This project is funded by the European Union Breaking the Cycle of Roma Exclusion in the Western Balkans E xecutive Summary Breaking the Cycle of Roma Exclusion in the Western Balkans Monica Robayo-Abril Natalia Millán March 2019 This project is funded by the European Union © 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. This publication was produced with the financial support of the European Union. Its contents are the sole responsibility of World Bank staff and do not necessarily reflect the views of the European Union. For the European Union, the designation of Kosovo is without prejudice to positions on status and is in line with UNSCR 1244/1999 and the ICJ Opinion on the Kosovo Declaration of Independence. 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. Cover photo: Jutta Benzenberg Typesetting: Budy Wirasmo BREAKING THE CYCLE OF ROMA EXCLUSION IN THE WESTERN BALKANS Acknowledgments This work was prepared by a multisectoral team from the Poverty and Equity Global Practice and the Social Protection and Jobs Global Practice. The lead authors are Monica Robayo-Abril and Natalia Millán. The authors acknowledge Céline Ferré and Paul Corral for their significant contributions to Chapter 4 and thank Trinidad Saavedra and Wenqing Zhu for excellent research assistance. The team is indebted to Timothy Johnston, Jamele Rigolini, Marco Hernandez, Marco Mantovanelli, Stephen Ndegwa, Maryam Salim, and Emanuel Salinas Munoz for their support and advice, and to Roberta Gatti, Joost de Laat, Ana María Muñoz, and Dena Ringold for their peer review comments. The team would also like to thank useful discussion and comments from (in alphabetical order): Reena Badiani-Magnusson, César Cancho, Dorothee Chen, María Dávalos, Ian Forde, Diego Garrido Martin, Maddalena Honorati, Marijana Jasarevic, Sandor Karacsony, Stefanie Koettl-Brodmann, Luis Felipe Lopez-Calva, Denis Mesihovic, Valerie Morrica, Miriam Mueller, Bojana Naceva, Trang Nguyen, María Beatriz Orlando, Mariam Sherman, Susanna Smets, and Stavros Stavrou. Administrative support was provided to the team by: Armanda Carcani, Amara Khiev, and Ngoc-Dung Thi Tran. Editing was carried out by Robert Zimmermann. The team received invaluable guidance from Katarina Mathernova, Marta Garcia-Fidalgo, and Enrica Chiozza in Brussels, as well as, in each jurisdiction, from desk officers and members of the European Commission Delegations for the Western Balkans from the European Commission Directorate- General for Neighbourhood and Enlargement Negotiations (DG NEAR). The material also reflects several consultations and exchanges, including the World Bank and United Nations Development Programme (UNDP) Regional Workshop on Methods and Tools for Welfare Measurement and Diagnostics, the World Bank cross-sectoral global practice workshops, the UNDP country dissemination events, and several presentations at events in Brussels during the 2017 and 2018 European Parliament Roma Weeks attended by the partner organizations (the World Bank, UNDP, Council of Europe, the Roma Integration 2020 Regional Cooperation Council, the Roma focal points from the European Commission delegations in the Western Balkans and desk officers from the respective countries, plus several Roma civil society organizations from the Western Balkans). The report also draws on background papers based on a complementary qualitative study commissioned by the authors. The papers were prepared by Shruti Majumdar, Andrea Woodhouse, Ivan Dordevic, and Srdan Radovic. The authors are grateful for funding from the Umbrella Facility for Gender Equality which made the qualitative study and background papers possible. This report was made possible due to the generous funding from the Europe 2020 Trust Fund by DG NEAR. The report was prepared under the guidance of Linda Van Gelder (Country Director for the Western Balkans), Carlos Silva-Jauregui (Practice Manager, Poverty and Equity Global Practice, Europe and Central Asia Region), and Cem Mete (Practice Manager, Social Protection and Jobs Global Practice, Europe and Central Asia Region). Acknowledgments iii BREAKING THE CYCLE OF ROMA EXCLUSION IN THE WESTERN BALKANS Executive Summary The Roma1 are the largest ethnic minority in Europe, as well as one of the most deprived and socially excluded groups; they typically have only limited access to basic services and economic opportunities. 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. Roma have only limited access to education, economic opportunities, health care, housing and essential services, and documentation even relative to their non-Roma neighbors who live in close physical proximity. Yet, Roma inclusion is not only a moral imperative; demographic aging in Europe means that it is also smart economics. The benefits of Roma inclusion are not negligible and include the productivity gains associated with higher employment rates and labor earnings, and they include fiscal benefits through greater tax revenue and lower social assistance spending. This is particularly important in aging societies because absorbing Roma entrants into the labor force can help counteract shrinking working-age populations. Roma are a young population, and this youth bulge can be turned into a demographic dividend through proper investment in education and basic services. Without the additional investments, governments will miss a demographic window of opportunity. Because of the lack of high-quality data, research on Roma inclusion to inform evidence-based policies is scarce, and accurate data on programs implemented in the Western Balkans are needed. Ethnicity is generally not a topic in nationally representative household surveys, and administrative data are not ordinarily disaggregated by ethnicity. Roma sometimes do not like to self-identify as Roma. As a result, there is generally undercounting in censuses and undersampling in household surveys. This report aims to fill this knowledge gap and inform policy making by relying on data 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. The 2011 round of the RRS, implemented in 12 countries of Central and Eastern Europe, including five countries in the Western Balkans, is a multitopic household survey representative of communities in which the share of Roma is larger than the share of Roma in the national populations.2 It encompasses both Roma population and neighboring non-Roma. After the first wave of the RRS was implemented in 2011, the European Commission (EC) Directorate-General for Neighbourhood and Enlargement Negotiations (DG NEAR), in an effort to explore changes in core development outcomes among marginalized Roma and non-Roma who lived nearby, commissioned the United Nations Development Programme (UNDP) and the World Bank to carry out another round of the RRS in the countries of the Western Balkans in 2017. 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. In the 2011 wave, the sample was limited to Roma who live in areas in which the concentration of Roma is greater than 40 percent; in the 2017 wave, areas with a 1 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). 2 The five Western Balkan countries were Albania, Bosnia and Herzegovina, North Macedonia, Montenegro, and Serbia. Executive Summary 1 concentration of Roma ranging between 10 percent and 40 percent were also included. In both waves, non-Roma who were living in physical proximity to the sampled Roma were also interviewed; referred to as “non-Roma”, they are not representative of the national population in each country.3 Not included in the first round, Kosovo was added to the survey at this time. Parallel qualitative research conducted by the World Bank, in collaboration with the Institute of Ethnography in Serbia, also helps in understanding the underlying mechanisms behind gaps between Roma and non-Roma in education and labor markets, with special attention to gender .4 The results show that marginalized Roma in the Western Balkans do not have the endowments and assets they need nor the ability to use the assets they have efficiently and intensively to generate economic gains and climb the socioeconomic ladder. Roma face multiple barriers and constraints that hinder their ability to accumulate human capital, participate in the labor market on an equal basis, and generate economic gains. The insufficient stock and accumulation of human, physical, financial, and social capital have hindered the ability of Roma households to generate income over the life cycle. Marginalized Roma are not well endowed; the returns to education among them are low; and their ability to accumulate assets is constrained. Weak labor market engagement is also a persistent phenomenon among them and has not improved; so, their ability to generate labor income is narrow. A comparison of the 2011 and 2017 RRS data shows that little progress toward Roma inclusion was achieved in the years between the two survey rounds in the five priority areas identified by DG NEAR: education, labor markets, health, housing and essential services, and documentation. This report relies on the 2011 and 2017 rounds of the RRS and qualitative work among Roma to identify the key barriers and constraints faced by Roma in the five priority areas of Roma inclusion. The findings are summarized below, along with relevant policy recommendations. The Five Priority Areas Education Several dimensions of education are analyzed across the life cycle, including enrollments in preprimary and compulsory education and completion rates in compulsory, upper-secondary, and tertiary education. Across the region, the coverage of education among Roma is narrow, and ethnic gaps are wide. Though there was some improvement in 2011–17, gaps between Roma and their non-Roma neighbors remain substantial, especially in upper-secondary and tertiary education. Inequalities between Roma and non-Roma become apparent early in life. The financial costs are cited as the most important barrier limiting preprimary enrollment among Roma children in the Western Balkans. So, removing preschool fees and providing other financial incentives may encourage enrollment among younger Roma children. Although all governments include measures to raise 3 Members of approximately 750 Roma and 350 neighboring non-Roma households were interviewed in each country in each survey year. 4 The qualitative study was funded by the Umbrella Foundation for Gender Equality. See Appendix A for details on the methodology and design of the quantitative survey and the qualitative research. 2 Executive Summary BREAKING THE CYCLE OF ROMA EXCLUSION IN THE WESTERN BALKANS preprimary enrollment in their National Action Plans for Roma Inclusion, access to preprimary education among Roma is low, and there was little change between 2011 and 2017. In 2017, preprimary enrollment among marginalized Roma ages 3–5 ranged from only 3 percent in Bosnia and Herzegovina to 33 percent in Albania. Less early childhood stimulation at home and restricted exposure to the majority language also mean Roma children are less prepared when they enter primary school. A large majority of caregivers say they do not send their children to school because they cannot afford to. This is consistent with recent supply and demand assessments in the Western Balkans that reveal the lack of affordable services as an important constraint on school enrolment. The provision of childcare subsidies or other financial incentives to low-income households can help alleviate the affordability issue. Recent evidence from a randomized control trial implemented across 236 poor settlements in Bulgaria shows that removing kindergarten costs was the most cost-effective strategy to boost kindergarten participation. Compulsory education is not the great equalizer; there are still significant gaps in enrollments in compulsory schools, mostly driven by discrimination and restrictive social norms. In 2017, the gaps between Roma and their non-Roma neighbors in compulsory education enrollment ranged from 10 percentage points (North Macedonia) to 29 and 31 percentage points (Montenegro and Albania, respectively), even though enrollments among Roma children expanded in all countries. In Albania and Montenegro, at least a third of Roma children ages 7–15 were outside the school systems in 2017. In most countries, a large share of the gaps between Roma and their non-Roma neighbors cannot be explained by differences in observed characteristics, such as demographics and other household characteristics, suggesting that discrimination and social norms play an important role. Among Roma, family background is associated with the level of enrollment in compulsory education. The association among family background, enrollment, and academic achievement is an essential driver of inequality of opportunity. Even after controlling for income, one finds that parental background—mainly the educational attainment of the mother—is associated with higher enrollment among students of compulsory school age. More well educated mothers may be providing more inputs into their children’s education, and they may also face lower liquidity constraints, even after one controls for household income. This means that raising the educational attainment of mothers, improving parenting skills, and relaxing liquidity constraints among less well educated parents should also lead to higher compulsory school enrollment rates. Completion rates in compulsory education among Roma ages 18–21 are generally low, and the gender gaps are large. Completion rates in compulsory education range from 34 percent and 70 percent in Montenegro and North Macedonia, respectively, whereas, among non-Roma neighbors, completion rates are above 90 percent in all countries. In Bosnia and Herzegovina, Kosovo, and North Macedonia, young Roma women ages 18–21 are significantly less likely than young Roma men to complete compulsory education. Among older cohorts, gaps in upper-secondary and tertiary education completion rates relative to neighboring non-Roma are even wider; tertiary completion is almost nonexistent among Roma. In upper-secondary education, completion among Roma ages 22–25 shows a wide range across countries, from only 3 percent in Montenegro to 32 percent in North Macedonia. Progress between 2011 and 2017 was observed only in some countries. In tertiary education, there was little change between 2011 and 2017; completion was still out of reach. Executive Summary 3 The financial costs are a main barrier to enrollment in compulsory school and higher levels of education, but child marriage is also an important barrier among Roma females. Across countries, around one-half of individuals ages 6–24 who are not in school and who have, at most, completed compulsory education (International Standard Classification of Education [ISCED] 2) report that they are not attending school because of economic factors, namely, the cost of education or related expenses such as transport and books. Among girls and women ages 6–24 who have not finished compulsory education and are not in school, between 11 percent (Bosnia and Herzegovina) and around one-third (North Macedonia and Serbia) report that they are out of school because they have married. Among women ages 20-49, there is a negative correlation between marriage as a child and lower educational attainment. Travel distance to a school is not a particular barrier. According to the 2017 round of the RRS, more than 90 percent of the members of Roma households who are in kindergarten or primary school require less than 15 minutes to reach the nearest school. Lower educational aspirations among Roma seem consistent with the lower returns to education, but also the difficult path to success and the lack of role models. In all countries, the great majority of non-Roma respondents—generally, well above 70 percent—said that tertiary education is sufficient for a child. Among Roma, this was the case among fewer than 50 percent of respondents in Albania, Bosnia and Herzegovina, North Macedonia, Montenegro, and Serbia. Likewise, in all countries except North Macedonia, a significant share of Roma respondents said that ISCED 1 and 2, that is, up to lower- secondary education, represent sufficient education for a child. This view was rather more rare among non-Roma. However, these differences in aspirations are not sufficiently large to explain the observed ethnic gaps in educational attainment. For example, the share of Roma respondents who said that tertiary education is sufficient, though lower than the corresponding share among their non-Roma neighbors, was still substantial, ranging between slightly more than one-quarter in Serbia to around 60 percent in Kosovo. Yet, the tertiary education completion rates among Roma are almost null. The lack of trust in education providers affects the use of education services. This issue can potentially be addressed by relying on Roma mediators, but implementation challenges must be carefully considered. (The mediator is a valued Roma community member who connects members of the Roma community with service providers.) Across countries, most Roma students report that they attend integrated schools, although a large share still attend majority Roma schools, possibly signaling lower-quality education in the latter. Efforts to integrate schools should reflect a consideration of the implementation challenges, such as a lack of safety or a need to travel longer distances to attend an integrated school, which may be a deterrent, especially among Roma girls. A large share of Roma children ages 7–15 in the region still attend schools with a high concentration of Roma students. Among marginalized Roma students across countries in 2017, between 10 percent (in Serbia) and 40 percent (in North Macedonia) reported that they attended majority Roma schools. In contrast, their non-Roma counterparts were less likely to attend such schools. North Macedonia stands out, with a 28 percentage point gap, that is, only 12 percent of non-Roma attend majority Roma schools. Efforts to integrate schools should be combined 4 Executive Summary BREAKING THE CYCLE OF ROMA EXCLUSION IN THE WESTERN BALKANS with active antidiscrimination measures because better outcomes in integrated schools may not be forthcoming if discrimination is pervasive. Evidence of the RRS suggests that, unlike in other countries such as the Czech Republic and the Slovak Republic, Roma children are not disproportionately represented in special schools or schools for disabled children in the Western Balkans. According to the 2017 round of the RRS, only around 1 percent of marginalized Roma ages 7–15 across the region were attending special schools, while the share of non-Roma children attending special schools was similar, showing no inequality by ethnicity. The shares were also low in 2011. However, these results may reflect substantial underreporting. Though there is a lack of data on school quality and learning outcomes, education quality among students from disadvantaged backgrounds is a problem in the region, suggesting that Roma may be especially affected. Students from disadvantaged backgrounds tend to exhibit worse performance. The Western Balkan countries have a much smaller share of resilient students than other countries in Europe. In Kosovo, only 2.5 percent of students from disadvantaged backgrounds perform among the top quarter of students on the tests of the Program for International Student Assessment. Labor markets Low labor force participation and a high incidence of unemployment mean that the employment prospects of working-age Roma are poor. Roma, especially Roma females, are much less likely to participate in the labor market than their non-Roma counterparts, and they are also less likely to participate than the majority population. Unemployment is especially high among Roma. Over 50 percent of economically active working-age Roma are unemployed in Albania and in Bosnia and Herzegovina. Employment among Roma ranges from only 13 percent in Kosovo to 22 percent in North Macedonia. These are much lower than the rates among non-Roma neighbors and national averages, which are already much lower than the average employment rate in the 28 members of the European Union (EU28) of 67.6 percent.5 The data show clearly that attention should be paid to the labor market integration of Roma. On the supply side, social assistance and interhousehold transfers do to not seem to have created work disincentives among marginalized Roma in the Western Balkans. Because of the lack of data, there is no evidence on the disincentive effects of remittances on economic activity among Roma. The data of the 2017 round of the RRS show no significant differences in labor force participation between Roma living in households receiving child social benefits and social allowances relative to nonrecipients. Inactivity rates among individuals living in Roma households receiving financial assistance or alimony from other households in the country or abroad are lower among Roma households receiving transfers, except in Bosnia and Herzegovina. In some countries, substantial reliance on remittance income may have contributed to low labor force participation rates and high unemployment because of the possible impact on reservation wages, that is, the lowest wage at which an individual would be willing to accept a particular type of job. However, this cannot be tested using the available RRS data. 5 The EU28 is the current membership of the EU. Executive Summary 5 The gender gap in labor force participation among Roma is large. The gender gaps in labor force participation among Roma range from 23 percentage points in Albania to 33 percentage points in Kosovo. In all countries, the presence of children in the household and family responsibilities are the main reasons behind female inactivity, while, among males, inactivity mostly arises because of insufficient labor demand. A significant share of the gender gap remains unexplained after controlling for individual and household characteristics, point to the important role of unobserved factors, which may include social and community norms as well as discrimination. Several overlapping constraints limit employability among Roma, including lack of skills and experience, time, access to services (childcare and eldercare), limited flexible work arrangements, and adverse attitudes and social norms. Constraints to employability constitute the first and most important set of barriers faced by Roma. In addition to low educational attainment, there is limited access to vocational skills, given the lack of on-the-job training among Roma, limited entrepreneurship and social enterprise development programs, and lack of access to microfinance schemes. Lack of access to labor market information and networks, and other information constraints may affect the matching of Roma job-seekers with job vacancies, especially among those living in Roma settlements. Restrictive labor legislation, limited flexible work arrangements, and high labor taxes among low- wage and part-time workers affect the employability of Roma. Roma females face adverse attitudes in the labor market, despite the fact that they share similar employment aspirations with their male counterparts. The availability and use of institutional childcare centers are generally low in the Western Balkans, and expanding good-quality, affordable childcare may boost female labor force participation. Social and community norms are also an important constraint, as they relegate women to household and caretaking activities and also curtail female educational attainment through child marriage. Narrowing the human capital gap between Roma and non-Roma may not be sufficient to provide fair chances on the labor market because different returns to human capital signal unequal treatment. Average returns to schooling among Roma vary between 0 percent and 3.4 percent, well below the returns associated with their neighboring non-Roma counterparts as well as the average in Europe. Governments need to continue investing in education among Roma, but getting Roma into education must also be a good investment option. So, sensible policies tackling discrimination and affecting the returns to schooling are fundamental, for instance, addressing education quality and financial constraints. The timing of the policies is also important. The returns to schooling tend to emerge slowly, which means that, if policies are undertaken now to increase significantly the returns to schooling among Roma, it may take many years for Roma to catch up with their non-Roma neighbors. In the short run, returns to schooling tend to change little, and by no more than 2 percentage points in each decade. The highest returns to schooling among Roma are in tertiary education; expanding tertiary education is therefore important to improving labor market outcomes among Roma. However, this should not come at the expense of primary and secondary education because primary education is a fundamental service, and access to primary and secondary education is a prerequisite for entry into tertiary. In addition to tackling supply-side barriers to employability, labor demand constraints must also be addressed. If barriers from the worker side alone are lifted, for example, through higher educational attainment, training, and other initiatives, but demand-side constraints are not confronted, outcomes may not improve as expected. Relevant measures might involve the private sector dealing with 6 Executive Summary BREAKING THE CYCLE OF ROMA EXCLUSION IN THE WESTERN BALKANS discrimination and stereotyping, thereby supporting the hiring of Roma by companies. It might include broader affirmative action policies, wage and employment subsidies, carrot and stick measures, and so on. Information campaigns among employers would also be useful. Apart from the legal requirements, employers may possess insufficient information about the practical benefits of employing a diverse workforce. The labor market is the priority area that lags the most, mainly because of substantial deterioration in labor force participation and employment among Roma across all countries, except North Macedonia, and a lack of progress in narrowing the gaps relative to non-Roma neighbors. Bosnia and Herzegovina, Kosovo, and Montenegro exhibit especially dismal employment outcomes among Roma. This highlights the need for a larger agenda in the creation of more and better jobs throughout the Western Balkans, but it also calls for a focus on the needs of Roma, women, and vulnerable groups and on the labor market barriers these groups face, including the barriers caused by discrimination and social and community norms. This would mean (1) identifying policies and interventions that would generate jobs for disadvantaged Roma and other vulnerable groups, (2) supporting and evaluating targeted labor market interventions in favor of Roma to prioritize and implement only those that we know work, and (3) identifying and scaling up cost-effective interventions, including interventions directed at Roma women, who are much less likely than their non-Roma neighbors to participate in the labor market or to be employed. Health This section analyzes several dimensions of health and health care, including self-perceived health status, the unmet need for health care, health insurance coverage, and the use of preventive health care services. In most countries, there was little change in coverage or access to health services between 2011 and 2017, and there was also little change in inequality. Self-perceived health is poorer among marginalized Roma than among neighboring non-Roma, possibly indicating lower objective health outcomes among Roma. The RRS does not include objective health outcomes, but RRS data do show that, although the Roma population is younger than neighboring non-Roma in all countries, Roma still perceive lower average health status than their non-Roma counterparts. In all countries, the prevalence of the self-reported unmet need for medical care is substantial, and affordability and lack of health insurance are the single most important self-reported barriers contributing to health inequalities. In countries on which national data are available, unmet health needs range from only 3 percent of total health needs in North Macedonia to 9 percent in Montenegro. Among Roma, the share of unmet needs is generally larger, ranging from 16 percent in North Macedonia to 36 percent in Albania. Across countries, the majority of Roma report high costs or the lack of health insurance as the reasons behind their unmet need for medical care, and they are more likely than their non-Roma neighbors to report these reasons. The availability of health care centers does not seem to be a major barrier to access to health care services. Across countries, the vast majority of Roma report that there is a health care center in their vicinity. Executive Summary 7 In line with the most common reason cited for unmet needs for medical care, Roma are less likely to have health insurance and much more likely to report that they cannot afford the medicine needed by household members. Lack of information, differences in beliefs and attitudes about health care, and discrimination are also major factors, and observable socioeconomic characteristics also account for some of the gaps in health insurance coverage. The size of the health insurance coverage gap is larger in Albania, Bosnia and Herzegovina, and Montenegro. Health insurance coverage is particularly low in Kosovo, mostly because there is no mandatory health insurance. In all countries, though differences in individual and household characteristics help explain part of the gap in health insurance coverage, a significant part remains unexplained. In Montenegro, the lack of access to identity cards greatly affects the chances of obtaining health insurance. The use of preventive health services among marginalized Roma is limited, and cost is always the leading factor behind gaps in the reliance on preventive health care services. Across countries, only 50 percent of Roma ages 16 and above report that they have used preventive health care services in the previous 12 months. This contrasts with a higher rate of use among neighboring non-Roma, which ranges from 63 percent in Kosovo to 71 percent in North Macedonia and Serbia. Having health insurance increases the likelihood of using preventive care services in all countries except Kosovo. Housing and Access to Essential Services Many Roma live in slums or informal settlements and experience severe overcrowding; ethnic gaps persist in access to essential services, although there has been progress in some countries. The expansion of access to services, such as electricity, piped water inside the dwelling, connections to public sewerage or waste water tanks, and waste collection, has been inclusive in Bosnia and Herzegovina and Serbia, but there are still large gaps, and some areas remain underserved. The lack of secure land titling in informal settlements contributes to shortages in housing and services among many marginalized Roma. Access to other productive assets, such as financial capital, is limited, as reflected by the low use of financial services; insecure property rights also restrict the productive use of land endowments. In all six countries, relative to their non-Roma neighbors, a substantially larger share of Roma household heads report they never use banks, automated teller machines (ATMs), or post offices. This is consistent with the general shortage in financial access in the Western Balkans. Insecure property rights limit the productive use of land endowments and limit the ability of Roma households to generate income. Substandard housing and insufficient access to services may be affecting human capital accumulation. Documentation The survey rounds show a high coverage of birth certificates and identity cards, but qualitative evidence indicates Roma often lack adequate documents—such as proof of residency—to gain access to social services. The coverage of identity cards is relatively narrow in Montenegro, where it may be affecting the ability of Roma to access social services and benefits. Lack of documentation, such as school certificates from abroad is also a problem among Roma returnees from the EU, leading to 8 Executive Summary BREAKING THE CYCLE OF ROMA EXCLUSION IN THE WESTERN BALKANS delays in the enrollment of children in the formal education system or of students in classes that do not reflect their prior schooling in the EU. Trends in Access to Services and Economic Opportunities and Inequality Overall, education and housing emerge as more well performing priority areas; average coverage rose in most countries between 2011 and 2017, though the evidence on progress in inequality is inconclusive.6 Nonetheless, across priority areas, education and labor markets exhibit the poorest indicators and the widest gaps. The changes in coverage and inequality in health were small, and the evidence on improvement or worsening is inconclusive, whereas access to economic opportunities through labor markets worsened in most countries, while the evidence on inequality in labor markets is generally inconclusive. The evidence on documentation is generally inconclusive or mixed, though the coverage among Roma is generally extensive. Policy Agenda That Roma should become healthy, productive members of society who are well integrated into the labor market is the ultimate objective. This implies an integrated, life-cycle approach that helps surmount the barriers many Roma have faced since early childhood. Given the factors that constrain Roma inclusion in the Western Balkans, policies and interventions that may generate improved outcomes among marginalized Roma are identified in this report. The first is the data agenda because accurate data on the programs implemented in the Western Balkans are scarce, and adequate policy monitoring and evaluation is not possible without proper data. Policy recommendations in the five priority areas and two cross-cutting themes, gender and discrimination, are then examined. The need for an integrated approach to breaking the cycle of Roma exclusion is subsequently described. Evidence-based policy making is fundamental if policy makers and practitioners are to enhance their policy decisions and the management of programs aimed at Roma inclusion; evidence-based evaluation is a key tool in the improvement of programs. Applying data, evidence, and evaluation to decision making involves the following. (1) Rigorous evidence should be built up on what works, including costs and benefits. Rigorous evaluation may encompass randomized control trials and quasi-experimental studies to measure program impacts as well as an analysis of the policy costs per outcome to enhance the allocation of public resources. The absence of rigorous evaluation does not necessarily mean that programs are ineffective, but it does mean one is uncertain about true program impacts. (2) Program delivery should be monitored, and impact evaluation should be used to measure program effectiveness. This is key to basic accountability to ensure that programs are operating as intended and to identify opportunities for improvement. If the collection of proper survey data is costly or impractical, administrative data can also be exploited to evaluate program outcomes. (3) Use rigorous evidence to improve programs, scale up what works, and redirect funds from consistently ineffective programs. (4) Encourage innovation and test new approaches. 6 The evidence is considered inconclusive if the changes are not sufficiently large to establish a statistically significant difference between the two survey years. Executive Summary 9 Current policy monitoring and evaluation are inadequate because of the lack of official ethnic- disaggregated data. To target policies at marginalized Roma communities more effectively and evaluate the impacts, governments in the Western Balkans need to improve the quality of official data to measure ethnic disparities reliably. The lack of the topic of ethnicity in official survey and administrative data hinders the ability to monitor Roma inclusion outcomes effectively. Many national action plans do not provide for concrete, measurable performance indicators, milestones, or data sources, nor do they identify the authorities responsible for monitoring and evaluation. With the goal of more effective monitoring and evaluation and the efficient implementation of antidiscrimination law, the EC has advocated for the collection of household data that are disaggregated by self-identified ethnic origin. However, this poses significant challenges. First, many Roma are reluctant to self-report their ethnicity. To address this problem, survey questionnaires should rely on dual ethnicity reporting, that is, allow respondents to identify with more than one ethnicity, while including questions on mother tongue. Grassroots campaigns, such as those undertaken in Montenegro and Serbia to encourage ethnicity self-reporting among Roma, can be beneficial by tending to reduce underreporting. Additional challenges include political and legal obstacles and the existence of attitudes toward the privacy implications of collecting ethnic-disaggregated data in household surveys. A pragmatic approach toward the short-term enhancement of official data might focus on improving administrative data, which is based on the voluntary provision of information, the privacy of which is protected. Such data may also represent a useful tool in monitoring and evaluation. Examples of administrative data that could cover ethnicity include social security, education, health, and public employment service data. Several other strategies might be considered to improve the reliability of ethnic data in the medium and long term. Data on ethnicity might be gathered across several data systems, oversampled, collected through targeted periodic surveys, or linked across multiple data sources using indirect methods. A comprehensive vulnerability diagnostic is critical to the identification of regional sources of vulnerability and the policy options available to address them. To support evidence-based policy making, the development of a geographic information system based on the merging of administrative data, census data, and poverty maps might provide a more accurate disaggregated picture of the sources of vulnerability among Roma. A potential second step would consist of working with selected Western Balkan governments to generate a geographic information system in which information would be automatically updated. Such a proposal could be realized in the context of supporting the development of social registries—a database of potential program beneficiaries of multiple social assistance programs—in the Western Balkans. Poverty maps offer insights into the potential targeting performance of geographic interventions aimed at Roma. The results of the research conducted for this report indicates that geographically targeted projects will reach most Roma in Serbia, but not in Albania. Combining the poverty maps of Albania and Serbia with census information on the distribution of Roma in each country showed that, in Serbia, Roma are concentrated in municipalities with high poverty rates. Targeting resources on areas in Serbia with high poverty rates might therefore also benefit Roma. In Albania, Roma are not concentrated in the poorest areas, which suggests that targeting the poorest municipalities would not 10 Executive Summary BREAKING THE CYCLE OF ROMA EXCLUSION IN THE WESTERN BALKANS necessarily reach Roma; thus, additional targeting methods must be carefully designed, piloted, and implemented in Albania. While poverty maps help identify the distribution of poverty across small subnational areas, they cannot be used at this time to identify pockets of marginalized Roma within discrete rural areas or towns and cities. If geographical targeting is used, there may be Roma communities within municipalities in which poverty rates are low that are missed (error of exclusion). Reaching such communities will require a different approach. An ethnic-informed policy agenda that guides practitioners, development partners, and policy makers seeking to advance Roma inclusion must support and evaluate interventions aimed at Roma, including Roma women, to scale up what works. Policy makers should focus on selected communities to design, pilot, and evaluate interventions based on international best practice that have the potential to be scaled up nationally and regionally. Based on a diagnostic of the critical constraints that Roma face as well as a review of successful interventions designed to tackle these constraints, the research conducted for this report has identified several policy measures to narrow ethnic gaps in each priority area. Policy Measures in Education The agenda on education and skills is perhaps the most important game changer among Roma in the long term. Early childhood development interventions can become the basis for cognitive development and long-term health and productivity. Improving quality, equity, and access in compulsory education is also important, as well as the implementation of targeted programs to increase Roma enrollment and learning in upper-secondary education, such as the conditional cash transfer (CCT) program in North Macedonia. Eight policy measures have been identified that can help narrow ethnic disparities in education in the Western Balkans. These include the following: (1) promote early inclusion of Roma children through affordable, high-quality preprimary education, (2) provide additional educational support for Roma children and introduce collaborative teaching techniques for diverse classrooms, (3) provide financial incentives to promote enrollment; (4) provide mentoring support and role models for students in transition to higher education, (5) change in mindsets and socioemotional skills to improve academic performance, (6) promote the use of Roma mediators at all levels of education, (7) ensure that schools attended by Roma and other vulnerable children receive adequate funding; and (8) address segregation and promote nondiscriminatory practices in schools. Policy Measures Aimed at Labor Markets Four main policy measures can help improve employment outcomes among the working-age Roma population. These are the following: (1) improve the skills and work experience of Roma job-seekers by introducing vocational educational opportunities in growth sectors, along with remedial and second- chance education or apprenticeship schemes; (2) adapt active labor market programs (ALMPs) and public employment services (PESs) to offer more efficient services to Roma by focusing on interventions Executive Summary 11 that help Roma workers access various labor markets, overcoming sectoral and spatial mismatches, and increasing the outreach of PESs toward Roma and the financial incentives for the hiring of Roma; (3) address discrimination and stereotyping against Roma on the job market by implementing broader affirmative action programs in employment, conducting information campaigns to increase employer awareness, providing information on working with vulnerable groups, and encouraging public institutions to become role models for nondiscriminatory practices; and (4) apply innovative schemes to tackle labor demand constraints and foster entrepreneurship, which may include entrepreneurship and social enterprise development programs and programs to foster access to microfinance schemes. Policy Measures in Health In health, three main policy measures can help improve health outcomes among the Roma population. These include the following: (1) continue to promote health knowledge and awareness through Roma health mediators; (2) support universal health coverage and endorse the financial coverage of medical bills among the poor and vulnerable to incentivize the use of formal health services, with an emphasis on preventive care; and (3) provide discrimination awareness training to medical providers to reduce and prevent discriminatory practices. Policy Measures in Housing and Access to Essential Services In housing and access to basic services, three main policy measures can help improve access to decent housing among Roma. These measures require a broadening of the range of tools available to the government, including projects and programs, and thus the implementation of a new model that goes beyond the provision of social housing and housing allowances. Such measures would include the following: (1) improve housing conditions among the least well off Roma living in slum areas; (2) help poor families move into better housing through holistic and participatory approaches; and (3) legally certify construction and occupancy; a good recent example is the law on the treatment of illegal construction in North Macedonia. Policy Measures on Documentation Two main policy measures can help improve the access to documentation among the Roma population. These are the following: (1) reduce the burden and the costs associated with civil registration, paying particular attention to internally displaced persons (IDPs) and returning Roma, among whom the share of individuals lacking proper documentation is larger; (2) raise awareness about the benefits of civil registration through information campaigns in both majority languages and the Romani language; this is especially critical in Montenegro because of the high incidence of IDPs and the relatively narrow coverage of identify cards. 12 Executive Summary BREAKING THE CYCLE OF ROMA EXCLUSION IN THE WESTERN BALKANS Policy Measures on Gender as a Cross-Cutting Theme Because Roma females face overlapping barriers and disadvantages, policy measures should focus on narrowing gender gaps among the Roma population. These measures would include the following: (1) provide additional financial incentives in favor of the school attendance of girls, such as CCTs that are more generous in the case of girls than boys; (2) focus on ALMPs for Roma women, for instance, by offering training and private sector employment programs that are attractive to them and do not lock them into nongrowth sectors; (3) reduce gender-biased social norms by focusing on self-help groups and targeted information and normative messaging campaigns aimed at influencing the decision process among Roma girls on continuing in education, delaying marriage, and entering the labor market; (4) overcome the constraints related to young women’s care responsibilities, including the provision of affordable, higher-quality childcare services and CCTs conditional on school attendance to delay marriage and childbearing; (5) promote social cohesion using the proven model of self-help groups. In addition, measures to increase the supply, use, and affordability of preprimary education, childcare, and eldercare would also have an impact on female labor force participation. Policy Measures on Discrimination as a Cross-Cutting Theme Discrimination, prejudice, and stereotyping require interventions across all sectors. Some of the measures proposed above address discrimination in specific sectors. Other measures that have been proposed by the World Bank in Romania can be piloted elsewhere in the Western Balkans. These measures include the following: (1) establishing assistance and service systems in local administrative units for victims of discrimination that can also help them navigate their legal options and (2) introducing a surveillance mechanism to detect and reduce stereotyping against Roma in the media. The Way Forward: Breaking the Cycle of Roma Exclusion through an Integrated Approach To break the cycle of Roma exclusion, a comprehensive and integrated approach that tackles barriers throughout the life cycle is needed. An integrated approach ensures that the numerous barriers that Roma face are addressed in a holistic manner through the provision of essential services, social benefits, and focused interventions. Such an approach requires a coordinated effort across multiple sectors, including governments and community members. To achieve a coordinated effort, service management and delivery must ensure the timely identification of vulnerable groups, rapid needs assessment, and the tailored provision of services. A case management approach to service delivery is also necessary in an integrated approach. Actively reaching out to Roma communities is essential. Vulnerability diagnostics are a necessary step in achieving tailored services. The identification of regional sources of vulnerability across education, health, labor markets, housing, and so on is key to the design of an optimal policy, program, and service mix for the integration of Roma and other vulnerable groups. An effort must be made to evaluate whether the current policy mix is well designed and if it is aligned with the vulnerability profile of the target population, that is, Roma and other vulnerable groups. The analysis of social program inventories can contribute to a more systematic approach to social protection. Such an analysis can help establish (1) whether program Executive Summary 13 design is aligned with the barriers and constraints faced by the target population and (2) whether any vulnerable groups are excluded, for example, because of stringent eligibility requirements or unfavorable geographical location. 14 Executive Summary Breaking the Cycle of Roma Exclusion in the Western Balkans | March 2019