has not yet been Prime Minister. adopted. Strategy for the Protection against Domestic Violence 2016-2020 North Macedonia Law on Equal National Equality and Non- Sector for Equal Opportunities for Women Discrimination Strategy (2016-2020). Opportunities in the and Men (2012). National Action Plan for the Ministry of Labor and Implementation of the Law on the Social Policy, Prevention and Protection Against Discrimination (2015-2020). Strategy for introducing Gender Responsive Budgeting in the Republic of Macedonia (2012-2017). Serbia Law on Gender Equality National Strategy and Action Plan for Coordination Body for (2009). Gender Equality (2016-2020). Gender Equality. 42 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Similarly, all countries have specific legislation and policies catered to the specific situation of the Roma minority. All countries legally prohibit discrimination on the grounds of sex and ethnicity (see table 5 below). Kosovo and Bosnia and Herzegovina have passed specific laws for the protection of the rights of minorities. All countries have also developed strategies and action plans for Roma inclusion20. All these, with the exception of Bosnia and Herzegovina, feature women and girls as a specifically vulnerable collective within Roma communities, and therefore make provisions targeted at them, especially with regards to access to basic services, gender- based violence and employment. Rreferences to the gender dimension of Roma inclusion are more obvious than those to Roma inclusion in instruments focused on gender equality. Table 5. Legal, Institutional and Policy Framework for Roma Inclusion in Western Balkan Countries Country Legislation Policy Institution Montenegro Law on Prohibition of Strategy for social Ministry for Human and Discrimination (amended in inclusion of Roma and Minority Rights. 2014). Law on the Protection Egyptian communities of Human Rights and (2016 – 2020). Freedoms. Albania Law on Protection from National Action Plan for Ministry of Social Welfare Discrimination (2010). Integration of Roma and and Youth. Egyptians (2016-2020). Bosnia and Herzegovina Law on the Protection of Action Plan for addressing Ministry for Human Rights of Members of Roma issues in the fields Rights and Refugees. National Minorities of the of employment, housing Federation of BiH, of and health care (2017- Republika Srpska. And of 2020). Brcko District. Law on Special Action Plan on the Prohibition of Discrimination educational needs of (2009). Roma. Kosovo Law on Protection of Strategy and Action Plan Office for Good Communities Rights and their for inclusion of Roma and Governance within the members. Law on Protection Ashkali communities in Office of the Prime against discrimination. the Kosovo society (2017- Minister. 2021). North Macedonia Law on Prevention and Strategy for the Roma National Coordination Protection Against (2014-2020). Body in the Ministry of Discrimination (2010). Labor and Social Policy. Serbia Law on Prohibition of Strategy and Action Plan Roma Social Inclusion Discrimination (2009). for Social Inclusion of Coordination Body. Roma (2016-2025). All Western Balkan countries have institutions in charge of the Roma agenda in place. Montenegro and Bosnia and Herzegovina, indeed, have a Ministry responsible for Human Rights; in the case of Montenegro, the Ministry is explicitly dedicated to Minority Rights. In the case of Kosovo, the main agency – the Office for Good Governance – is part of the Prime Minister Office. In both North Macedonia and Serbia national coordination bodies exist. Other 20 For a review of Roma action plans, see Robayo-Abril and Millan 2019. 43 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS relevant institutions in this area across countries include anti-discrimination commissions, minority councils and the Ombudsman. The EU institutions (e.g., European Commission) and CSO organizations (e.g., Open Society Foundation) are also especially active in this area. Despite the myriad of institutions and policies on issues related to gender and Roma inclusion in Western Balkan countries, the persistence of gaps may be an indication of the inadequate capacity of implementation and enforcement bodies. Indeed, there are various reports that the financial and human capacity of these institutions is in many cases insufficient for them to fulfill their mandates and influence policy (Dokmanovic 2016, Government of Albania 2016, Lazarević and Tadić 2018, EC 2018). In addition, it must be noted that coordination and coherence challenges – i.e. overlaps, gaps and potential inefficiencies – are likely to exist given the clear institutional separation between the Roma and gender agendas across countries. The compound result may be that Roma women oftentimes would not only not benefit from but also would even be adversely affected by both Roma policies and gender equality ones (Kóczé and Raluca 2009). 4.2 The Road Forward: What Works for Roma Women Inclusion Targeted programmatic interventions should tackle the multiple barriers and disincentives that affect Roma females across all countries in the Western Balkans. With regards to markets, for instance, equalizing access to quality basic services among the Roma will likely require improvements in the coverage and quality of services, and their ethnic and gender- sensitiveness. Achieving gender parity in employment among Roma communities may also demand specialized interventions to tackle constraints that affect significantly more young Roma females. On the side of (informal) institutions, working on social norms underpinning the especially large gaps in outcomes experienced by Roma women and girls appears to be fundamental across all areas studied, most likely through a combination of education, information and awareness raising activities targeting young people and parents. In addition, the formal legal and institutional framework for Roma and gender issues should be improved across countries to better reflect and address the needs of this particularly vulnerable group. This assessment confirms that important knowledge and data gaps exist. First, although much information and research exist with regards to demand-side constraints to Roma women´s inclusion, little is known about the supply-side constraints faced by this group. This is particularly the case with regards to those that go beyond the traditional structural and institutional barriers of rational choice theory (see below). Furthermore, most analysis on women’s issues within the Roma community treat ethnicity and gender as separate social categories, and the disadvantages rendered by it as additive.21 In addition, and as noted in the diagnostic, although parental expectations on the education of Roma boys and girls are similar, large gaps exist in enrolment and attainment that contradict such result. More research and analysis on what is driving these contradictory data would be of special interest to understand the dynamics behind differenciated outcomes. Finally, a better understanding of what is 21Several feminist scholars call this critique one of ‘missing intersectionality’, i.e. while scholars and practitioners alike have emphasized the unique ways women experience race / ethnicity and gender, during analysis the idea is widely cited but unevenly applied (see Kóczé and Raluca 2009, Oprea 2004, Schultz 2012, Corradi 2017) 44 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS preventing the effective functioning of institutions and policies in this area would be helpful moving forward. Box 3. Behavioral Economics and Roma Exclusion A recent paper concludes that people are prevented from participating fully in society through mechanisms that go beyond the structural and institutional barriers identified by rational choice theory (poverty, exclusion by law or force, taste- based and statistical discrimination, and externalities from social networks), and discusses four additional mechanisms that bounded rationality can explain: (i) implicit discrimination, (ii) self-stereotyping and self-censorship, (iii) “fast thinking” adapted to underclass neighborhoods, and (iv) “adaptive preferences” in which an oppressed group views its oppression as natural or even preferred. Stable institutions have cognitive foundations—concepts, categories, social identities, and worldviews—that function like lenses through which individuals see themselves and the world. Abolishing or reforming a discriminatory institution may have little effect on these lenses. Groups previously discriminated against by law or policy may remain excluded through habits of the mind. Behavioral economics recognizes forces of social exclusion left out of rational choice theory and identifies ways to overcome them (Hoff and Walsh 2017). Based on the analysis presented in this note, specific policy recommendations on gender and Roma include the following: Provide additional financial incentives to girls to attend school, such as more generous conditional cash transfers. As seen above, one of the major barriers to equitable access to education is lack of financial resources. A large literature finds that, if parents are provided with more financial assistance for the education of daughters than for the education of sons, the impact on enrollment rates among girls is large. An example of good practice is Mexico’s Prospera program, which offered larger cash transfers for girls than for boys. The initial analysis of Prospera’s impact on education showed that the program significantly enlarged the enrollment of both boys and girls, and particularly so of girls in secondary school. CCTs are also effective in postponing marriage and reproduction among school-age girls (Skoufias and McClafferty 2001). Provide legal protection against child marriage. Early marriage is one of the key barriers for girls to access the same opportunities as boys in Roma communities, starting with education. Globally, girls married before the age of 18 are at a greater risk to experience gender-based violence, poorer health outcomes, early pregnancies, to drop out of school and consequently earn less over their lifetimes, and to live in poverty (Wodon et al. 2017). Girls under 18 years could legally be married with their parents’ consent in several countries including Kosovo, Bosnia and Herzegovina, Albania and Macedonia (UNFPA 2014, Girls not Brides 2014, UNFPA Macedonia22). Even in cases where early marriage is prohibited, such as in Montenegro or Serbia, the incidence of these phenomenon is large among the Roma and remains largely unpunished. Empowering girls and engaging their families and communities to prevent child marriage. The lack of regulation or enforcement of laws in this area combines with the generally weak 22 https://eeca.unfpa.org/sites/default/files/pub-pdf/unfpa%20macedonia%20overview.pdf 45 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS empowerment of women and girls and the prevalence of traditional and patriarchal norms in Roma communities. Especially promising interventions in this regard appear to be the safe space programs, which help girls to build their life skills, health and financial literacy and provides them with information on services; these programs offer a space where girls can meet regularly and learn from each other and from designated mentors in an informal environment. The Abriendo Oportunidades program in Guatemala offers an example of such programs.23 Enagaging the families, parents and wider community in the discussion over the negative implications of child marrieage and the importance of girls´ education would also be central in this area.24 Change labor laws reinforcing gender inequality. Western Balkan countries must ensure that labor legislation treats males and females in an equitable way. Enabling all parents to balanced paid work is important, facilitating an equal role for men in caregiving. Laws should ensure that parents have equal access to meet their children and adult family members´ needs. Providing and encouraging the use of effectively shared parental leave policies – for instance through ¨daddy¨ quotas or bonuses – and ensuring the effective protection of women´s work can play an especially important role in this area. Part-time and flexible work formulas need to be legally allowed and not discouraged. In addition, the low progressivity of labor taxes needs to be addressed. However, it must be noted that it is not the lack of legislation in general that undermines the situation of Roma women, but mostly its lack of enforcement due to, among others, potential discrimination based on patriarchal social and gender norms. Implement measures to reduce gender-biased social norms. There is evidence that social norms related to the role of women and girls vis-á-vis that of men and boys are in a state of flux. For instance, a study from Bulgaria shows that despite traditional gender roles being formally upheld, de facto male authority in the family seems to be increasingly challenged by women in many Roma communities.25 In addition, a qualitative study from Serbia also concludes that some norms and aspirations around early marriage, time use, and education are shifting, which offers a relevant entry point for policy interventions in this area. 26 Education and information are key in this regard. A combination of self-help measures for women (see below) and promotional campaigns that for instance highlight the success stories of Roma women could be particularly helpful. Engaging men and boys in any initiative aimed at changing social norms would be necessary (see for instance OSCE and UNFPA 2018). Promote social cohesion and women´s empowerment using the proven model of self-help groups. Coordination among the poor and the vulnerable is generally costly, and women face particularly high barriers to collective action in communities. Self-help groups are membership- based organizations that have been widely used to encourage social cohesion through a combination of education, access to finance, and links to development programs. This formula relies on women’s collective capacity to break barriers, increase access to assets, and improve their voice and agency. Self-help groups have showed to lead to higher solidarity between peers, 23 https://www.girlsnotbrides.org/how-can-we-end-child-marriage/#empower-girls; https://www.unicef.org/protection/files/Global_Programme_Brochure_2018(1).pdf 24 https://www.unicef.org/protection/files/Global_Programme_Brochure_2018(1).pdf 25 World Bank (2014) 26 World Bank (2019) 46 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS more independent financial decision-making, and greater respect for women within households and communities (Brody et al. 2015). Implement measures to overcome constraints related to young female Roma’s care responsibilities. Interventions to overcome constraints related to young women’s care responsibilities can be classified into two types: (1) among females who do not have children, interventions that seek to delay marriage and reduce the fertility rate, raising enrollment and the retention of girls in school; and (2) interventions that alleviate the time constraints on women who already have children, for example, through high quality affordable childcare service provision. With regards to the second type of measures, there is a large body of evidence that shows that the provision of quality accessible and affordable childcare services tends to increase the labor force participation of women throughout countries in the world (Lokshin 2004 and Fong and Lokshin 2000, Scholsser 2011, Berlinski and Galiani 2007, Barros et al. 2011). Strengthening institutional capacity and ensuring adequate coordination. Coordination between the multiple institutons involved in the two areas of gender equality and Roma inclusion will be crucial for effectiveness. In some cases, such as Montenegro or Serbia, the main agencies in charge of both gender and Roma inclusion belong to the same main umbrella organisms; this could facilitate coordination. In other cases, however, they are completely separate, which could often lead to duplications, overlap and incoherence, and ultimately to the waste of limited resources. In addition, such institutional set-up may lead to Roma women not benefiting or even being disadvanagated by both Roma and gender policies. Strengthening both these agendas in an integral manner would require further and concerted efforts at the top governmental level. Setting up adequate monitoring and evaluation mechanisms in this area. Disagregated data and evidence on what drives the large gaps for Roma women faced with a double exclusion factor – gender and ethnicity – is particularly scarce and necessary to better address the vulnerabilities experienced by them. Adequate monitoring and evaluation mechanisms would allow to assess progress, and gain a better understanding of what works and what does not to address the existing gaps. The results of these efforts could then be used to inform future policy developments and therefore would contribute to the design and implementation of increasingly effective measures in this area. They could also be helpful to inform policies in other European countries confronting similar challenges. 47 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Bibliography EBRD (European Bank for Reconstruction and Development). (2016). Life In Transition Survey. London: EBRD. A. Chevalier & T. K. Viitanen. (2002). The causality between female labour force participation and the availability of childcare. Applied Economics Letters, 9(14), 915-918. doi:10.1080/13504850210138469 Allin, S., & Masseria, C. (2009). Unmet need as an indicator of health care access. EuroHealth, 15, 7-9. Anderson, M., Dobkin, C., & Gross, T. (2010, March). The E ffect of Health Insurance Coverage on the Use of Medical Services,. NBER Working Paper No. 15823. Arandarenko, M., & Bartlett, W. (2012). Labor market and Skills in the Western Balkans. Foundation for the Advancement of Economics and LSE European Institute. Arias, O. S., Sánchez-Páramo, C., Dávalos, M. E., Santos, I., Tiongson, E. R., Gruen, C., . . . Cancho, C. A. (2014). Back to Work: Growing with Jobs in Europe and Central Asia (Vol. Europe and Central Asia Reports). Washington, DC: World Bank. Armand, A., and Carneiro, P. (2018). Impact Evaluation of the Conditional Cash Transfer Program for Secondary School Attendance in Macedonia. 3ie Impact Evaluation Report 69 (January), International Initiative for Impact Evaluation, New Delhi. Asai, Y., Kambayashi, R., & Yamaguchi, S. (2015). Childcare availability, household structure, and maternal employment. Journal of the Japanese and International Economies, 38 , 172-192. Asenov, R., Buhler, D., Jasarevic, M., & Millan, N. (2018). Roma inclusion in the Western Balkans: Policy brief. Washington, DC: The World Bank. Attanasio, O. & Vera-Hernandez, M. (2004). Medium- and long-run effects of nutrition and child care: Evaluation of a community nursery programme in rural Colombia. Working Paper Series, Institute for Fiscal Studies. Attanasio, O., & Székely, M. (1999). An Asset-Based Approach to the Analysis of Poverty in Latin America. IDB Working Paper No. 109. Austin, B., Glaeser, E., & Summers, L. (2018). Saving the heartland: Place-based policies in 21st century America. Brookings Papers on Economic Activity Conference Draft. Ayanian, J. Z., Weissman, J. S., Schneider, E. C., Ginsburg, J. A., & Zaslavsky, A. M. (2000). Unmet Health Needs of Uninsured Adults in the United States. Journal of the American Medical Association, 284 (16), 2061-2069. Baird, S., Chirwa, E., Mcintosh, C., & Ozler, B. (2010). The short-term impacts of a schooling conditional cash transfer program on the sexual behavior of young women. Health Economics, 19(S1), 55-68. Barrera-Osorio, F., Garcia-Moreno, V., Patrinos, H. A., & Porta, E. (2011). Using the Oaxaca-Blinder Decomposition Technique to Analyze Learning Outcomes Changes over Time: An Application to Indonesia’s Results in PISA Mathematics. Policy Research Working Paper No 5584. Berlinski and Galiani. (2007). The effect of a large expansion of pre-primary school facilities on preschool attendance and maternal employment. Labour Economics, 14(3), 665-680. 48 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Bertrand, M., & Crépon, B. (2016). Teaching Labor Laws: Results from a Randomized Control Trial in South Africa. Center for Research in Economics and Statistics, Palaiseau, France. Betcherman, G., Godfrey, M., Puerto, S., Rother, F., & Stavreska, A. (2007). A Review of Interventions to Support Young Workers: Findings of the Youth Employment Inventory. Social Protection discussion paper no. SP 0715. Washington, DC: World Bank Blinder, A. S. (1973). Wage Discrimination: Reduced Form and Structural Estimation. Journal of Human Resources, 8, 436-455. Bodewig, C., & Sethi, A. (2005). Poverty, Social Exclusion and Ethnicity in Serbia and Montenegro: The Case of the Roma . (p.63). Belgrade: World Bank. Borra, Cristina. (2010). Childcare cost and Spanish mother’s labour force participation. Hacienda Pública Española, 194(3), 9-40. Boudet, A. M., Petesch, P., & Turk, C. (2013b). On norms and agency: Conversations about gender equality with women and men in 20 countries. Washington DC: World Bank. Brody, C., T. de Hoop, M. Vojtkova, R. Warnock, M. Dunbar, P. Murthy, and S. Dworkin. (2015). Economic Self-Help Group Browne, E. (2017). Gender in the Western Balkans. K4D Helpdesk Report. Brighton, UK: Institute of Development Studies. Brüggemann, C., & Friedman, E. (2017). The Decade of Roma Inclusion: Origins, Actors and Legacies. Journal of European Education, Vol. 49, Issue 1. Bussolo, M., Koettl, J., & Sinnott, E. (2015). Golden Aging : Prospects for Healthy, Active, and Prosperous Aging in Europe and Central Asia. Washington, DC: World Bank. Butler, J. (2004). Undoing Gender. Psychology Press. Cahuc, P., & Zylberberg, A. (2004). Labor Economics. Cambridge, Massachusetts: MIT Press. Campos, F., Goldstein, M., McGorman, L., Pimhidzai, O., & Boudet, A. M. (2016). Breaking the Metal Ceiling: Female Entrepreneurs who Succeed in Male-Dominated Sectors. World Bank Policy Research Working Paper No. 7503. Card, D., Kluve, J., & Weber, A. (2015). What Works? A Meta-Analysis of Recent Active Labor Market Program Evaluations. Cambridge, MA: NBER Working Paper No. 21431. CARE. (2011). Situational Analysis of Education and Social Inclusion of Roma Girls in Serbia. Serbia, Belgrade: Report, CARE. Carter, M., & Barrett, C. (2006). The economics of poverty traps and persistent poverty: An asset-based approach. Journal of Development Studies, Taylor & Francis Journals, vol. 42(2), 178-199. Chakravarty, S., Das, S., & Villant, J. (2017). Gender and Youth Employment in Sub-Saharan Africa: A review of Constraints and Effective Interventions. Policy Research Working Paper No. 8245. 49 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Cho, Y., & Honorati, M. (2014). Entrepreneurship Programs in Developing Countries: a Meta Regression Analysis. Cojocaru, A. (2017). Kosovo jobs diagnostic. Washington, D.C.: World Bank Group. Contreras, D.; Plaza, G.;. (2010). Cultural Factors in Women's Labor Force Participation in Chile. Feminist Economics, 16(2), 27-46. Corradi, L. (2017). Gypsy Feminism: Intersectional Politics, Alliances, Gender and Queer Activism. Routledge. Council of Europe. (2012). Protecting the Rights of Roma. Retrieved from http://www.strassburg- europarat.diplo.de/contentblob/3392612/Daten/1862070/CoE_publication__Roma.pdf Cuberes, D; Teignier, M. (2015). How costly are labor gender gaps? Estimates for the Balkans and Turkey. World Bank Policy Research Working Paper No. 7319. Cukrowska, E., and Kóczé, A. (2013). Interplay between gender and ethnicity: exposing structural disparities of Romani women. Analysis of the UNDP/World Bank/EC regional Roma survey data. Roma Inclusion Working Papers. Bratislava: UNDP. Dao, M. (2014). Female Labor Force Participation in Developing Countries - La partecipazione della forza lavoro femminile nei paesi in via di sviluppo. Economia Internazionale / International Economics, 67 (3), 337- 350. De Laat, J., & Bodewig, C. (2011). Roma inclusion is smart economics. Washington, DC: World Bank . Deaton, A. (2002). Policy Implications of The Gradient of Health And Wealth. Health Affairs, 21(2). Dokmanovic, M. (2016). Gender Analysis for Serbia. IBF Interntional Consulting. Drezgić, R. (2010). Religion, politics and gender in the context of nation-state formation: the case of Serbia. Third World Quarterly, 31(6), 955-970. Duflo, E., Dupas, P., & Kremer, M. (2014). Education, HIV, and early fertility: Experimental evidence from Kenya. MIT, working paper. Retrieved from Retrieved from http://economics.mit.edu/files/6951. Durst, J. (2002). Fertility and Childbearing Practices among Poor Gypsy Women in Hungary: The intersections of class, race and gender. Communist and Post-Communist Studies, Vol. 35. EBRD (European Bank for Reconstruction and Development). (2010). Life in Transition Survey. London: EBRD. EBRD (European Bank for Reconstruction and Development). (2016). Life in Transition Survey. London: EBRD. Elbers, C., Fujii, T., Lanjouw, P., Yin, W., & Ozler, B. (2004). Poverty alleviation through geographic targeting: How much does disaggregation help? The World Bank. EPTISA. (2016). Report on State of Service Delivery within Local Public Administration Institutions. Mimeo. ERGO. (2017). Investing in our Future: What Work(s) for Young Roma? Top 10 Best Practices of Roma Youth employment. European Grassroots Organizations Network. 50 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS ERRC. (2017). Statelessness, Discrimination and Marginalization of Roma in the Western Balkans and Ukraine. Budapest: European Roma Rights Center (ERRC). ERRC. (2017). Written Comments of the European Roma Rights Centre Concerning Serbia For Consideration by the Human Rights Council, Working Group on the Universal Periodic Review. Retrieved from http://www.errc.org/cms/upload/file/serbia-hrc-submission-30-june-2017.pdf ESARM. (2017). Number of Roma Participants per program from the Employment Service Agency of the Republic of Macedonia. European Commission (2011). An EU Framework for National Roma Integration Strategies up to 2020: Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee, and the Committee of the Regions.” Document COM(2011) 173/4, European Commission, Brussels. European Union. (2013). Empowerment of Roma Women within the European Framework of National Roma Inclusion Strategies. European Parliament, Directorate General for Internal Policies. European Commission. (2015). Employment, Social Affairs and Inclusion: Measuring Skill Mismatches. Analytical Note. European Commission (2018). A credible enlargement perspective for and enhanced EU engagement with the Western Balkans. Communication from the Commission European Parliament, the European Council, the Economic and Social Committee and the Committee of the Regions. Fairlie, R. W. (2005). An Extension of the Blinder-Oaxaca Decomposition Technique to Logit and Probit Models. Journal of Economic and Social Measurement, 30, 305-316. Fong and Lokshin. (2000). Child Care and Women's Labor Force Participation in Romania. Policy Research Working Paper;No. 2400. World Bank, Washington, DC. FRA (European Union Agency for Fundamental Rights). (2009). Case Study: Improving Roma housing and eliminating slums, Spain. mimeo. Retrieved from https://fra.europa.eu/sites/default/files/fra_uploads/605- 090210-ROMA_Housing_Case-final-ENES.pdf Gatti, R., & Flabbi, L. (2018, January). A Primer on Human Capital. Washington, DC: World Bank. Gatti, R., Karacsony, S., Anan, K., Ferre, C., & de Paz Nieves, C. 2. (2016). Being Fair, Faring Better: Promoting Equality of Opportunity for Marginalized Roma. (Vols. Directions in Development--Human Development.). Washington, DC: World Bank. Gatti, R., Karacsony, S., Anan, K., Ferre, C., & de Paz Nieves, C. (2016). Being Fair, Faring Better: Promoting Equality of Opportunity for Marginalized Roma. (Vols. Directions in Development--Human Development.). Washington, DC: World Bank. Government of Albania (2016). National Strategy and Action Plan on Gender Equality 2016-2020. Government of the Republic of Serbia. (2010). Strategy for Improvement of the Status of Roma in the Republic of Serbia. Belgrade. Retrieved from http://www.seio.gov.rs/upload/documents/ekspertske%20.misije/protection_of_minorities/strategy_for_r oma.pdf 51 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Group, W. B. (2014). Perceptions and demand for childcare and eldercare services in the Western Balkans : a qualitative study (English). Washington, DC: World Bank. Retrieved from http://documents.worldbank.org/curated/en/807501521459331060/Perceptions-and-demand-for- childcare-and-eldercare-services-in-the-Western-Balkans-a-qualitative-study Hallegatte, S., Bangalore, M., Bonzanigo, L., Fay, M., Tomaro, K., Narloch, U., . . . Vogt-Schilb, A. (2014). Climate Change and Poverty. Policy Research Working Paper No 7126. Hausmann, R., Tyson, L. D., & Zahidi, S. (Eds.). (2014). The Global Gender Gap Report 2013. Geneva, Switzerland: World Economic Forum. Hausmann-Muela, S., Muela Ribera, J., & Nyamongo, I. (2003). Health seeking behaviour and the health system response. Disease Control Piroirities Project working paper No14. Hoff, K. and Walsh, J. (2017). The Whys of Social Exclusion Insights from Behavioral Economics. Policy Research Working Paper 8267. Hughson, M. (2014). Gender Country Profile For Bosnia And Herzegovina. HTSPE Limited. https://europa.ba/wp-content/uploads/2015/05/delegacijaEU_2014070314432045eng.pdf Huillery, E., de Laat, J., & Gertler, P. (2017). Supporting Disadvantaged Children to Enter Kindergarten: Experimental evidence from Bulgaria. Washington, DC: The World Bank. Idzerda, L., Adams, O., Patrick, J., Schrecker, T., & Tugwell, P. (2011). Access to primary healthcare services for the Roma population in Serbia: a secondary data analysis. BMC Int Health Hum Rights. 11, 10. Iqbal, S. (2015). Women, business, and the law 2016 : getting to equal Vol. 1 of Women, business, and the law 2016 : getting to equal. Washington, D.C. : World Bank Group. https://hubs.worldbank.org/docs/imagebank/pages/docprofile.aspx?nodeid=25018263 ILO (2016). Promoting decent work opportunities for Roma youth in Central and Eastern Europe. An ILO resource guide. Jain-Chandra, Sonali et al. (2018). Gender Equality: Which Policies Have the Biggest Bang for the Buck? IMF Working Paper No. 18/105. Retrieved from https://www.imf.org/en/Publications/WP/Issues/2018/05/10/Gender-Equality-Which-Policies-Have-the- Biggest-Bang-for-the-Buck-45823 Kahanec, M. a. (2010). Intergenerational Transfer of Human Capital under Post-War Distress: The Displaced and the Roma in the Former Yugoslavia. In G. S. Epstein, & I. N. Gang, Migration and Culture. Bingley: Emerald Group Publishing Limited. Kendall, S., Straw, S., Jones, M., Springate, I., & Grayson, H. (2008). Narrowing the gap in outcomes for vulnerable groups. A review of the research evidence. National Foundation for Educational Research. Retrieved from http://www.nfer.ac.uk/nfer/publications/LNG01/LNG01.pdf Kezdi, G., & Suranyi, E. (2009). A Successful School Integration Program. Budapest: Roma Education Fund. Kling, J., & Kagin, J. (2012). Integrated household surveys among Roma populations: one possible approach to sampling used in the UNDP-World Bank-EC Regional Roma Survey 2011. Roma Inclusion Working Papers. 52 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Klugman, J., Hanmer, L., Twigg, S. , Hasan, T. , McCleary-Sills, J. and Santamaria, J. (2014). Voice and Agency: Empowering Women and Girls for Shared Prosperity. Washington: World Bank. Kluve, J., Puerto, S., Robalino, D., Romero, J., Rother, F., Stöterau, J., . . . Witte, M. (2016). Do Youth Employment Programs Improve Labor Market Outcomes? A Systematic Review. Kóczé, A., & Raluca, M. P. (2009). Missing Intersectionality: Race/Ethnicity, Gender and Class in Current Research and Policies on Roma Women in Europe. Center for Policy Studies. Kolin, M. (2010). Gender Inequalities and the status of women in the labor market in transitional Serbia. Anthropology of East Europe Review, 28(1), 119-137. Kozaric-Fanning, M. (2017). Analysis of the Priorities and Budget of the Official Roma Integration Policies in the Areas of Employment and Housing in the Western Balkans and Turkey. Belgrade: Roma Cooperation Council. Lazarević, M. and Tadić, K. (2018). Gender Issues in the Western Balkans. Civil Society Forum of the Western Balkan Summit Series. Levin, H. M., & Belfield, C. R. (2007). Educational Interventions to Raise High School Graduation Rates. In C. M. Belfield, & H. M. Levin (Eds.), The Price We Pay: Economic and Social Consequences of Inadequate Education (pp. 177–199). Washington, DC: Brookings Institution Press. Levin, V., Guallar Artal, S., & Safir, A. (2016). Skills for Work in Bulgaria : the Relationship Between Cognitive and Socioemotional Skills and Labor Market Outcomes. Washington DC: World Bank. Levin, V., Munoz Boudet, A. M., Rosen, B. Z., Aritomi, T., Flanagan, J., & Rodriguez-Chamussy, L. (2016). Why Should We Care about Care?: The Role of Childcare and Eldercare in Serbia. Washington, DC: World Bank. Retrieved from https://openknowledge.worldbank.org/handle/10986/29547 Loichinger, E., & Prskawetz, A. (2017). Changes in economic activity: The role of age and education. Demographic Research, 36(40), 1185-1208. Lokshin, Michael. (2004). Household Childcare Choices and Women's Work Behavior in Russia. The Journal of Human Resources, 39(4), 1094-115. doi:10.2307/3559040. Lorber, J. (1994). Paradoxes of Gender. Yale University Press. Lokshin. (2004). Household Childcare Choices and Women's Work Behavior in Russia. The Journal of Human Resources 39(4):1094-1115. Majumdar, S. and Woodhouse, A. 2019. "The Problem Is 'They See Us As A Whole': Explaining Gender- and Ethnicity-Based Inequalities among the Roma in Serbia." Background paper, World Bank, Washington, DC. Magnuson, K., & Waldfogel, J. (2016). Trends in Income-Related Gaps in Enrollment in Early Childhood Education: 1968 to 2013. AERA Open 2(2), 1-13. Mahmound, O., & Roberts, R. (2018). One camp, three refugee groups: challenges for local NGOs. Forced Migration Review(57), 7-9. 53 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Manley, James and Vásquez Lavín, Felipe, . (2013). Childcare Availability and Female Labor Force Participation: An Empirical Examination of the Chile Crece Contigo Program. Working Papers No 2013-03, Towson University, Department of Economics. Martinez, S., Naudeau, S., & Pereira, V. (2012). The promise of preschool in Africa: A randomized. World Bank and Save the Children. Retrieved from http://siteresources.worldbank.org/INTAFRICA/Resources/The_Promise_of_Preschool_in_Afric Maurer-Fazio M, Conelly R, Chen L, Tang L. (2011). Childcare, Eldercare, and Labor Force Participation of Married Women in Urban China, 1982–2000. Journal of Human Resources, 46(2), 261-294. McKenzie, D. (2017). How Effective are Active Labor Market Policies in Developing Countries? A Critical Review of Recent Evidence. Bonn.: IZA Discussion Paper Series, No. 10655. Messing, V., Beremenyi, L., Kurekova-Mytna, J., Konstekova, J., Pamporov, A., & Pop, F. (2013). From Benefits to Brooms: Case Studies Report on the Implementation of Active Labor Market Policies for Roma at Local Level. NEUJOBS. Mihajlović, M. (2004). Needs assessment study for the Roma Education Fund –Serbia (without Kosovo). Budapest: Roma Education Fund. Milcher, S. (2011). Decomposing income differentials between Roma and non-Roma in South East Europe. Romanian Journal of Regional Science, 5(1), 27–53. Milcher, S., & Fischer, M. M. (2011). On labour market discrimination against Roma in South East Europe. Papers in Regional Science, 90(4), 773-88. Milenković, N. (2018). Nowhere to turn: Gender-based violence against Roma women. UNDP. Ministry of Human and Minority Rights. (2016). “The Strategy for Social Inclusion of Roma and Egyptians in Montenegro – 2016-2020. Podgorica: Government of Montenegro. Ministry of Labor and Social Policy. (2014). Strategy for the Roma in the Republic of Macedonia – 2014-2020. Skopje: Government of the Republic of Macedonia. Ministry of Social Welfare and Youth. (2015). National Action Plan for Integration of Roma and Egyptians in the Republic of Albania. Tirana: Government of Albania. Mojsoska-Blazevski, N. a. (2015). Impact Evaluation of Active Labor Market Programs in FYR Macedonia: Key Findings. Budapest: International Labor Organization, ILO DWT and Country Office for Central and Eastern Europe. Montenegro, C. E., & Patrinos, H. A. (2014). Comparable estimates of returns to schooling around the world (English). Washington, DC: World Bank Group. Ñopo, H. (2008). Matching as a tool to decompose wage gaps. Review of Economics and Statistics 90, 290–299. O’Higgins, N. (2012). Roma and non-Roma in the Labour Market in Central and South Eastern Europe. Roma Inclusion Working Papers. O’Higgins, N. (2013). Ethnicity and Gender in the Labor Market in Central and South East Europe. Bonn.: IZA Discussion Paper No. 7667. 54 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS O’Higgins, N. (2015). Ethnicity and gender in the labour market in central and south -east Europe. Cambridge Journal of Economics, 39(2), 631–654. O’Higgins, N., & Brüggemann, C. (2013). The influence of educationa l attainment, special schooling and labour market discrimination on employment of Roma in Central Europe. IZA Discussion paper No. 7668. . Oaxaca, R. (1973). Male-female wage differentials in urban labor markets. International Economic Review, 14, 693–709. Oaxaca, R. L. (2007). The challenge of measuring labor market discrimination against women. Swedish Economic Policy Review, 14, 199. Oaxaca, R. L., & Ransom, M. R. (1994). On discrimination and the decomposition of wage differentials. Journal of Econometrics, 61, 5–21. OECD. (2001). The Well-being of Nations, The Role of Human and Social Capital. Paris: OECD. OECD. (2010). OECD Employment Outlook 2010: Moving beyond the Job Crisis. Paris . OECD. (2016). PISA 2015 Results (Volume I): Excellence and Equity in Education. Paris: OECD. Office of Good Governance and Office of the Prime Minister. (2017). Strategy for Inclusion of Roman and Ashkali Communities in the Kosovo Society – 2017-2021. Pristina: Government of Kosovo. Open Society Foundation. (2012). Roma Feel Less Fear and More Hope After Census. Retrieved from https://www.opensocietyfoundations.org/voices/roma-feel-less-fear-and-more-hope-after-census Oprea, A. (2004). Re-envisioning social justice from the ground up: Including the experiences of Roma women. Essex human rights review, 1(1), 29-39. Oprea, A. (2015). Re-envisioning Social Justice from the Ground Up: Including the Experiences of Romani Women. Essex Human Rights Review Vol. 1 No. 1. OSCE (Organization for Security and Co-operation in Europe). (2012). Access to Civil Registration in Kosovo. mimeo. OSCE and UNFPA (2918). A men´s perspective on gender equality in Kosovo. Oxfam and Argument. (2001). The Roma Livelihood in Belgrade Settlements. Belgrade. Paes de Barros, R., Olinto, P., Lunde, T. and Carvalho, M. (2011). The Impact of Free Childcare on Women’s Labor Force Participation: Evidence from Low-Income Neighborhoods of Rio de Janeiro. Conference Paper, World Bank Economists’ Forum. Patrinos, H. A. (2016). Estimating the return to schooling using the Mincer equation. IZA World of Labor 278. Pešić, J. (2006). Persistence of traditionalist value orientations in Serbia. Sociologija, 48(4), 289-307. Pleis, J. W. (2010). Vital and Health Statistics: Summary Health Statistics for U.S. Adults: National Health Interview Survey. National Health Interview Survey, 2009, 10(249). Powell, J. A. (2013, September/October). Deepening Our Understanding of Structural Marginalization. Poverty & Race, 22(5). 55 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Praxis. (2011). Analysis of the Main Problems and Obstacles in Access of Roma in Serbia to the Right to Education. Praxis. (2017). Joint Submission to the Human Rights Council at the 29th session of the Universal Periodic Review (Third cycle, 15 - 26 January 2018). Institute on Statelessness and Inclusion European Network on Statelessness and European Roma Rights Centre. Retrieved from http://www.institutesi.org/UPR29_Serbia.pdf Praxis and UNHCR. (2017). Stop to Child Marriage. Retrieved from http://praxis.org.rs/images/praxis_downloads/Stop_Decijim_Brakovima.pdf Programs for Improving Women’s Empowerment: A Systematic Review. Campbell Systematic Reviews 11, no 19. Ransom, M. R. (1999). Identification in detailed wage decompositions. Review of Economics and Statistics, 81, 154–157. Rawlings, L., Sherburne-Benz, L. D., & VanDomelen, J. (2004). Evaluating Social Funds: A Cross-country Analysis of Community Investments. Washington, DC: World Bank Publications. Reva, A. (2012). Gender Inequality in the Labor Market in Serbia. World Bank Policy Research Working Paper No. 6008. Washington DC: World Bank. Revenga, A., Ringold, D., & Tracy, W. M. (2002). Poverty and ethnicity: A cross-country study of Roma poverty in Central Europe. Washington DC: World Bank. Ridgeway, & L., C. (1991). The social construction of status value: Gender and other nominal characteristics. Social Forces, (pp. 70(2): 367-386). RisiAlbania. (2014). Employment Promotion Programs in Albania: An Assessment of its Quality in the Formulation and Implementation Processes (2008-2014). Albania: ILO-EU IPA 2010 Project on Human Resources. Risman, & Barbara, J. (2004). Gender as a Social Structure: Theory wrestling with activism. Robayo-Abril, M; Millan, Natalia. (2019). Breaking the Cycle of Roma Exclusion in the Western Balkans. Washington D.C.: World Bank. Roder,Ann and Elliott, Mark. (2013). Stimulating Opportunity: An Evaluation of ARRA-Funded Subsidized Employment Programs (. New York: Economic Mobility Corporation. Sadiq, K. (2008). Paper Citizens: How illegal immigrants acquire citizenship in developing countries. Oxford University Press. Salazar, M., Karacsony, S. I., Bodewig, C., Kuznetsova, D., Millan, N., Mayer Gukovas, R., & Buhler, D. C. (2018). Improving access to essential services for vulnerable groups in Bulgaria: technical report. Washington, DC: World Bank Group. Sarkhel, Sukanya et al. (2014). Culture, discrimination and women’s work force participation: a study on Indian labor market. Mimeo,Indian Statistical Institute, Delhi Centre. Retrieved from https://www.isid.ac.in/~epu/acegd2014/papers/SukanyaSarkhel.pdf 56 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Schultz, D. L. (2012). Translating intersectionality theory into practice: A tale of Roma-Gadže feminist alliance. Signs: Journal of Women in Culture and Society, 38(1), (pp. 37-43). Secretariat Foundation (2015). Roma Inclusion Index 2015 Shane, D., & Trivedi, P. (2012). What Drives Differences in Health Care Demand? The Role of Health Insurance and Selection Bias. Health, Econometrics and Data Group (HEDG) Working Papers. HEDG, c/o Department of Economics, University of York. Skoufias, E., & McClafferty, B. (2001). Is PROGRESA Working? The Summary Results of an Evaluation by IFPRI. (IFPRI, Ed.) FCND Discussion Paper No. 118. Statistical Office of the Republic of Serbia and UNICEF. (2014). Serbia Multiple Indicator Cluster Survey and Serbia Roma Settlements Multiple Indicator Cluster Survey, 2014 Final Reports. Belgrade: Statistical Office of the Republic of Serbia. Sukanya, S. e. (2014). Culture, discrimination and women’s work force participation: a study on Indian labor market. Mimeo,Indian Statistical Institute, Delhi Centre. Surdu, M. (2016). Those Who Count. Central European University Press. The World Bank and Kosovo Agency of Statistics. (2017). Consumption Poverty in the Republic of Kosovo, 2012- 2015. The World Bank and Statistical Office of the Republic of Serbia. (2016). Poverty Map of Serbia. Washington, DC. UN Women. (2015). Assessment of the Structural Barriers that Hinder the Employment of Roma Women. Strengthening Social Cohesion in the Labor Market through Support to Disadvantaged and Vulnerable Groups Project. UN, Economic Commission for Europe. (2003). Youth in the UNECE Region: Realities, Challenges and Opportunities. UNDP, World Bank, and EC. (2011). Regional Roma Survey. UNDP, World Bank, and EC. (2017). Regional Roma Survey. UNFPA. (2014). Child Marriage in Kosovo (overview). UNHCR. (2016). Persons at Risk of Statelessness in Serbia: Progress Report 2010 – 2015. Belgrade: Slobodan Cvejić. Retrieved from http://www.refworld.org/docid/57bd436b4.html UNHCR/UN OCHA. (2014). Analysis of the Situation of Internally Displaced Persons from Kosovo in Serbia and Montenegro: Law and Practice. A Legal Analysis. IDP Interagency Working Group. UNICEF (2014). Realizing the rights of Roma children and women in Bosnia and Herzegovina, the former Yugoslav Republic of Macedonia, and Serbia. UNICEF Insights Issue 2/ 2014. https://www.unicef.org/ceecis/Insights2014_2.pdf UNICEF. (2015). The Rights of Roma Children and Women in Bosnia and Herzegovina, the former Yugoslav Republic of Macedonia, and Serbia: A comparative review and further analysis of findings from MICS 57 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS surveys in Roma settlements in the three countries. UNICEF. Geneva. Retrieved from https://www.unicef.org/eca/21352_UNICEF_Roma_Woman_web_V2.pdf UNICEF. (2017). Child Marriage Among the Roma Population in Serbia: Ethnographic Research. UNICEF. Belgrade. Retrieved from https://www.unicef.org/serbia/Child_marriage_among_the_Roma_population_in_Serbia_web.pdf UNICEF, MICS. (2011). FYR Macedonia Multiple Cluster Indicator Survey. UNICEF, MICS. (2013). Montenegro Multiple Indicator Cluster Survey. UNICEF, MICS. (2014). Serbia Multiple Indicator Cluster Survey. Vella, Francis; Farre, Lidia. (2013). The Intergenerational Transmission of Gender Role Attitudes and its Implications for Female Labour Force Participation. Economica, 80(318), 219-247. Verick, S. (2014). Female labor force participation in developing countries. IZA World of Labor. Retrieved from https://wol.iza.org/uploads/articles/87/pdfs/female-labor-force-participation-in-developing-countries.pdf West, C., & Zimmerman, D. H. (1987). Doing Gender. Gender & Society 1(2), (pp. 125-151). Wodon, Q.T., Tavares, P.M.T., Fiala, O., Nestour, A.L. and Lisa, W. (2017). Ending child marriage: Child Marriage Laws and their Limitations. The World Bank, Save the Children, Children´s Investment Fund Foundation, Global Partnership for Education. Wodon, Q. T., De La Briere, B. 2018. Unrealized potential : the high cost of gender inequality in earnings (English). The cost of gender inequality. Washington, D.C. : World Bank Group. http://documents.worldbank.org/curated/en/172021527258723053/Unrealized-potential-the-high-cost-of- gender-inequality-in-earnings World Bank. (2010). Roma Inclusion: An Economic Opportunity for Bulgaria, Czech Republic, Romania and Serbia. Washington DC: World Bank. Word Bank. (2012a). Toward an Equal Start: Closing the Early Learning Gap for Roma Children in Eastern Europe. Washington, DC: The World Bank. World Bank. (2012b). Reducing vulnerability and Promoting Self-Employment of Roma in Eastern Europe through Financial Inclusion. Policy Research Working Paper No 72331. World Bank (2012c). Kosovo: Gender Gaps in Education, Health and Economic Opportunities. The World Bank. World Bank (2012d). World Development Report 2012 : Gender Equality and Development. World Bank. © World Bank. https://openknowledge.worldbank.org/handle/10986/4391 License: CC BY 3.0 IGO. World Bank (2013a). FYR Macedonia: Gender Diagnostic: Gaps in Endowments, Access to Economic Opportunities and Agency. The World Bank. World Bank (2013b). On norms and agency: Conversations about gender equality with women and men in 20 countries. Ed. by Boudet, A. M. M., Petesch, P., & Turk, C. World Bank Publications, Washington DC. 58 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS World Bank. (2014a). Diagnostics and Policy Advice for Supporting Roma Inclusion in Romania. Washington, DC: The World Bank Group. World Bank. (2014b). Handbook for improving the living conditions of Roma (English). Washington, DC: World Bank. World Bank (2014c). Gender Dimensions of Roma Inclusion: Perspectives from Four Roma Communities in Bulgaria. Report, World Bank, Washington, DC. World Bank. (2014d). Global Case studies: Lessons from interventions for Improving the living conditions of marginalized communities. Washington,DC: World Bank. World Bank. (2014e). Handbook for improving the living conditions of Roma (English). Washington, DC: World Bank. World Bank. (2015a). Republic of Serbia: Country Partnership Framework 2016 – 2020. (pp. Report No. 100464- YF). Washington DC: World Bank. Retrieved from http://documents.worldbank.org/curated/en/650701467993479160/Serbia-Country-partnership- framework-for-the-period-FY2016-20 World Bank. (2015b). Roma in Serbia – A Generation of Opportunities: The Economic and Fiscal Benefits of Roma Inclusion in the Western Balkans. mimeo. The World Bank. World Bank. (2015c). Serbia Systematic Country Diagnostic. Washington, DC: World Bank. World Bank (2015d). Bosnia and Herzegovina. Gender Disparities in Endowments, Access to Economic Opportunities and Agency. The World Bank. World Bank. (2015e). Why should we care about care?The role of childcare and eldercare in Bosnia and Herzegovina. Washington, DC: World Bank. World Bank. (2015f). Why should we care about care?The role of childcare and eldercare in former Yugoslav Republic of Macedonia. Washington, DC: World Bank. World Bank. (2015g). Why should we care about care?The role of childcare and eldercare in Kosovo. Washington, DC: World Bank. World Bank Group. (2015h). Why Should We Care About Care? : The Role of Informal Childcare and Eldercare in Aging Societies. Washington, DC: World Bank. Retrieved from https://openknowledge.worldbank.org/handle/10986/22814 World Bank. (2016a). Montenegro Systematic Country Diagnostic: Achieving Sustainable and Inclusive Growth amidst High Volatility. Washington, DC: World Bank. World Bank. (2016b). Youth in Action for Inclusive Growth - Improving Employment Opportunities for Vulnerable Youth in Bulgaria. Policy Note. Washington DC: World Bank. World Bank. (2016c). Women´s Access to Economic Opportunities in Serbia. The World Bank. World Bank. (2016d). Bosnia and Herzegovina. Economic mobility, Jobs and Gender. The World Bank. 59 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS World Bank. (2016e). Why should we care about care?The role of childcare and eldercare in former Yugoslav Republic of Macedonia. Washington, DC: World Bank. World Bank (2016f). What works for Women in the Western Balkans: a stocktaking of economic empowerment project, June 2016. World Bank Group. (2017a). Promoting Women's Access to Economic Opportunities: Bosnia and Herzegovina. Washington, DC: World Bank. Retrieved from https://openknowledge.worldbank.org/handle/10986/29590 World Bank Group (2017b). Promoting Women´s Access to Economic Opportunities. FYR Macedonia. Washington, DC: World Bank. World Bank Group. (2017c). Promoting Women's Access to Economic Opportunities: Kosovo. Washington, DC: World Bank. Retrieved from https://openknowledge.worldbank.org/handle/10986/29587 World Bank Group. (2017d). Promoting Women's Access to Economic Opportunities: Serbia. Washington, DC: World Bank. Retrieved from https://openknowledge.worldbank.org/handle/10986/29591 World Bank. (2017e). Serbia - Inclusive Early Childhood Education and Care Project (English). Washington DC: World Bank. Retrieved from http://documents.worldbank.org/curated/en/918021487905271830/Serbia- Inclusive-Early-Childhood-Education-and-Care-Project World Bank Group. (2017f). The Western Balkans: Revving Up the Engines of Growth and Prosperity. Washington, DC: World Bank. Retrieved from https://openknowledge.worldbank.org/handle/10986/28894 World Bank. (2018a). Serbia Policy Note: Social Protection and Labor. Belgrade: World Bank. World Bank (2018b). Western Balkans Labor Market Trends 2018. Washington, DC: World Bank. Retrieved from http://documents.worldbank.org/curated/en/565231521435487923/pdf/124354-WP-P158508-PUBLIC- 16-3-2018-10-13-39-WBLabormarkettrends.pdf World Bank (2019). Reform Momentum Needed. Western Balkans Regular Economic Report. Washington, DC: World Bank. World Bank. (forthcoming). Ready, Set, Go! Supporting early childhood development of Roma children by the Roma Education Fund, Project Evaluation Report. Washington, DC: World Bank. World Bank. (forthcoming). Supporting the Effective Reintegration of Roma Returnees in the Western Balkans. Washington, DC: The World Bank. 60 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Appendix. Regional Overview Tables by Gender Table 6. Regional Overview at a Glance: 2011 Roma Levels by Sex Roma Females Roma Males Roma Gender Gap Priority Area Indicator ALB BIH MKD MNE SRB ALB BIH MKD MNE SRB ALB BIH MKD MNE SRB Net pre-primary enrollment rate (ages 3-5) 30 5 13 13 9 38 8 11 4 7 8 3 -2 -9 -2 Adjusted net compulsory education enrollment rate (ages 7-15)† 49 59 72 57 80 49 62 74 55 80 0 3 2 -2 0 Compulsory education completion rate (ages 18-21) 22 43 53 28 45 25 45 60 41 61 3 2 7 13 16 Upper secondary education completion rate Education (ages 22-25) 4 12 14 2 10 3 19 16 9 16 -1 6 2 7 6 Tertiary education completion rate (ages 26- 29) 0 0 0 0 1 1 0 0 0 0 1 0 0 0 -1 Percentage of students attending majority Roma schools (ages 7-15)†† 17 6 24 10 8 11 5 23 7 8 -6 -1 -1 -3 0 Percentage of students attending special schools (ages 7-15)†† 0 1 2 3 6 2 2 6 8 8 2 0 4 5 2 Labor force participation rate (ages 15-64) 37 26 35 27 38 71 57 63 71 65 33 31 28 43 28 Employment to population ratio (ages 15- 64) 25 5 11 9 12 60 31 34 48 40 35 25 24 39 28 Unemployment rate (% of total labor force, Labor Markets ages 15-64)†† 34 80 70 69 67 15 46 45 32 39 -18 -34 -24 -37 -28 Informal employment (% of total employment)†† 80 77 68 64 76 88 81 63 53 74 8 4 -6 -10 -1 Not in employment, education, or training (NEET) (ages 15-24)†† 80 84 79 85 78 49 69 61 55 57 -31 -15 -18 -30 -21 Health insurance coverage (ages 16+) 35 74 94 88 95 28 65 90 91 88 -7 -9 -3 4 -7 Self-reported unmet need for medical care (% of population ages 16+) 58 43 35 10 33 47 35 30 13 28 -11 -8 -5 3 -5 Health Self-perceived health (% of population ages 16+ reporting good or very good health) 63 52 65 76 51 70 59 70 79 59 7 7 5 2 8 Use of preventive health care services (% of population ages 16+) 43 53 63 44 60 44 39 52 34 57 1 -14 -11 -10 -4 Electricity (% of population) 94 83 97 91 84 93 84 97 90 83 0 1 0 -1 -1 Piped water inside the dwelling (% of population) 46 87 89 79 70 44 87 90 80 72 -2 0 1 1 2 Connection to public sewerage or waste water tank (% of population) 71 69 83 65 59 71 70 82 66 59 1 0 0 1 0 Housing Waste never collected (% of population)†† 30 25 18 18 25 30 25 17 19 24 0 1 -1 1 -1 Rooms per household member††† 0.4 0.6 0.7 0.8 0.7 0.4 0.6 0.6 0.6 0.6 0.0 0.0 -0.1 -0.2 -0.2 Overcrowding rate (% of population) 84 74 67 72 74 81 74 68 71 75 -2 0 0 -1 0 Birth certificate (% of population) 97 97 98 94 99 97 98 99 96 99 0 0 0 2 0 Documentation ID card (% of population ages 16+, MKD 18+) 86 91 95 76 94 84 92 97 86 93 -2 1 2 10 -1 Source: World Bank estimates based on weighted 2011 UNDP-WB-EC Regional Roma Survey data. Notes: (1) All values shown are in percent, except for rooms per household member. (2) Cells in blue refer to gaps that are statistically significant at the 10 percent level. † Compulsory education refers to ISCED levels 1 and 2. †† A positive (negative) change in the gap for this indicator implies a reduction (increase) in inequality. ††† This indicator is calculated at the level of the head of household. 61 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Table 7. Regional Overview at a Glance: 2011 Non-Roma Levels by Sex Non-Roma Neighbor Females Non-Roma Neighbor Males Non-Roma Neighbor Gender Gap Priority Area ALB BIH MKD MNE SRB ALB BIH MKD MNE SRB ALB BIH MKD MNE SRB Net pre-primary enrollment rate (ages 3- 5) 55 10 35 20 41 50 0 15 13 29 -6 -10 -20 -8 -12 Adjusted net compulsory education enrollment rate (ages 7-15)† 93 99 86 92 96 90 91 94 97 94 -3 -7 7 5 -2 Compulsory education completion rate (ages 18-21) 87 94 89 98 100 82 89 90 91 100 -4 -5 1 -7 0 Upper secondary education completion Education rate (ages 22-25) 53 84 65 88 85 43 83 83 90 79 -10 -1 18 2 -6 Tertiary education completion rate (ages 26-29) 14 19 34 21 12 5 15 11 8 13 -9 -4 -23 -13 1 Percentage of students attending majority Roma schools (ages 7-15)†† 0 11 13 0 5 3 7 12 1 8 3 -4 0 1 3 Percentage of students attending special schools (ages 7-15)†† 4 2 3 1 2 2 0 5 4 5 -2 -2 2 3 4 Labor force participation rate (ages 15- 64) 40 36 39 50 50 70 62 63 74 69 31 26 23 24 19 Employment to population ratio (ages 15- 64) 31 24 26 32 32 60 45 49 55 55 29 22 23 23 23 Unemployment rate (% of total labor Labor Markets force, ages 15-64)†† 22 35 34 35 37 15 28 22 25 20 -7 -7 -12 -9 -17 Informal employment (% of total employment)†† 55 17 17 10 22 67 18 25 18 29 12 0 8 8 7 Not in employment, education, or training (NEET) (ages 15-24)†† 44 30 42 41 35 26 30 29 42 23 -18 0 -14 1 -12 Health insurance coverage (ages 16+) 52 94 97 100 96 57 97 96 98 91 5 2 -1 -2 -6 Self-reported unmet need for medical care (% of population ages 16+) 34 24 21 5 23 29 20 16 5 17 -5 -5 -5 1 -6 Health Self-perceived health (% of population ages 16+ reporting good or very good 70 61 73 76 59 79 65 79 80 67 10 5 6 4 8 health) Use of preventive health care services (% of population ages 16+) 56 69 77 68 72 59 56 63 52 63 3 -13 -13 -16 -9 Electricity (% of population) 96 98 95 99 98 95 98 94 99 98 0 0 0 0 0 Piped water inside the dwelling (% of population) 71 99 99 94 98 70 99 98 93 98 -1 0 -1 -1 0 Connection to public sewerage or waste water tank (% of population) 78 90 91 94 87 75 90 90 94 85 -3 0 -1 0 -2 Housing Waste never collected (% of population)†† 31 27 9 28 21 31 24 10 29 21 0 -3 1 0 0 Rooms per household member††† 0.9 1.2 1.3 1.3 1.3 0.6 1.0 1.0 1.1 1.0 0.0 0.0 -0.1 -0.2 -0.2 Overcrowding rate (% of population) 60 30 36 37 33 59 30 33 36 31 0 0 -3 -1 -2 Birth certificate (% of population) 99 100 100 98 100 99 100 99 97 100 0 0 0 -1 0 Documentation ID card (% of population ages 16+, MKD 18+) 88 97 98 95 98 92 95 98 94 98 3 -1 0 -1 0 Source: World Bank estimates based on weighted 2011 UNDP-WB-EC Regional Roma Survey data. Notes: (1) All values shown are in percent, except for rooms per household member. (2) Cells in blue refer to gaps that are statistically significant at the 10 percent level. † Compulsory education refers to ISCED levels 1 and 2. †† A positive (negative) change in the gap for this indicator implies a reduction (increase) in inequality. ††† This indicator is calculated at the level of the head of household. 62 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Table 8. Regional Overview at a Glance: 2017 Roma Levels by Sex Priority Area Indicator Roma Females Roma Males Roma Gender Gap ALB BIH MKD MNE SRB KSV ALB BIH MKD MNE SRB KSV ALB BIH MKD MNE SRB KSV Net pre-primary enrollment rate (ages 3- 5) 30 1 9 19 8 12 35 6 11 23 10 15 4 5 2 5 2 3 Adjusted net compulsory education enrollment rate (ages 7-15)† 66 70 77 61 85 72 66 73 78 63 83 74 0 2 1 2 -3 2 Compulsory education completion rate (ages 18-21) 39 36 63 30 63 55 47 50 77 38 72 67 8 14 14 7 9 12 Upper secondary education completion Education rate (ages 22-25) 14 18 32 2 9 15 15 23 33 4 25 27 1 4 1 2 16 12 Tertiary education completion rate (ages 26-29) 1 0 4 0 2 3 2 0 2 0 0 2 1 0 -2 0 -2 0 Percentage of students attending majority Roma schools (ages 7-15)†† 14 15 32 15 9 15 23 9 46 15 12 12 9 -6 14 0 3 -3 Percentage of students attending special schools (ages 7-15)†† 1 0 2 0 1 2 1 0 3 1 4 2 0 0 1 0 2 1 Labor force participation rate (ages 15- 64) 29 13 30 5 19 9 51 38 56 33 49 42 23 25 25 28 30 33 Employment to population ratio (ages 15- 64) 11 3 13 3 11 4 24 19 31 26 33 22 13 15 18 23 22 18 Unemployment rate (% of total labor Labor Markets force, ages 15-64)†† 63 73 58 33 45 53 53 51 45 21 32 47 -10 -22 -13 -12 -13 -5 Informal employment (% of total employment)†† 47 57 31 73 51 43 69 62 41 57 69 75 22 5 11 -16 17 31 Not in employment, education, or training (NEET) (ages 15-24)†† 86 86 75 88 75 77 67 77 60 70 51 63 -20 -10 -15 -18 -24 -14 Health insurance coverage (ages 16+) 28 74 95 78 92 10 26 72 93 82 92 10 -2 -2 -2 3 -1 0 Self-reported unmet need for medical care (% of population ages 16+) 37 40 22 27 30 27 34 24 11 18 23 25 -3 -16 -11 -9 -7 -2 Health Self-perceived health (% of population ages 16+ reporting good or very good health) 57 56 56 60 47 67 63 59 59 65 58 70 6 3 3 5 11 3 Use of preventive health care services (% of population ages 16+) 48 51 60 48 64 55 39 47 50 47 52 54 -9 -3 -10 -2 -12 -2 Electricity (% of population) 82 90 94 92 88 89 85 90 93 92 87 89 3 0 -1 1 0 0 Piped water inside the dwelling (% of population) 46 87 91 73 81 84 47 86 90 73 80 85 2 -1 -2 0 -2 1 Connection to public sewerage or waste water tank (% of population) 62 69 84 55 66 75 63 68 83 56 64 75 1 -1 -1 1 -2 0 Housing Waste never collected (% of population)†† 6 20 5 20 21 9 5 18 6 21 24 8 -1 -1 0 1 3 -2 Rooms per household member††† 0.9 0.8 1.0 0.6 0.9 0.8 0.6 0.7 0.7 0.5 0.7 0.6 -0.2 -0.1 -0.3 -0.1 -0.2 -0.1 Overcrowding rate (% of population) 65 64 61 77 66 71 66 66 60 78 64 71 1 2 -1 1 -2 0 Birth certificate (% of population) 98 98 98 96 99 97 98 99 99 97 99 97 0 1 1 0 0 0 Documentation ID card (% of population ages 16+, MKD 18+) 91 94 95 81 94 88 90 93 94 85 94 91 -1 0 -1 4 0 3 Source: World Bank estimates based on weighted 2017 UNDP-WB-EC Regional Roma Survey data. Notes: (1) All values shown are in percent, except for rooms per household member. (2) Cells in blue refer to gaps that are statistically significant at the 10 percent level. † Compulsory education refers to ISCED levels 1 and 2. †† A positive (negative) change in the gap for this indicator implies a reduction (increase) in inequality. ††† This indicator is calculated at the level of the head of household. 63 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Table 9. Regional Overview at a Glance: 2017 Non-Roma Neighbor Levels by Sex Priority Area Indicator Non-Roma Neighbor Females Non-Roma Neighbor Males Non-Roma Neighbor Gender Gap ALB BIH MKD MNE SRB KSV ALB BIH MKD MNE SRB KSV ALB BIH MKD MNE SRB KSV Net pre-primary enrollment rate (ages 3- 5) 68 0 33 30 39 5 56 4 21 41 19 27 -12 4 -12 11 -20 22 Adjusted net compulsory education enrollment rate (ages 7-15)† 96 93 81 92 99 90 97 88 94 91 98 95 1 -5 13 -2 -1 5 Compulsory education completion rate (ages 18-21) 96 94 96 96 96 94 100 96 91 98 97 96 4 2 -6 2 1 2 Upper secondary education completion Education rate (ages 22-25) 72 82 88 83 95 71 77 91 86 77 89 83 6 9 -2 -6 -6 12 Tertiary education completion rate (ages 26-29) 37 23 35 32 26 33 12 18 26 17 6 17 -25 -6 -9 -15 -19 -16 Percentage of students attending majority Roma schools (ages 7-15)†† 14 6 12 12 4 3 8 5 12 14 18 7 -6 0 0 2 14 4 Percentage of students attending special schools (ages 7-15)†† 3 0 5 3 0 1 0 0 8 0 0 2 -3 0 3 -3 0 1 Labor force participation rate (ages 15- 64) 34 27 36 28 42 13 49 56 60 55 63 45 15 29 24 28 21 32 Employment to population ratio (ages 15- 64) 21 18 26 26 35 6 30 42 52 51 53 32 9 24 26 25 18 26 Unemployment rate (% of total labor Labor Market force, ages 15-64)†† 37 32 26 6 16 49 39 25 14 7 16 27 1 -7 -12 1 1 -22 Informal employment (% of total employment)†† 11 12 10 27 11 27 30 17 19 30 16 49 19 5 9 3 5 22 Not in employment, education, or training (NEET) (ages 15-24)†† 26 41 27 37 31 46 46 45 27 30 36 35 20 3 0 -7 6 -11 Health insurance coverage (ages 16+) 43 95 97 98 99 10 44 94 97 98 96 13 0 -2 0 0 -3 3 Self-reported unmet need for medical care (% of population ages 16+) 22 16 11 11 20 28 22 15 5 7 12 17 0 -1 -6 -4 -8 -11 Health Self-perceived health (% of population ages 16+ reporting good or very good health) 67 64 61 66 56 71 72 67 62 73 63 81 5 3 1 7 8 10 Use of preventive health care services (% of population ages 16+) 65 74 76 81 82 65 69 57 67 77 59 62 4 -17 -10 -4 -23 -3 Electricity (% of population) 93 99 97 98 98 97 95 98 97 98 97 97 1 -1 0 0 -1 -1 Piped water inside the dwelling (% of population) 89 97 97 95 98 97 90 96 97 96 95 96 1 -1 0 0 -3 -1 Connection to public sewerage or waste water tank (% of population) 75 81 92 84 80 91 78 80 93 84 79 91 3 -1 1 0 -1 0 Housing Waste never collected (% of population)†† 2 6 2 19 14 2 3 7 2 18 16 2 0 1 0 -1 1 -1 Rooms per household member††† 1.2 1.5 1.3 1.3 1.6 0.8 1.1 1.2 1.0 1.1 1.1 0.8 -0.1 -0.3 -0.4 -0.2 -0.5 0.0 Overcrowding rate (% of population) 32 23 31 41 22 53 27 20 26 40 23 49 -6 -3 -5 -1 1 -4 Birth certificate (% of population) 99 100 99 100 100 98 100 100 100 100 100 98 1 0 1 0 0 0 Documentation ID card (% of population ages 16+, MKD 18+) 96 96 96 96 97 95 98 96 97 96 98 94 2 -1 1 0 1 -1 Source: World Bank estimates based on weighted 2017 UNDP-WB-EC Regional Roma Survey data. Notes: (1) All values shown are in percent, except for rooms per household member. (2) Cells in blue refer to gaps that are statistically significant at the 10 percent level. † Compulsory education refers to ISCED levels 1 and 2. †† A positive (negative) change in the gap for this indicator implies a reduction (increase) in inequality. ††† This indicator is calculated at the level of the head of household. 64 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Table 10. Regional Overview at a Glance: Roma 2011-2017 Percentage Point Changes by Sex Priority Area Indicator Change in Roma Female Coverage Change in Roma Male Coverage Change in Roma Gender Gap ALB BIH MKD MNE SRB ALB BIH MKD MNE SRB ALB BIH MKD MNE SRB Net pre-primary enrollment rate (ages 3- 5) 1 -4 -4 5 -1 -3 -2 0 19 3 -3 2 4 13 4 Adjusted net compulsory education enrollment rate (ages 7-15)† 17 12 5 4 6 17 11 4 8 3 0 -1 -1 5 -2 Compulsory education completion rate (ages 18-21) 17 -7 10 3 18 23 5 16 -3 11 6 12 7 -6 -7 Upper secondary education completion Education rate (ages 22-25) 10 6 17 0 -1 12 4 17 -5 9 2 -2 0 -4 10 Tertiary education completion rate (ages 26-29) 1 0 4 0 1 1 0 2 0 0 0 0 -2 0 -1 Percentage of students attending majority Roma schools (ages 7-15)†† -2 9 8 5 1 12 4 23 8 4 14 -5 15 3 3 Percentage of students attending special schools (ages 7-15)†† 1 -1 0 -2 -5 -1 -2 -3 -8 -4 -2 -1 -3 -5 1 Labor force participation rate (ages 15- 64) -8 -13 -4 -22 -18 -19 -19 -7 -38 -16 -11 -6 -2 -15 2 Employment to population ratio (ages 15- 64) -14 -2 2 -5 -2 -35 -12 -4 -22 -7 -21 -10 -6 -17 -5 Unemployment rate (% of total labor Labor Market force, ages 15-64)†† 29 -7 -12 -36 -22 37 4 0 -11 -6 8 11 12 25 16 Informal employment (% of total employment)†† -33 -20 -38 9 -24 -19 -19 -21 3 -6 15 2 16 -6 18 Not in employment, education, or training (NEET) (ages 15-24)†† 6 3 -4 3 -3 18 8 -1 15 -6 12 5 4 12 -3 Health insurance coverage (ages 16+) -7 0 2 -9 -3 -2 7 3 -10 3 4 7 1 0 6 Self-reported unmet need for medical care (% of population ages 16+) -20 -3 -13 17 -3 -12 -11 -19 5 -4 8 -7 -6 -11 -2 Health Self-perceived health (% of population ages 16+ reporting good or very good health) -7 4 -9 -16 -4 -8 0 -11 -13 -1 -1 -4 -2 3 3 Use of preventive health care services (% of population ages 16+) 5 -2 -2 5 3 -5 8 -2 13 -5 -10 10 0 8 -8 Electricity (% of population) -11 7 -3 0 4 -8 6 -4 2 4 3 -1 -1 2 0 Piped water inside the dwelling (% of population) 0 0 2 -6 12 3 -1 0 -7 8 4 -1 -3 -1 -4 Connection to public sewerage or waste water tank (% of population) -9 -1 2 -11 7 -8 -2 1 -10 6 1 -1 -1 0 -2 Housing Waste never collected (% of population)†† -24 -5 -13 2 -5 -25 -7 -12 3 0 -1 -2 2 0 5 Rooms per household member††† 0.4 0.2 0.3 -0.2 0.2 0.2 0.1 0.1 -0.1 0.2 -0.3 0.0 -0.2 0.1 0.0 Overcrowding rate (% of population) -18 -10 -6 5 -9 -15 -7 -7 7 -10 4 3 -1 2 -2 Birth certificate (% of population) 1 1 0 2 0 1 1 0 0 0 0 0 1 -1 0 Documentation ID card (% of population ages 16+, MKD 18+) 5 3 0 4 0 6 1 -3 -1 1 1 -1 -3 -6 1 Source: World Bank estimates based on weighted 2011 and 2017 UNDP-WB-EC Regional Roma Survey data. Notes: (1) All values shown are in percent, except for rooms per household member. (2) Green, red and yellow cells denote changes in the indicators between 2011 and 2017. Green denotes an improvement; red denotes a worsening; yellow denotes no statistically significant change at the 10 percent level. † Compulsory education refers to ISCED levels 1 and 2. †† A positive change in the gap for this indicator implies a reduction in inequality. ††† This indicator is calculated at the level of the head of household. 65 CLOSING THE GENDER GAPS AMONG MARGINALIZED ROMA IN THE WESTERN BALKANS Table 11. Regional Overview at a Glance: Non-Roma neighbor 2011-2017 Percentage Point Changes by Sex Change in Non-Roma neighbor Female Change in Non-Roma neighbor Male Change in Non-Roma neighbor Gender Priority Area Indicator Coverage Coverage Gap ALB BIH MKD MNE SRB ALB BIH BIHMKD MNE SRB ALB BIH MKD MNE SRB Net pre-primary enrollment rate (ages 3- 5) 13 -10 -2 10 -1 7 4 6 28 -10 -6 14 8 18 -8 Adjusted net compulsory education enrollment rate (ages 7-15)† 3 -6 -5 0 3 7 -3 1 -7 4 4 2 6 -7 1 Compulsory education completion rate (ages 18-21) 9 -1 8 -2 -4 18 6 1 7 -3 9 7 -6 9 1 Upper secondary education completion Education rate (ages 22-25) 19 -2 23 -5 10 35 8 3 -13 10 16 9 -21 -8 0 Tertiary education completion rate (ages 26-29) 23 5 1 11 14 7 3 15 9 -7 -16 -2 14 -2 -20 Percentage of students attending majority Roma schools (ages 7-15)†† 14 -5 -1 12 0 5 -2 -1 12 11 -9 3 0 1 11 Percentage of students attending special schools (ages 7-15)†† -1 -2 2 2 -2 -2 0 3 -4 -5 -1 2 1 -6 -4 Labor force participation rate (ages 15- 64) -5 -10 -3 -22 -9 -21 -7 -2 -19 -6 -16 3 1 3 3 Employment to population ratio (ages 15- 64) -9 -6 1 -7 3 -30 -3 3 -4 -2 -20 2 3 2 -6 Unemployment rate (% of total labor Labor Market force, ages 15-64)†† 15 -3 -8 -28 -21 24 -3 -8 -18 -4 8 0 0 10 17 Informal employment (% of total employment)†† -44 -6 -7 17 -12 -37 -1 -6 13 -13 7 5 2 -5 -2 Not in employment, education, or training (NEET) (ages 15-24)†† -18 11 -15 -4 -4 20 15 -2 -12 13 38 3 14 -8 18 Health insurance coverage (ages 16+) -9 1 0 -1 3 -14 -3 0 1 5 -5 -4 0 2 3 Self-reported unmet need for medical care (% of population ages 16+) -12 -9 -10 7 -3 -6 -5 -11 2 -5 5 4 -1 -5 -2 Health Self-perceived health (% of population ages 16+ reporting good or very good health) -3 3 -12 -10 -4 -7 2 -17 -7 -4 -4 -2 -5 3 0 Use of preventive health care services (% of population ages 16+) 9 4 0 12 10 11 1 3 25 -4 1 -4 4 12 -14 Electricity (% of population) -2 0 2 -1 0 -1 0 3 -1 -1 2 0 1 0 -1 Piped water inside the dwelling (% of population) 18 -2 -2 1 0 19 -3 -1 2 -2 1 -1 0 1 -3 Connection to public sewerage or waste water tank (% of population) -3 -9 1 -10 -7 3 -10 3 -9 -6 6 -1 2 0 1 Housing Waste never collected (% of population)†† -28 -20 -6 -9 -6 -28 -17 -8 -10 -5 0 3 -1 -1 1 Rooms per household member††† 0.3 0.3 0.1 0.0 0.4 0.5 0.2 0.0 0.0 0.1 0.2 -0.1 -0.1 0.0 -0.2 Overcrowding rate (% of population) -27 -7 -4 3 -11 -33 -10 -6 4 -9 -5 -3 -2 0 2 Birth certificate (% of population) 0 0 -1 2 0 1 0 0 3 0 1 0 1 1 0 Documentation ID card (% of population ages 16+, MKD 18+) 8 0 -2 1 -1 6 0 -2 1 0 -2 1 0 0 1 Source: World Bank estimates based on weighted 2011 and 2017 UNDP-WB-EC Regional Roma Survey data. Notes: (1) All values shown are in percent, except for rooms per household member. (2) Green, red and yellow cells denote changes in the indicators between 2011 and 2017. Green denotes an improvement; red denotes a worsening; yellow denotes no statistically significant change at the 10 percent level. † Compulsory education refers to ISCED levels 1 and 2. †† A positive change in the gap for this indicator implies a reduction in inequality. ††† This indicator is calculated at the level of the head of household. 66