"The Problem Is They See Us As A Whole" Explaining Gender and Ethnicity-based Inequalities Among the Roma in Serbia “The Problem Is They See Us As A Whole”: Explaining Gender and Ethnicity-based Inequalities Among the Roma in Serbia © 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 Umbrella Facility for Gender Equality. Its contents are the sole responsibility of the authors and do not necessarily reflect the views of the World Bank. 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 design: Budy Wirasmo 2 Acknowledgments This report was prepared by a multi-sectoral team from the Social Protection and Jobs Global Practice, Poverty and Equity Global Practice and the Social, Urban and Resilience Global Practice. The lead authors are Shruti Majumdar and Andrea Woodhouse. The report was designed as a qualitative complement to the quantitative evidence from the Regional Roma Survey 2017, and commissioned by Natalia Millán and Monica Robayo-Abril. The team is indebted to the Institute of Ethnography in Serbia, in particular to Ivan Dordevic and Srdan Radovic for the field analysis and background paper, and Ljijana Gavrilović, Dragana Radojičič, Miloš Rašić and Milesa Stefanović-Banović for invaluable data collection and fieldwork support. Within the World Bank, the team would like to thank Jamele Rigolini, Marco Hernandez, and Stephen Ndegwa for their advice, and to Miriam Muller and Maria Beatriz Orlando for their peer review comments. The team would also like to thank the following colleagues for comments in several stages of the report (in alphabetical order): Marijana Jasarevic, Sandor Karacsony, Valerie Morrica, Miriam Mueller, Ana Maria Munoz, María Beatriz Orlando, Hoda Osman, and Stavros Stavropoulou. This report was made possible due to the generous funding from Umbrella Facility for Gender Equality. It 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). 3 Contents Acknowledgments........................................................................................................................... 3 Acronyms ........................................................................................................................................ 5 Section 1: Introduction ................................................................................................................... 6 Section 2: Status of Education and Labor Market Outcomes among the Roma in Serbia ............. 9 Section 3: Methodology: Finding the Roma ................................................................................. 18 Section 4: Key Findings ................................................................................................................. 24 Section 5: Kamendin, Belgrade ..................................................................................................... 32 Section 6: Novi Bečej .................................................................................................................... 43 Section 7: Vranje ........................................................................................................................... 49 Section 8: Conclusion and Recommendations ............................................................................. 54 References .................................................................................................................................... 62 Annex 1: Sampling Strategy for Qualitative Study........................................................................ 68 Annex 2: In-depth Interview Discussion Guide - Roma Women and Men ................................... 70 4 Acronyms ALP Accelerated Learning Programs CDD community driven development CoE Council of Europe CPF Country Partnership Framework EC European Commission EPD Education for pupils with disabilities ECD early childhood development ERRC European Roma Rights Centre EU European Union FGD Focus Group Discussion IDP internally displaced persons ISCED International Standard Classification of Education MICS Multiple Indicator Cluster Survey DG-NEAR Directorate-General for Neighbourhood and Enlargement Negotiations OCHA United Nations Office for the Coordination of Humanitarian Affairs OECD Organization for Economic Co-operation and Development PPP Preparatory Preschool Program RAE Roma, Ashkali, and Egyptians REF Roma Education Fund RRS Regional Roma Survey SCD Systematic Country Diagnostic SHG self-help group SPC Serbian Orthodox Church UNDP United Nations Development Programme UNHCR United Nations High Commissioner for Refugees UNICEF United Nations Children's Fund 5 Section 1: Introduction The objective of this report is to analyze the gender norms of Roma and non-Roma men and women in marginalized neighborhoods of Serbia, and assess the impact of these norms on schooling, work, and household decision-making. Through in-depth interviews and focus group discussions conducted in three regions of the country—Belgrade, Novi Bečej, and Vranje—the report aims to provide sociologically nuanced evidence as a complement to the Regional Roma Survey 2017 1. More broadly, this report aims to generate and promote the use of evidence on living conditions and human development outcomes among marginalized Roma communities of the Western Balkans and their neighboring non-Roma counterparts and provide evidence-based recommendations for policy and programming. The primary universe for the Regional Roma Survey is: (i) all households in Roma settlements or areas of compact Roma population (referred to as “marginalized Roma” 2 ); and (ii) non-Roma communities living in close proximity to marginalized Roma. The results of the RRS 2017 show that marginalized Roma in Serbia, much like the rest of 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 up the socioeconomic ladder 3. Roma face multiple barriers and constraints that hinder their ability to accumulate human, physical, financial, and social capital. Lack of endowments leads to weak labor market engagement, and Roma show a persistent inability to generate income over the life cycle in the survey. Marginalized Roma have low levels of endowments; the returns to education among them are fairly low; and their ability to accumulate assets is constrained. Roma women are particularly disadvantaged and face overlapping barriers in education and household decision-making. Whereas the RRS 2017 underscores these overarching trends, this report aims to explore these trends further and provide explanations for some of the sustained and accumulating disadvantages of the Roma. This qualitative report makes three main contributions. First, the gaps between Roma and non- Roma in schools, work places, and household decision making revealed by the RRS 2017 are placed within a structural marginalization or social exclusion framework. Embedding our understanding of gender or social norms within this framework can expand our investigation 1. Robayo-Abril and Millan 2019 2 The marginalized Roma in the survey are defined as Roma living in areas where the share of Roma population is greater than the national share. The sampling frame for Roma settlements was based on information from the most recent population census available in the country, using the lowest administrative units with the equal or higher than national average proportion of Roma population on its total population. See more on the survey and qualitative sampling in Section III: Methodology 3 Robayo-Abril and Millan 2019 6 from individuals and groups to structures, and from tangible to invisible forms of marginalization. Second and most critically, we find a broad range of diversity within the Roma population in Serbia. Unfortunately, Roma are often homogenized in both academic and policy literature. As Roma themselves point out, most of their problems stem from the fact that the “world sees them as a whole,”; as a result, policies and solutions are designed for the Roma as an undifferentiated population. The findings of this report move beyond flattening histories and attributing norms and beliefs to a static and singular “identity” and allegedly common “culture.” We make room for varying degrees of “Roma-ness” and aim to shed light on the variation in gender norms, beliefs, and experiences within this group. Third, we find that social norms fluctuate and shift as the wider context shifts – that within Serbia, there is significant variation among the Roma living in three different contexts (urban, post-industrial and agricultural). Using this perspective, we illuminate the mechanisms of change with respect to gender and ethnicity that remain invisible in the RRS 2017 but result in the diverging impacts captured by it. Strategic Fit 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. To fill this knowledge gap and inform policy making, RRS 2017 was commissioned by the European Commission (EC) Directorate-General for Neighborhood and Enlargement Negotiations (DG NEAR). DG NEAR explored changes in core development outcomes among marginalized Roma and non-Roma who lived nearby. While the survey (Robayo-Abril and Millan 2019) gives us trends across the region, this qualitative report drills down and unpacks some of the findings and trends in Serbia specifically. This report also helps fill gaps in our understanding of Roma issues. These gaps were highlighted by the 2015 Serbia Systematic Country Diagnostic (SCD), which stated, for example, that “the agenda to enhance inclusion is not simple; continuing to experiment, evaluate and learn is necessary to find out what works best.” 4 The SCD notes that enhancing shared prosperity constitutes the main challenge for Serbia; the country has pockets of extreme poverty, in particular among the Roma population. Development partners have allocated resources to address Roma exclusion and lack of opportunity, but these have not yet had the desired impact. Government has so far not given this issue sufficient priority, in light of the numerous other challenges it faces. A more comprehensive approach is needed to address the main constraints preventing this community from building assets and creating opportunities. To this end, one of the 10 priority areas identified by the SCD is to focus on education and skills, particularly 4 World Bank (2015b): p. 10. 7 marginalized groups such as Roma and women. To address this priority, a far more nuanced understanding of gender gaps in outcomes is necessary. The strategic objective of the Country Partnership Framework (CPF) is to support the Serbian government’s goal of reducing barriers to labor market institutions. Under the overall job creation objective of the CPF, the priorities are (i) to revise the minimum social security contribution policy, and (ii) to reform the design of social benefits in order to bring into the labor markets vulnerable groups such as Roma and women, thereby enhancing shared prosperity. To address the latter priority the CPF suggests additional measures, like providing care facilities for children and elderly, which help women reconcile care duties with career goals. These priorities are complemented by the proposed diagnostic in this report, which aims to investigate other barriers to Roma women’s participation in the labor force. The diagnostic in this report is also in line with the regional strategic emphasis on Roma issues. The European Union (EU) Framework for National Roma Integration Strategies 5 (for nine countries) focuses on four priority areas for achieving Roma inclusion: employment, housing, access to education, and access to health care. A consensus on these priorities already existed among key actors working with Roma issues, as demonstrated by the work of the Second Decade for Roma Inclusion (2015–2025) 6 . The EU Framework also explicitly called upon EU member states and Enlargement countries to design national inclusion strategies that include monitoring of the impact of Roma integration actions, along with review mechanisms for the adaptation of strategies. Within Serbia, the EU and the government are both keen on implementing of the General Action Plans as well as the Roma Action Plans associated with the Second Decade of Roma Inclusion. Situating the findings within these broader frameworks, the paper is structured as follows. Section II reviews the literature and current knowledge on marginalized Roma and social exclusion in Serbia, and notes the gaps in the literature. Section III gives a description of the methodology of the study and the sample used for analysis. Section IV presents key findings from the study on gender norms in work, schooling, and household decision making. Section V, VI, and VII present the three case studies and a detailed overview of the location- and context-specific barriers for Roma versus non-Roma households demonstrating how these three locations – one urban, one post-industrial and one agricultural – offer important insights into the variation in constraints and forms of exclusion of the Roma. Section VIII draws on the findings to provide concrete recommendations for policy interventions and concludes. 5 EU COM (2018) 6 See link for more details: https://www.rcc.int/romaintegration2020/pages/4/roma-decade-and-the-eu 8 Section 2: Status of Education and Labor Market Outcomes among the Roma in Serbia According to Council of Europe (CoE) estimates, the Roma population in Serbia is approximately 600,000, or about 8.3 percent of the population. 7 This makes the Roma the second largest minority in Serbia, after Hungarians. However, an accurate count of the population remains elusive. In addition to the “autochthonous” groups, an estimated 50,000 Roma fled during and after the conflict in Kosovo to Serbia, only half of whom registered as internally displaced persons (IDPs). 8 However, it is not known if all of them remained in Serbia or left for destinations in Western Europe. Furthermore, thousands of Roma have been returned to Serbia from Western European countries in the last few years as failed asylum seekers; including Roma who were originally from Kosovo. One of the biggest barriers to arriving at an accurate estimate is chronic statelessness of the Roma population. According to UNHCR (2016), while most Roma, Ashkali, and Egyptians (RAE) have identification cards and birth registration, few have their residence registration and registration in the vital record of citizens. Obtaining the necessary paperwork for “documentary citizenship” 9 requires procedures that are often too costly and complex for Roma households, especially those illiterate or with insufficient command of the Serbian language. Often lacking the complete set of documentation 10 required to access certain public social services or benefits, such as social assistance, many Roma remain out of the reach of public poverty alleviation efforts. 11 Moreover, many Roma settlements are located on non-designated land and the inhabitants lack a formal residential address, without which residential registration becomes impossible. A new judicial procedure of registration in the registry of births, together with 7 CoE (2012); Robayo-Abril and Millan (2019). 600,000 is the mean estimate. CoE estimates range from a minimum of 400,000 to 800,000. 8 ERRC (2017). 9 Sadiq (2008). 10 Documents that are often missing for Roma families are as follows: birth certificate (for registration for school, obtaining citizenship card, and key to many other documents); ID card or Lična Karta (for proof of residency, access to services, proof of identity); marriage license (for proof of marriage, legal rights of married persons); work booklet (proof of work history and qualifications, obtaining new employment, registering at Employment Bureau as unemployed, claiming pension); residence registration; and registration in the vital record of citizens (Bodewig and Sethi 2005: 63; UNHCR 2016). 11 According to UNHCR (2016), 1 percent of RAE living in “Roma” settlements (of around 700 people) are not registered in birth registry books, 5.2 percent of them have no identity cards, 38 percent have no residence registration, and 50 percent are not registered in the vital records of citizens. The same research also revealed that 8 percent of children under the age of four in “Roma” settlements are not registered in birth registry books. Five percent of Roma children from informal settlements under the age of five are not registered in birth registry books. Thus, a considerable number of children are still not registered immediately after birth. This presents a circular/intergenerational problem: in order to obtain basic citizenship documents, one needs to provide evidence that one was born in Serbia; however, such proof is impossible if the parents were not registered in the first place. 9 amendments to the regulations governing residence, 12 have made it possible to obtain personal documents, and the overall share of persons without residence registration has dropped in the past five years (from 53 to 38 percent). Roma as a subgroup of the population are typically underrepresented in survey work, and accurate data on the situation and causes of poverty of Roma are extremely scarce. However, Serbia household surveys with their large booster samples of Roma (with questionnaires comparable to those used in the general population surveys) and a handful of anthropological studies on the Serbian Roma offer a unique opportunity to understand the conditions and causes of poverty within this group. 13 In the following sections, we draw on policy and academic literature to understand the current state of knowledge on normative enablers and barriers faced by marginalized Roma (in comparison to neighboring non-Roma) men and women across three broad outcome themes: access to livelihoods, education, and household decision making. Access to Livelihoods The consequences of transitioning to a market economy and of the global financial crisis have not been the same for women and men in Serbia. Jobs that women held during socialist times, such as jobs in state administration, the textile industry, education, and social welfare, were the first either to be lost or to become low-paying jobs as the country transitioned.14 Women who lost their jobs during economic reconstruction in Serbia often found themselves performing traditional roles at home, including domestic duties, such as cooking and caring for children and the elderly. Since the outbreak of the global financial crisis the fall in employment has been dramatic in general, but even more so for women and vulnerable groups. Since the onset of the 2008 financial crisis, and up until the recent present, 15 progress on poverty reduction and shared prosperity has been reversed, driven largely by losses in employment and labor income. 16 Between 2008 and 2009 the number of workers (ages 15–64) decreased by over 180,000 individuals, about 7 percent, of which over 65,000 were women.17 12 The 2012 Law on Permanent and Temporary Residence of Citizens allowed registration of permanent residence at the addresses of social welfare centers for those who have no registered permanent residence or could not register it because they live in informal settlements without adequate addresses (UNHCR 2016). 13 UNHCR (2016). 14 Kolin (2010) 15 Because of the global financial crisis, poverty peaked at 15.1 percent in 2010 but then dropped to 13.6 percent in 2016 due to the recent economic recovery and labor market trends. Relative poverty, or the fraction of population living below 60 percent of the median income, was estimated at 25.4 percent in 2015 (World Bank, forthcoming, “Serbia Social Protection Policy Note”). 16 World Bank (2015b). 17 Marjanovic (2015). 10 This represents a decline of the employment-to-population ratio from 53.7 percent to 50.4 percent, and for women from 45.4 percent to 43 percent. 18 According to the same survey, in the same period the number of unemployed grew by almost 57,000 individuals, of which slightly over 17,000 are women; this represented an increase in the unemployment rate from 14.4 percent to 17 percent (and from 16.7 to 18.6 percent for women). The remarkably lower increase in unemployment compared to the sharp decline in employment in the period 2008–09 suggests that many persons who lost their job shifted to inactivity rather than unemployment. 19 The economic crisis additionally worsened the situation of the bottom 40 percent of the income distribution (B40). 20 The B40 constitute the bulk of the jobless, where the lack of labor income is a major obstacle to escaping poverty. In Serbia, the B40 is mostly composed of two groups. The first group is the pre-transition generation, aged 40 or more, who either lost their jobs during the transition from socialism and never found a proper new job, or entered the labor market as the transition was beginning but never gained a firm foothold. The second group is people who have never been strongly attached to the labor market; minorities like the Roma and women generally are highly represented in this population. These B40 groups now live either on public and private transfers or subsistence farming and are largely excluded from labor markets, especially formal jobs. Women are affected by unemployment more severely than men. Serbia fairs poorly compared to the rest of the EU member states when it comes to female labor force participation. According to the Gender Equality Index, gaps are pronounced in participation, segregation, and quality of work. 21 Women are less frequently employed in full-time equivalent jobs and less frequently can 18 Marjanovic (2015). 19 According to the most recent World Bank SCD for Serbia, “Numerous factors account for Serbia’s sluggish labor market. On the demand side, plateaued or even negative economic growth and limited private job creation have reduced income-generating opportunities for all Serbians but especially for the bottom 40 percent of the income distribution (B40). On the supply side, while human development indicators are generally good, issues remain on the quality and on the types of skills the education and TVET systems produce. Ineffective labor institutions also make it difficult to match labor and jobs efficiently. Until a 2014 reform, labor regulations restricted part-time work. A number of other constraints—ranging from poor public employment services to a high tax wedge for low- wage workers and to work disincentives in social welfare—also constrain labor market performance” (World Bank 2015b: v). 20 Serbia Social Protection Policy Note (World Bank, forthcoming). 21 Gatti et al. (2016). 11 access flexible working hours. 22 Their working life is five years shorter than men’s. 23 The unemployment rate for women is significantly higher than for men and the employment-to- population ratio of women is considerably lower than men’s (40.1 percent and 54.9 percent, respectively, in 2013). 24 The current labor market integration of Roma is particularly poor. According to the RRS, the employment gap between Roma and neighboring non-Roma populations is approximately 22 percentage points25. Moreover, Roma with jobs earn much less than non-Roma. 26 A 2015 UN Women report on structural barriers to Roma labor force participation highlights four impediments. Two affect both men and women: (i) low levels of education 27 and vocational training, and (ii) discrimination by potential employers. When it comes to Roma women, two additional barriers are cited: (iii) patriarchal norms within the Roma society, and (iv) the difficulty in accessing information and services/support. In the following sections, we unpack some of the supply side barriers, in particular the education gap as well as gender norms within the Roma society. Education Overall, access to early childhood education and care is both low and extremely inequitable. Only 9 percent of Roma children aged 3–5 are enrolled in pre-primary enrollment, compared to 28 percent of neighboring non-Roma 28 . Increased access to preschool among these children 22 According to Reva (2012), there are no significant gender differences in the characteristics of workers employed in the informal sector. Most of them live in rural areas and work in agriculture (70 percent of men and 76 percent of women). While Serbian women are more likely to hold part-time jobs than men, the percentage of women engaged in part-time work in Serbia is significantly lower than in OECD and EU-15 countries. Furthermore, both men and women who work part time in Serbia cite inability to find full-time employment as a major reason for working fewer hours; but part-time employment in OECD countries is largely voluntary (OECD 2010). Regarding entrepreneurship, men are almost twice as likely to engage in entrepreneurial activities as women. In fact, men represent 72 percent of the people who identify themselves as self-employed. Furthermore, self-employed men are significantly more likely to hire workers than self-employed women. Men comprise over two-thirds of entrepreneurs with employees. 23 Reva (2012). 24 Marjanovic (2015). 25 Robayo-Abril and Millan 2019: 197, table 2 26 Marjanovic (2015). 27 In a survey of Roma girls in Serbia (CARE International 2011), the hostility of the school environment and teachers’ prejudices were ranked as the most common reasons why girls leave school early (69 percent of respondents); the teachers often ignore or fail to take measures to end bullying and harassment against Roma students by their non-Roma peers. This, in part, explains why Roma women’s educational attainment and outcomes are much lower than Roma men’s. In 2011, 80 percent of marginalized Roma women aged 15–24 responded that they could read or write, in comparison to 92 percent of their male counterparts (2011 RRS; https://cps.ceu.edu/article/2012-10-16/undpworld-bankec-regional-roma-survey-2011). 28 Robayo-Abril and Millan 2019: 193, table C2 12 (especially when on a full-time basis) would also promote increased opportunities for parents, particularly mothers, to participate more actively in the labor force. The gap between Roma and non-Roma households in school completion and performance is stark. The compulsory education completion rate (ages 18–21) among Roma is only 67 percent, compared to 96 percent for their non-Roma neighbors, and only a staggering 19 percent of Roma complete upper secondary education, compared to 93 percent among neighboring non-Roma 29. According to the 2017 RRS, the primary school completion rate for non-Roma children is 94.5 percent and the transition rate to secondary school is 96.5 percent; for Roma children the primary school completion rate is considerably lower (63 percent), and for secondary school transition lower still (55.5 percent). 30 Not only are school completion rates lower, but Roma children also fare worse in school performance compared to their non-Roma peers (Robayo-Abril and Millan 2019). The 2017 RRS also highlights significant socioeconomic and geographic disparities in learning outcomes, especially according to school location, size, and parental education. Thus, students in urban areas perform much better than students in rural areas, and students in the top 20 percent cohort perform the equivalent of two school years over students in the bottom 20 percent income cohort. 31 Roma children are also disproportionately represented in “special schools” (or schools for children with disabilities) excluding them from equal access to quality education. 32 The enactment of positive legislative measures, such as the Law on Foundations of the Educational System in 2009, led to some improvements, such as to a decrease of the proportion and overall number of Roma pupils in special education (or EPD schools i.e. Education of pupils with disabilities). Nevertheless, higher number of Roma pupils are still transferred from mainstream schools to EPD schools. General steps have been taken to reduce the use of EPD schools, but no particular measures have been taken to assist Roma pupils. 33 29 Robayo-Abril and Millan 2019: 193, table C2 30 2011 RRS (https://cps.ceu.edu/article/2012-10-16/undpworld-bankec-regional-roma-survey-2011). 31 2011 RRS. 32 Praxis (2011: 24). 33 ERRC (2017). As we will see later, there is a discrepancy between the qualitative data and the survey on this finding i.e. on the prevalence of being transferred from mainstream to EPD schools. The RRS 2017 finds that percentage of students (age 7 – 15) going to EPD schools is 2% only, while an overwhelming majority of the adults and youth in the qualitative data report having attended special schools. There could be several reasons for this discrepancy – the qualitative did not interview any minors, so the status of students currently enrolled in middle and high school was not captured by the qualitative. Secondly, the qualitative finds that it is not at the enrollment stage that Roma students claim facing difficulties, rather that once enrolled, language and other structural factors prevent them from being able to continue in mainstream schools, and this is hard to capture in the quantitative survey. Thirdly, the survey had two separate categories – attendance in majority Roma schools (which is significantly high for Roma at 10%), and attendance in special schools which are not mutually exclusive categories, and hence the data may be messy and hard to interpret. And finally, as the qualitative data revealed, several respondents under-reported when first asked if they attended special schools, but on further probing by the name 13 A major bottleneck is the lack of disaggregated data concerning Roma pupils and the lack of a systematic and uniform approach to recording data on national belonging. 34 The problem has been partly addressed by the adoption of the 2009 Law on Fundamentals of Education,35 which is seen as a major step forward. The law introduces the concept of inclusive education and demands schools to keep track of “pupils or children, their academic achievement, exams, educational process and employees.” However, monitoring of the law’s implementation is limited, which reduces the education system’s efficacy in preventing social exclusion and marginalization of ethnic minorities. Another important systemic measure that supported Roma school enrolment and completion was the introduction of one compulsory year of Preparatory Preschool Program (or PPP) for all children ages 5.5 to 6.5 years. PPP lasts four hours a day over nine months, and is free of charge when provided by a public kindergarten or school. A third important measure under the same 2009 law allows enrollment of Roma children without any personal documents and abolishes testing at school entry. 36 Despite these measures, there are serious structural barriers to access to education for Roma. Even when schooling is free, textbooks, clothes, and transport can be costly. Children in school are unable to earn as much for their families through work. Additional barriers, particularly to preschool education, include: (i) insufficient and 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 ECD, especially when they can take care of the child at home (for example, if they are not working) or through relatives, and/or their dissatisfaction with the quality, sensitivity (for example, to the specific needs of individual children), and practical aspects (for example, hours of operation) of the service provided; and (iv) other perceived 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). 37 of the school they confirmed attendance, implying the survey question may have been misunderstood and under- reported on. 34 GoRS (2010): p. 11, as cited in ERRC (2017) 35 Law on Fundamentals of Education (Official Gazette of RS, no. 55/13), as cited in ERRC (2017). 36 UNICEF (2015). 37 World Bank (2017). 14 Further barriers for girls in the Roma community include: increased risk of peer violence against them; violence by school personnel; and, most critically, gender norms that encourage early marriage and commitment to housework and family rather than women’s education.38 In the following section, we unpack some of the basic factors that influence the existence and reproduction of the practice of child marriage, and the related social and cultural norms that inhibit education in particular, and eventually labor force participation. Early Marriage According to the RRS, 29 percent of Roma women ages 20–49 in Serbia were married before turning 18 years old, whereas among the neighboring non-Roma this figure is only 9 percent 39. A recent study by UNICEF 40 on child marriage among the Roma in Serbia explores attitudes, opinions, and motives of male and female members of Roma communities of diverse ages. The study finds that the lack of motivation for further education, primarily among girls, makes marriage seem desirable because there is no alternative model. The gender role of a woman as a housewife, mother, and wife is imposed on girls early on, denying them the idea of different life choices. Education is perceived therefore as a generally desirable, but essentially unnecessary, resource. Integrated Roma populations offer examples of alternatives to child marriage and a wider range of life options. However, a discriminatory social environment and the current difficult economic situation in Serbia can pressure children, including those from integrated families, to marry early. Social and economic difficulties help explain the increase in the rate of child marriages in the past decade. The UNICEF (2017) study also provides variation to this narrative: in places where there are opportunities for employment, going to school is considered important. The impact of economic prospects on the length of formal education—and the preferable age for marriage— fully corresponds to the results of research in post-socialist Hungary. In the period after 1989, as a result of losses of jobs and job opportunities, the Roma population suddenly became impoverished, the education level decreased, and, at the same time, the number of underage marriages increased. 41 In other words, the UNICEF (2017) study points to the relevance of understanding norms in response to wider structural constraints. The study suggests that in many 38 CARE International (2011). 39 Robayo-Abril and Millan (2019) 40 UNICEF (2017). 41 Durst (2002) as cited in UNICEF (2017). 15 Roma communities, early marriage is a response to widespread unemployment, which is a demotivating factor to pursue further education or potential job opportunities. Intra-household Decision Making Gender norms within the household also affect women’s education and employment. The evidence on this however is relatively mixed. While some scholars argue that gender norms within Roma households are sticky and have remained traditional over time, others find evidence of a re-patriarchalization of gender roles within the domestic realm in Serbia overall as a result of increased de-secularization of society. This started in Serbia in the late 1980s and peaked during the Yugoslav wars in the 1990s. According to the UNICEF (2017) study on child marriage 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. It is taken for granted that, ideally, men work and support the woman and family, while women maintain the home and take care of the children. In this model, neither girls nor married women should be employed. The possible employment of a girl would mean that a girl is not taken care of by her family (her “honor” above all is not taken care of). A 2013 study by the World Bank of gender norms and agency in 20 countries found a similar pattern: “in the interviews in the Roma community in Serbia, the defense of highly unequal gender relations and the institutions that support these relations was very strong. The young Roma men did not appear to be open to questioning their privileges or to seeing any value in more equal relations.” 42 Another study suggests that what we see today is a more traditional ascription of gender roles than before. Drezgic (2010) highlights the peculiarity of the Serbian case in the Western Balkans in this regard: “From a highly secularized society, in which religion and religious institutions have been even more marginalized than in other parts of former Yugoslavia, over the course of only two decades Serbia has become a society with high rates of religious identification, while religion and the Serbian Orthodox Church (SPC), representing the major denomination, have gained a prominent place in public life” (Drezgic 2010: 956). 43 She argues, using examples of the anti- abortion campaign, increased presence of religious content in the media, and the incorporation of religious education in public schools, that the discourse gradually subverted the ideology of women’s emancipation and gender equality developed during state socialism, and led to the 42 World Bank (2013: 34). 43 According to the 2017 RRS, about 53 percent of Roma in Serbia report being Orthodox, and about 30 percent report being Muslim. In the other five regional countries, the Roma are majority Muslim (over 80 percent—99 percent in Kosovo) (World Bank staff estimates based on weighted UNDP-World Bank-EC (2017) RRS data.) 16 reinforcement of traditional patriarchal values. In a survey conducted in 2004 on a representative sample of 3,639 respondents, 60.5 percent showed moderate to high patriarchal value orientations. 44 These data represent significant changes compared to the results of similar research conducted more than 10 years earlier, where 60 percent of the surveyed households showed a transitional division of domestic labor, 10 percent a modern division, and only 30 percent a traditional division. In other words, the 2004 survey suggests a nationwide reinforcement of traditional patriarchal values within the household.45 Gaps in the Literature As is evident from the literature, the Roma population faces multiple interrelated disadvantages in accessing basic opportunities in education, health, living conditions, and labor market. There is a strong research focus on addressing some of the demand-side constraints, including strengthening anti-discrimination legislation, ensuring documentation, collecting accurate data, expanding school and ECD coverage, and incentivizing public institutions to educate and employ the community. 46 However, fewer studies and public action have focused on supply-side constraints. We still have limited understanding of how broader structural constraints affect decision making, normative beliefs, and choices at the community, household, and individual level, particularly for Roma women. Therefore, in this report we aim to assess the normative enablers and barriers faced by marginalized Roma compared to neighboring non-Roma men and women. We study these enablers and barriers across three broad outcome themes: education, access to livelihoods, and household decision-making. Overall, in all three thematic domains the main gaps in the literature reflect a lack of disaggregation of norms among the Roma and situating it in their wider socioeconomic context. More research is needed to further unpack the different types of personal normative beliefs, social normative beliefs, and aspirations held by the Roma and non- Roma population in Serbia, and how these shape outcomes in education, access to livelihoods, and household decision making. 44 See Pešić (2006: 295). Analysis of patriarchal orientation relied on a composite index made up of empirical statements: if in marriage one person is employed, it should be the man; most of the housework is naturally a woman’s job; women and men should be equal in marriage, but it is best if men have the final say; men should do public and women do private work. 45 Interestingly, Kagin (2013) finds that unlike the general population, religion for Roma has a positive effect. Using econometric methods and data from the UNDP-World Bank-EC (2011) RRS of 12 Eastern European countries, the study finds that increases in churchgoing of parents significantly increases education rates of Roma children and lowers child marriage rates. 46 World Bank (2010). 17 Another limitation of research on women’s issues within the Roma community is that it tends to treat ethnicity, gender, and class as internally homogenous, separate social categories. Thus, structural barriers may be presented in an additive manner according to social category, without “intersectionality.” 47 The policies that ensue from such analyses risk “reproducing gender stereotypes of Roma women being chiefly responsible for the reproduction of and care for their families and communities. Thus, anti-racist interventions could end up recreating gender inequalities. At the same time, gender equality policies could reproduce ethnic biases. The compound result is that Roma women do not benefit from and sometimes are even adversely affected by both Roma policies and gender equality ones.” 48 In this report, we explicitly adopt therefore an intersectional methodology. We thus take into account how gender, ethnicity, and class are enmeshed and constructed by each other in order to understand the specificities of the condition of Roma women. Section 3: Methodology: Finding the Roma Consider the following six individuals who self-identify as Roma and reside in marginalized Roma neighborhoods across Serbia: Muharem is an internally displaced person (or IDP) from Kosovo. Muharem identifies himself as Turkish Muslim and his family has lived in Kosovo for generations. He moved two decades ago to Belgrade when the war broke out, making him one of 23,000 Roma displaced from Kosovo since 1999.49 In Belgrade, he has lived undocumented in several collective centers around the city center and is now living in an EU-funded public housing project on the outskirts of the city. He speaks Albanian and some Serbian and is struggling to integrate into a housing project with a diverse population of Roma and non-Roma IDPs, as well as refugees and low-income residents from around the city—none of whom Muharem identifies with. Muharem used to work in the city’s garbage disposal company but has now retired. Sanela is a young returnee from Italy, currently living in the same public housing in Belgrade as Muharem. Her parents sought asylum in the 1990s in a small town in Italy and she was born and raised there. She speaks Italian and having recently been deported back to Serbia is struggling to integrate. Her school certificate is foreign and invalid, which 47 Several feminist scholars call this critique one of “missing intersectionality.” Although scholars and practitioners alike have emphasized the unique ways women experience race/ethnicity and gender, in the research the idea is widely cited but unevenly applied. See Kóczé and Raluca (2009), Oprea (2004), Schultz (2012), and Corradi (2017). 48 Kóczé and Raluca (2009). 49 Allen (2016) 18 prevents her from continuing school. She was increasingly becoming a burden on her family. At 17 years of age, she was married to another Italian-speaking Roma resident from the neighborhood; they now have a child. Dejan considers himself a “true Belgrader” for having been in Belgrade for generations. He resides in a slum (or nehigijenska naselja) along the railway tracks in the heart of the city. He finds himself treated as a second-class citizen by Serbs, but distinguishes himself from the IDPs and returnees by reasserting that he is a “true Belgrader” unlike the new arrivals. Dejan has held the same job in a public utility funeral service company for 18 years and feels disappointed by his inability to move upward. He believes strongly that his opportunities and experiences at work, school, community, and even public spaces are shaped almost entirely by his Roma identity. He feels he is denied the same opportunities as non-Roma. Ivan has been evicted five times in his life from one location in Belgrade to another, and is currently in a container settlement (kartonska naselja) in Makiš. He is on welfare and has been coping by finding odd jobs throughout his life—cleaning, garbage collection or bin-diving (na kantu in Serbian), repairing cars, and construction work. He occasionally commits theft to make ends meet. He lives in a transient neighborhood that the city of Belgrade authorities used to maintain but do not anymore. The population is constantly in flux with residents coming in and out of the settlement constantly. Moving outside of Belgrade, we encounter Olivera, an “old settler” in the hills of Vranje in southern Serbia. Identifying as Pentecostal Christian, Olivera claims that her family was one of the first settlers in the area since the 1700s and it is in fact the non-Roma who are the outsiders in her city; they arrived only once factories and jobs cropped up in the late 1960s. The textile factory close by used to employ her family, and her husband was a musician. But as the factories closed down and traditional businesses such as music and dance have dwindled, both Olivera and her husband lost their jobs. Olivera has recently found a Roma teaching assistant position in a local public school in order to support the family. In order to cope, several of her neighbors and members of her family have begun migrating to urban areas or abroad (referring to themselves as “Gastarbeiters”, literally meaning guest workers or migrant workers). Danijel resides in the industrial-agricultural region of central Banat in northern Serbia. Danijel identifies with the Romanian Roma and claims that his family has been in the same neighborhood for at least a century; they settled here under the Austro-Hungarian Empire. Danijel’s family has always worked on the fields of their non-Roma neighbors. 19 However, as the latter began diversifying occupations in the 1990s, Danijel’s family was quickly unemployed and had difficultly finding new jobs in the vicinity. Danijel then began petty trade in secondhand car parts and his wife began selling bread in the neighborhood in order to keep the family afloat. These are six individuals with vast differences in place of birth, “origin”, language, religion, region, housing, schooling, work experience, and relationship with the majority population.50 Most importantly they diverge greatly in terms of how they identify with the ethnic marker “Roma.” Some actively disassociate themselves from this identity, some have a strong sense of Roma pride. Some prefer assimilation, some are more sensitive to and supportive of “difference.” Some are “ethnicity-neutral” while others construct their life histories and narratives almost entirely through the “Roma versus ethnic Serb” lens. To generalize their life experience based on their ethnicity alone greatly risks reifying their identity as if they are a sharply bounded, internally homogenous group. To this end, we carefully selected marginalized Roma communities that represent a wide range of life experiences, with an explicit awareness of the thin line between imposed categorization and self-identification of the Roma. A growing body of literature is now critical of previous research on Roma (including World Bank research), which fails to treat their ethnic identity as contextual and fluid. According to Surdu (2016), “the recently reemerged Roma-related research imports assumptions, classifications, and narrations from the social sciences and contributes through sampling strategies, interpretation of data, and generalization to reify and pathologize Roma ethnicity. Roma are relegated by experts to several types of determinism: to a social category, to a frozen culture, and to a homogenous biologized entity.” Rather than flattening histories and attributing norms and beliefs to a static and singular “identity” and allegedly common “culture” of the Roma, we aim to shed light on the variation in gender norms, beliefs, and experiences. Data were collected across five Roma settlements through 15 focus groups and 30 in-depth interviews. The qualitative sample at the settlement level was a sub-set of the quantitative sample from the RRS 2017, which selected the settlement in two stages: At the first stage of sampling, a list of settlements from the most recent population census available was used, and 50 According to the RRS 2017 Roma show considerable diversity in religion (53 percent Orthodox, 30 percent Muslim, 4 percent other, 3 percent Catholic, 2 percent no religion, 1 percent Protestant, 6 percent do not answer). Roma speak diverse languages (62 percent mainly speak Romani at home, 35 percent speak Serbian, 2 percent speak Albanian, 1 percent speak Montenegrin). According to the RRS 2017, 31 percent of Roma live in rural areas. Serbia was the birthplace of 91 percent of Roma, and 4 percent were born in Kosovo (the rest were born in various countries including North Macedonia, Germany, and Italy). Two percent are recent returnees (in the last 12 months) (UNDP-World Bank-EC 2017). 20 the sampling frame for Roma settlements was drawn using the lowest administrative units with the equal or higher than national average proportion of Roma population on its total population i.e. the “marginalised Roma” settlements. The “non-Roma” refer to non-Roma population living in close vicinity to the marginalised Roma and are not representative of the total population in the country. In a second stage, external or outsider’s identification (local people, NGOs, and experts) was used to pinpoint areas where Roma households are located in a given municipality. The qualitative sample at the settlement level was a representative subset of the quantitative sample of settlements and was aimed to capture as much diversity as possible in terms of location, age, gender, place of birth, type of housing and language (for more on the sampling strategy see Annex 1). Belgrade (the capital city), Vranje (a post-industrial town) and Novi Becej (an agricultural areas) were selected first in order to provide three types of contexts. Within each location, the neighborhoods selected represent five different housing types and ethnic compositions: Kamendin is an ethnically mixed settlement with apartment buildings in north- west Belgrade; Rakovica-selo is an ethnically mixed semi-urban neighborhood on the outskirts of Belgrade; Makiš is a temporary container settlement hosting only Roma families; Novi Bečej is a small town in the Autonomous Province of Vojvodina, northern Serbia; and the neighborhood of Gornja Čaršija in Vranje—the subregional urban center in southern Serbia—is almost exclusively inhabited by Roma. At the household level, the type of home ownership and demographic characteristics were taken into consideration – some respondents are homeowners of private houses (Rakovica-selo and Novi Bečej), others live in apartment blocks built within a public housing program (Kamendin), and a third group live in temporary containers in a temporary settlement (Makiš). The social and cultural characteristics of respondents in this study in fact vary considerably as well. The languages spoken by the majority of our respondents are Romani, Serbian, Albanian, and Romanian (according to the 2017 RRS, of the Roma sample 62 percent speak mainly Romani at home and 35 percent speak mainly Serbian).51 Many respondents are bilingual (often trilingual). A significant number were born or raised in Italy and Germany as a result of shorter or longer migrations and are fluent in Italian or German. Some younger respondents are even more fluent in Italian and German than Serbian. The study respondents also belong to different religions; the majority profess various Christian denominations and Islam; the Christians are predominantly Eastern Orthodox and increasingly Protestant (with Evangelical Christianity becoming more noticeably present in Vranje) 52. This does not by any means exhaust the diversity of the Roma population in Serbia, but it does constitute a wide array of social and economic disparities, living conditions, education, cultural traditions, and geographic locations (alongside the gender and age stratification of the respondents). 51 World Bank staff estimates based on weighted data from the 2017 RRS (UNDP-World Bank-EC (2017). 52 Diverse as they are, however, religious identities do not seem to draw lines of distinction within or between the investigated Roma communities. 21 What does seem to make a significant difference among Roma groups and families is their place of origin. Drawing on their life experiences and group characteristics, the respondents can be divided into three groups: returnees from the EU, internally displaced people from Kosovo, and members of the well-established urban and suburban Roma families. Some respondents could be located at the cross section of these approximate groups, but the most common cases of emigration are among the “old families” (in Novi Bečej and Vranje). Almost all respondents who have temporarily resided (or were born) in the EU contemplate returning to the EU, and many of them have plans for at least temporary emigration. Internally displaced Roma (mostly from Kosovo) and emigrants to the EU on the other hand showcase new modes of mobility (willing or otherwise) in Roma communities. Widespread marginalization of Roma by the majority population is more pronounced toward Roma “newcomers” and further exacerbated by intra-Roma marginalization. In Kamendin and Vranje a significant discrepancy has been observed between the “original residents” and “newcomers,” who are for the most part IDPs from Kosovo. The localities also represent a wide range in terms of age of settlement. Makiš and Kamendin are regarded as complex, recently created Roma communities of various origins and backgrounds. The residents of Makiš are predominantly IDPs from Kosovo. The localities of Novi Bečej, Vranje, and Rakovica-selo are considered old (ranging from the 1700s to the early 1900s). A particularly divisive factor is “rootedness” in the specific location (that is, a multigenerational presence and habitation in the same neighborhood or town). These parameters of spatial and social location need to be considered in interpreting the results of the study and formulating potential interventions in those communities. Data were collected to assess aspirations, opportunities, and barriers faced by both Roma and non-Roma across three broad themes: work, education (from early childhood development to tertiary school), and household decision making. The data collection 53 was guided by a questionnaire that was first piloted in two locations in Belgrade and then modified. A community- based group called Bibija 54 recruited participants, and all data was collected in the respondents’ living quarters. In each location, a mix of three types of interaction were employed: (i) focus group discussions with men, women, and youth; (ii) key informant interviews (that is, interviews 53 Respondents were given shopping vouchers for a drugstore chain as incentives. The respondents were informed of their right to leave the discussions and interviews, and to decline to answer the questions they found inappropriate. Only informants that have consented to their interviews being used verbatim have been quoted. For the purpose of keeping their identity anonymous, their names have been changed. 54 Bibija or Roma Women’s Center is an NGO with strong grassroots presence in marginalized Roma communities across the country. They work on strengthening Romani women and families as well as the visibility of Roma women and girl’s rights, the right to education, free choice of partners, employment, and adequate health care. 22 with community organizers, leaders, and elders) to provide an overarching perspective of the status of the neighborhood and relationships between communities; and (iii) individual in-depth interviews in order to investigate individual Roma and non-Roma life histories and narratives. Each form of data shed light on a different aspect of Roma life. Through the key informant interviews with community leaders and elders, a general overview of the community and its history was drawn. Through the focus group discussions, we sought data on the community- and location-specific barriers and enablers to work and school. We grounded these narratives within local employment and education trends, norms, and changing trends from the previous and next generation. During the focus group discussions, we asked questions about norms through vignettes and were interested in drawing out both “oughtness” and “regularity” norms—the former defined as moral judgments about behaviors that are internalized, generating internal sanctions like guilt and shame, and the latter defined as behavioral regularities that generate social expectations, which if violated trigger external sanctioning.55 Finally, through the individual in-depth interviews, we sought to drill down at the individual level retrospectively into specific life decisions on work and schooling, and encouraged respondents to reflect on the guiding norms as well as the broader context. The interviews were semi-structured in order to allow themes to emerge inductively (see Annex 2 for sample guide for IDIs) The interview and FGD guides underwent three rounds of pilots in Belgrade in order to finetune the guides, pilot the locations where interviews could be held, and test the framing, order and sensitivity of questions. Interviews were conducted in Serbian, Romani, and Albanian and a few Roma returnee interviews were conducted in German and Italian. After data collection, they were transcribed in English, and coded in QSR NVivo (a qualitative data analysis software). During the coding, some themes were preselected to match the themes of the questions asked, but we also allowed themes to emerge from the data in an inductive manner. From the data analysis, three themes consistently emerged as relevant across outcomes: structural marginalization, nested stigmatization, and regularity norms in a constant state of flux. In the next section, some key findings across the three outcomes of interest are illustrated through these three themes. 55 Hechter and Opp (2001: xiii). 23 Section 4: Key Findings Conceptual Framing: Structural Marginalization The qualitative data suggest that all norms—whether governing decisions on work, schooling, or intra-household decision making—have to be understood within the broader context of structural marginalization of the Roma community. Even though the Roma represent an extremely diverse and complex ethnic and social group in Serbian society, what is evident among all respondents is feeling a sense of “otherness” in relation to the majority non-Roma population (“ethnic Serbs”). How they conceptualize their “Roma-ness,” or how they define being a Roma, varies significantly. Census results on ethnic affiliation regularly show figures that are considered to be disproportionally low compared to the actual number of Serbian Roma. This is only partly because some Roma were not counted in the last census (for the lack of documentation, or temporary relocation to other places in Serbia and abroad). The main cause of the discrepancy is that many Roma declare other ethnic affiliation in official censuses (mostly declaring themselves Serb). However, despite the diversity in self-ascription within the community, from the outside they are often homogenized and experience very similar forms of marginalization at multiple levels. As one of our key informants noted, “our biggest problem is that they see us as a whole … being Romani means carrying a heavy burden on your back. We have also always had problems with others we grew up with. Why? Because we are Roma. And we have to resign to the burden we carry just because of our skin color.” Roma male, 40-45 years old February 2, 2018, male focus group, Rakovica, Belgrade Over the last decade, a growing literature on structural marginalization shows broad agreement on ways to analyze marginalization across diverse contexts—including expanding analytical focus from people and individuals to structures and institutions, and from tangible events to invisible processes. As Powell (2017) highlights, marginalization comprises a set of forces that may generate disparities or depress life outcomes without any racist actors: it is “a web without a spider.” Being Roma in the Serbian context means to endure a range of tangible and invisible forms of marginalization by the majority population. Instances of extreme violence and overt discrimination are less common than subtle, almost invisible forms of “other-ing.” Symbolic violence is rampant—every respondent in the study attested to experiencing discrimination in some form or the other through their life course. Consider the following quote from the mother of a young boy in Novi Bečej: 24 “Well, he’s a mixed marriage child, a child from a different marriage, and these children are… well, he told me, “my dad is white, let him go to the parent meeting, you don’t go,” and you know how difficult it is. It’s very difficult for both him and me. Until he finally overcame it. Now he sometimes says, “please, mom, you should go to the parent meeting too, don’t let dad go…” but he still does not declare himself as Roma. And I know it’s not his fault but the fault of the surroundings – which is as it is. He says, “I am not Roma.” Twice it happened and I always had a fight with him about it, so to say, well, I wanted to kill him, “what do you mean you are not Roma, Roma is also human,” I kept repeating, wanting to teach him properly… However, when I saw what happened, saw that he wound up in the ER twice, I told him, “be whatever you like….” Roma female, 40-45 years old February 16, 2018, Novi Becej, Vojvodina Direct violence and harassment of this nature in schools, work places, communities, and neighborhoods came up in several interviews (few first-hand and mostly second-hand experiences). But what is more common is a more systematic and invisible form of marginalization described by respondents, including a gradual sorting of Roma children into special schools, having to overperform in order to excel, acceptance of lower salaries and odd jobs as a way of life, and internalizing harassment and unsafe public spaces as the norm. Invisible and internalized marginalization are the hardest kinds to detect—especially statistically. For example, the RRS, using a Blinder-Oaxaca decomposition, finds little evidence of overt discrimination in labor market outcomes; it attributes these differences to educational attainment, but is unable to probe why differences in educational attainment exist. 56 The qualitative work shows that all respondents—Roma or non-Roma—are faced with the same constraints in marginalized neighborhoods in terms of access and distance to jobs and schools and both seem to consider job prospects to be dwindling across the country. Nevertheless, there is a stark contrast between how Roma versus non-Roma respondents articulate their aspirations and life chances. The non-Roma have greater educational and career aspirations than Roma respondents, who by and large accept their life circumstances and the stigma they experience as fixed. The long-term consequences of this are a deep-rooted normalization of injustice, a decrease in collective capacity, and a systematic erosion of human capital over generations. This internalization of stigma has ramifications for all outcomes of interest as described below in the sections on work, schools, and household decision making. 56 Robayo-Abril and Millan (2019). 25 The same structural framework can be applied in order to understand women’s unequal access to resources. A structural perspective on gender 57 views societies as having a gender structure that has implications at the individual, interactional, and institutional levels. “Gendered structures of constraints” 58 may prevent women and girls from accessing resources and more critically in translating those resources into improvements in their livelihoods and schooling outcomes. It produces gendered selves, generates interactional expectations from men and women, and gives rise to institutional forms of organizing and controlling men and women. Norms within this context have been an acknowledged force in shaping ideational aspects and in guiding interactions and actions. Norms have been understood not only as internalized behavioral prescriptions that have in them implied notions of worth (“oughtness norms”) but also as being coercive, that is, injunctions that can be reinforced by external sanctioning (“regularity norms”). In the following three sections, we highlight key findings on gender norms as embedded within a context of structural marginalization and argue that gender and social norms often interlock with civic institutions, the institutions of the state, the market, and intra- household bargaining dynamics to shape and sometimes reinforce the gender inequities of power—and impact the choices and freedom of women and girls. 59 These structural inequalities can persist even in the face of women’s individual agency. Work Employment opportunities vary drastically from one location to another in Serbia. The densest networks of livelihood opportunities are found in Belgrade, while the bleakest prospects were in Novi Bečej. Despite location, the Roma respondents claimed that there are a rare few “respectable” jobs available to them. Belgrade, as the capital of Serbia and the most developed part of the country, offers the widest variety of job opportunities in both the formal and informal sectors. On the other hand, Vranje, which used to be an important industrial center under socialism with a well-developed textile industry enabling the employment of mostly women (including Roma women), is now seeing a decline of work opportunities. Industries in Vranje were mostly devastated during the process of the transition, and only a handful of old and some new factories still work and employ Roma women. Novi Bečej, a much less developed town compared to both Belgrade and Vranje, was an agricultural center under socialism. The decline of agri-based jobs has forced local Roma to either depend on social welfare or find different strategies of generating income, mostly through emigration to developed states in Western Europe or through a liminal “grey economy.” Formal employment is rare and mostly revolves around working in public utility companies (such as city sanitation units in Belgrade), custodial staff in enterprises 57 Risman (2004). 58 Kabeer (2016) 59 Muñoz Boudet et al. (2013). 26 (women mostly take these jobs), and a handful of industrial workers (mostly in Vranje). Informal jobs mostly comprise trading in open-air markets, “bin-diving” or collection of recyclables (informal recycling), cleaning jobs, working on construction sites, and so forth. This seems to be in stark contrast with their non-Roma neighbors, who despite the same constraints seem to be more optimistic of their chances of finding employment and are in more stable jobs than their Roma counterparts. Overall, the types of jobs that the Roma are in rarely produce human capital (knowledge, skills, and health). 60 As Roma themselves claim, this lack of what they consider respectable jobs has set in an erosion of human capital and a steady decline in their “capacity to aspire” 61 over the last two generations. How does structural marginalization work in the labor force? All Roma respondents attested to the fact that they must work harder than others in order to sustain their job, they have shorter contracts with less benefits when compared to non-Roma with often the same levels of qualifications, and they are often treated differently at work which leaves them demotivated. “The conditions are different, they treat us differently as soon as they hear our name. I applied for a job once and there was another woman who applied too. Both of us got the job, but she signed a six-month contract and I only a one-month. You get different treatment, and that’s why I quit—whenever there’s something missing, they accuse us first.” Roma woman, 40-45 years old February 11, 2018, women’s focus group, Kamendin, Belgrade Several respondents consider emigration for this reason because they are likely to face less discrimination: “people are kinder abroad, they don’t ask who you are.” 62 Moreover, factors that are typically constraining for Roma, are not for the non-Roma. For instance, several non-Roma respondents claimed that even though they lack documents (for example, for high school degrees), they are rarely asked for such documents during job applications. They also seem well aware lack of documents will be a constraint for their Roma counterparts for the same types of jobs. “I know that despite a lack of documents, I can get a job, but they can’t” 63. There also seems to be a difference in how Roma respondents find jobs versus non-Roma. Non-Roma rely on social networks, public information, and role models, whereas Roma did not mention any of these as resources. Finally, of all respondents, IDPs and Roma returnees seemed more at a disadvantage than Roma who distinguish themselves as “having always lived in Serbia.” 60 World Bank (2019) 61 Appadurai (2004). 62 Interview with Roma woman, 40-45 years old (February 11, 2018, women’s focus group, Kamendin, Belgrade) 63 interview with non-Roma Serbo-Croatian refugee, 50-60 years old (February 21, 2018, Kamendin, Belgrade) 27 Among the Roma, women’s unequal access to jobs also must be understood within the framework of structural marginalization. The very idea of work is gendered such that it further excludes Roma women compared to men. Domestic/unpaid labor is rarely seen as work by men and women alike and is considered the domain of the women of the household. Men rarely participate in domestic chores and the burden of housework falls mostly on women—whether employed or unemployed. Common refrains are “my wife does the typically feminine chores” and “why should I work if my husband is providing for me.” Women’s freedom of movement is much more constrained than men’s. 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. Probing deeper into what is considered “working” at all, it is typically something that men in the family are said to do. Women, even when making a living in the informal sector, are said to be coping or making ends meet. Women are often engaged in informal entrepreneurial activities (selling produce or baking and selling goods, or hairdressing) but seldom see themselves as “working.” Anthropological research on women’s work in the Balkans suggests the same perspective. Roma women (much more than men) are using innovative ways to make ends meet, but identify these as “coping strategies of survival” rather than “work,” let alone “decent work.” “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 old February 11, 2018, women’s focus group, Kamendin, Belgrade There is a strong desire among Roma women to work decent jobs, but they find themselves increasingly caught up in coercive webs of appropriate gendered behavior in public spaces as well as within the household. Education Structural marginalization in education is a process that creates unequal opportunities for Roma children throughout the lifecycle of school attendance. At kindergarten level (age 3 to 5), Roma parents claim that extreme poverty leads them to prioritize paying bills and rent at the expense of kindergarten for younger children, whether boy or girl. The remoteness of educational institutions and the cost of attendance are major barriers. Even if parents receive subsidies for their children’s kindergarten and pre-school, the cost of transportation may still 28 present a barrier. For the few Roma parents that did recognize the value of kindergarten and wanted to send their children, the law became a binding factor since both parents must be employed for the child to get admission. The gap between Roma and non-Roma becomes even more stark and really manifests itself at primary and lower secondary levels of school. Roma children who have missed going to kindergarten and pre-school are less prepared for primary and secondary school; and they are less likely to know the Serbian language because their parents speak Romani at home. Even when compared to their non-Roma peers who did not attend kindergarten, Roma children are less prepared because their non-Roma peers speak Serbian at home. Moreover, several Roma respondents reported having been sorted into special schools, unlike their non-Roma counterparts. Adults claim to have begun at mainstream schools but were transferred to primary and secondary special schools or three-year vocational schools. the formal reasons for transfer were inadequate knowledge of Serbian, frequent failing of school years, or, in some cases of returnee children, administrative mishaps. This is also the primary reason why occupations such as hairdressing for Roma girls or basic manual laborer for boys are frequently mentioned as ideal jobs, because those are the skills taught in the special and vocational schools they attend. An oft-repeated sentiment among all Roma respondents is a major disconnect between education and work. They strongly believe that there are diminishing returns on formal education beyond primary school, unlike their non-Roma counterparts. “If you know in advance that your kid has no chances to succeed in life, you have no grounds on which to send the kid to schools.… If you are disparaged all your life and reduced to some level that belittles you, that this is not for you. Believe me, I experience these things.” Roma male, 50-60 years old February 2, 2018, male focus group, Rakovica, Belgrade Several respondents strongly attested that if the best jobs for them are in cleaning companies or as bin-divers, then a high school degree is of little use to them, and getting through primary school is sufficient. “Here’s my husband with his 12 years of school, and it’s no use to him. He’s a cleaning man with a broom. As if he didn’t have any education at all. And I’m a seamstress.” Roma woman, 50-60 years old February 22, 2018, interview, Kamendin, Belgrade This is corroborated by the 2017 RRS data as well, which finds that Roma have significantly lower educational aspirations for their children in general than do non-Roma. The majority 29 (generally above 70 percent) of non-Roma respondents said that tertiary education is sufficient for a child; among Serbian Roma, this was the case among less than 50 percent of respondents64. Likewise, a significant share of Roma responded that lower-secondary is enough education for a child; this view was rare among non-Roma. The low educational aspirations among Roma seem consistent with the lower returns to education. Also, their narratives suggest that their expectations about the types of jobs that will be available for their children are similar to what they have themselves achieved. The only exception is Roma returnees, who seem to be more confident in their aspirations for their children. While both Roma boys and girls are dropping out of school early—girls earlier than boys—the mechanisms are starkly different; preserving sexuality is a prominent reason among girls, though this sometimes plays out in unexpected ways. For girls, the most commonly cited reasons for dropping out of school are early marriage, preserving their sexuality, and helping with household chores. For boys, the commonly cited reason is to join the labor force. This is also corroborated by the survey data, which finds that 12 percent of males ages 15–24 reported dropped out because of joining the labor market, but only 2 percent of females. Among female drop outs, 35 percent reported marriage/pregnancy as the primary reason. 65 Both decisions are strongly tied into norms related to masculine and feminine “natures,” conferring upon these pathways a sense of “naturalness” and “rightness.” The road to masculinity as young men described is paved outside the classroom, and young Roma boys deprioritize education and learning much more than girls. “I was thinking in a different way. Ever since I was a kid 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 Although girls did attest to enjoying school, some of them were pulled out early in order to fulfill their household duties or as a precautionary measure to preserve their virginity for marriage. In fact, a stark difference between girls and boys leaving school is that boys leave voluntarily, but girls are sometimes forced to leave. Hence, as we will see later, young mothers typically aspire to send their girl child to school for longer because the mothers were likely forced to quit, unlike boys their age—even if both dropped out of school at about the same time. Much like in the job market, IDPs and Roma returnees are the most disadvantaged groups in schools. These groups reported having educational problems such as disruption of the 64 Robayo-Abril and Millan 2019 65 World Bank staff estimates based on weighted 2017 RRS (UNDP-World Bank-EC 2017). 30 educational process both in Serbia and abroad and inability to get foreign certificates recognized. Only 3 respondents out of 94 have started higher education, and the share of those who have never gone to school is higher than other Roma. Most of the IDP/returnee respondents (both male and female) had either dropped out of or only managed to go till primary school. In all locations men’s employment was higher (both formal and informal) but the differences in educational levels between men and women are smaller, albeit in men’s favor. Household Decision Making Within households, Roma men and women typically spoke of a strictly gendered division of labor. Femininity and being a woman are associated with the domesticated space of the home, where women take on most household chores (cooking, cleaning, taking care of the elderly and children, fetching water and fuels where necessary, and so forth). The road to masculinity is paved outside the home in public spaces and the labor force. This division follows from the traditional model of Balkan patriarchy, 66 which poses the single greatest barrier to reducing the gender gap when it comes to making decisions and sharing responsibilities within the household. This is the case in all the surveyed communities, but particularly those of Novi Bečej and in families of IDPs, where the influence of the husband’s parents is still crucial to the decisions made in the household. Households still predominantly consist of members of the extended family living together under the same roof. Unlike household labor, decision making within the household shows a slightly more mixed pattern. All respondents agreed that men and women should consult each other on all household-related decisions. In fact, most Roma women control the purse in the family and have more say in financial decision making than other aspects. However, for most other decisions men claim to have the last word. Men also have a strong say on decisions around early marriage and preserving women’s virginity prior to marriage. In Novi Bečej in particular, if a bride is found not to be a virgin, kris (common-law court) is still summoned sometimes to determine if the bride’s family should return all the money invested into the wedding to the family of the bridegroom. Such traditional arrangements used to be present in the Balkans in the non-Roma community as 66 This model of patriarchy was predominant in the Balkans until the second half of the twentieth century, and remained vital in some areas until its last decades. It implies a set of formalized rules of inheritance, obedience of children, organized relations, rules against immorality, and rules of obedience of women. Among other things, these rules imply the universality of marriage, patrilinearity, patrilocality, high esteem for the elderly (especially ancestors), and marriage at a young age. In many parts of Serbia, these rules were present until the mid-twentieth century. The modernist ideology of socialism led to a huge change in traditional roles for women in Serbia. Women got the opportunity to be involved in the educational system, as well as in the labor market. These changes were encourage by the process of industrialization, which necessitated a growing workforce, including women. 31 well, but during the socialist period these practices went through a continual process of dissolution. The model, however, persisted more visibly in the Roma communities. Early marriage however must be situated in wider structural constraints. While some families said the reasons for doing so were “customary,” others saw it as a response to widespread unemployment. Poor job prospects discouraged further education or a search for job opportunities and marriage was an exit strategy to unemployment. This is also evidenced by the fact that in places where perceptions of job opportunities are more robust, going to school is considered important and the age of marriage is pushed. Similarly, the emphasis on preserving young girls’ virginity must be embedded in the wider context of public spaces being deemed unsafe by the Roma for themselves. Although regularity norms (that generate external sanctions when pushed) are resistant to change, oughtness norms (that generate shame or guilt when pushed) are in flux and are being challenged. For instance, some oughtness norms and aspirations around early marriage, time use, and education are shifting. Among respondents, several young Roma mothers were pulled out of school and pushed into marriage and household duties earlier than they would have liked. However, they aspire to send their children, particularly the girls, to school for longer and push the age of marriage, against the dictates of regularity norms. Importantly, the change of norms is not a uniform process, and there are intra-Roma differences in the pace of normative change. Roma returnees, for instance, recognize that they themselves were not best prepared for primary school because they didn’t go to kindergarten. Hence, they desire to send the next generation to kindergarten if they can afford to. The predominant attitude toward kindergarten, despite their circumstances, is positive and the positive effects of it on the children, such as learning the majority language and early socialization, are acknowledged. Similarly on time use, while Roma returnee men rarely partake in household work, they claim that the right thing to do is for men and women to share in the household labor. In this sense there is a widening gap between oughtness and regularity norms in Roma households. In the case studies of the following three sections, we consider these findings in more detail. We also unpack some of the location-specific normative barriers and enablers to work, schooling, and household decision making. Section 5: Kamendin, Belgrade 32 The neighborhood of Kamendin began to be developed from the mid-1990s with the gradual erection of apartment blocks (which continues to this day, the newest blocks being built as late as 2017). Located in the northwestern outskirts of Belgrade, Kamendin hosts a mix of private residential homes (mostly owned by non-Roma) and apartment buildings (mostly social housing, inhabited primarily by Roma, and a few non-Roma). Most of the latter were built under the auspices of the city’s social housing program (also called social dwellings or socijalni stanovi in Serbian). These apartments are targeted at refugees, internally displaced persons (IDPs), and low-income families residing in Belgrade. Hence, among respondents, a significant number of Roma families were housed in the Kamendin apartment buildings (many of them internally displaced from Kosovo since the end of the twentieth century). They were housed here usually after the demolishing of the temporary “cardboard settlements” (kartonska naselja) in the central Belgrade area 67 or relocation from inadequate housing/“slums” (nehigijenska naselja). The container settlement of Makiš was in some cases a temporary home before resettlement to Kamendin. Roma living in this neighborhood thus originate from various parts of Serbia and neighboring countries (very often from Kosovo), and stem from differing cultural and religious traditions. Some are not fluent speakers of the Serbian language, or don’t speak it at all. Many of them are also returnees from EU countries, in many cases deported back to Serbia after a period of asylum seeking or a prolonged illegal stay in host countries. Kamendin as a local community is currently in flux given that both Roma and non-Roma inhabitants were given apartments in the last several years. For the most part, local non- Roma/Serbs are refugees from Bosnia or Croatia or IDPs from Kosovo; Roma are mostly IDPs from Kosovo or readmitted returnees from several EU countries. Non-Roma IDPs and refugees previously lived in collective (refugee) centers or other inadequate housing in other parts of Belgrade, while Roma IDPs from Kosovo were often living in cardboard and container settlements or other illegal dwellings. While the study did not directly ask any questions about housing, several respondents expressed discontent regarding their living situation. They felt that being moved to social housing in Kamendin has resulted in a situation where strangers without a common past or future are being forced to maintain commonly-held spaces and build a community. As all respondents—Roma and non-Roma—reported, transfer to apartment blocks did not prove to be the most convenient solution since most respondents are unable to cover the costs of utilities and building maintenance with their income (expenditures they did not have in either collective centers or illegal settlements). Some respondents even recalled their previous living and housing situation as happier and more carefree, with a stronger sense of community and greater proximity to informal job networks. They found that housing problems were a major issue that should be discussed. They also expressed concern on rising tensions between and 67 One such settlement was Belvil, which was demolished because of construction investments. 33 within communities and petty crimes between Roma and non-Roma residents in the apartment blocks. “There really was a lot of discrimination, especially here. I mean, when we moved to this neighborhood, to Zemun Polje, the discrimination was really active; all of them Serbs, so to say, were gathered yelling about how they don’t want Roma people in their neighborhood.” Roma male, 40-45 years old, arrived from Makiš container settlements February 2, 2018, male focus group, Kamendin, Belgrade “Yes, I am happy with the new apartment, but we are not accustomed to the new neighborhood. There is some mess every night—people throwing stones or sand, but we survive. Though we still want to go away from here. To a house maybe, we don’t like living in an apartment—it’s too noisy. There are always fights, the police and ambulance coming, always someone we have to reprove.” Non-Roma / Bosnian refugee, 50-60 years old February 2, 2018, Kamendin, Belgrade “They gave more houses here to the people from Krajina and Bosnian Croats. Serbia is our country, I don’t know what they’re doing here.” Roma male / IDP from Kosovo, 40-45 years old February 2, 2018, male focus group, Kamendin, Belgrade “The way things are today—God forbid! When we settled in these apartments, do you know what happened here? They didn’t dare to go out, our folks. Whoever dared was beaten up. They didn’t touch me because they knew me, I was born here.” Roma male, 60-70 years old, oldest Kamendin resident February 2, 2018, Kamendin, Belgrade As we will see later, these rising tensions and the general perception of the neighborhood as being unsafe to the Roma has had spillover effects on their ability to send children to school and get to jobs safely. Education Regarding public services, there is a primary school and a kindergarten in the neighborhood, as well as a local public health center. More centrally located Belgrade quarters host a range of high schools, colleges, medical institutions, and other public services aimed at all Belgrade 34 residents. Kamendin is connected to central Belgrade and neighboring quarters of the Zemun borough by a bus line of the city’s public transportation company. While both Roma and non- Roma parents are faced with the same infrastructural constraints, Roma children were slightly less likely to go to school at multiple levels. Moreover, Roma girls face multiple overlapping disadvantages that further deter them from continuing school and entering the labor force much more than their male counterparts. Post-Transition Schooling Experience: “In the Tito era when I went to school, no one could tell you that you’re a Gypsy” The older generation in Kamendin contrasts their schooling experience strongly with that of the younger generation—claiming that there was little discrimination against the Roma in schools and more equality in the former Yugoslavia. This in line with the general perception among Roma that in the post-transition era, as availability of jobs has declined and the Roma are further marginalized and pushed into poverty, discrimination against them has increased. At kindergarten level (ages 3 to 5), Roma parents claim that extreme poverty led them to prioritize paying bills and rent at the cost of kindergarten for younger children. The remoteness of the institution from Kamendin and the cost of their children attending are major barriers. Even if parents receive subsidies for their children’s kindergarten and pre-school, the cost of transportation may still present a barrier 68 Moreover, the joint family or multigenerational model of family is prevalent in Kamendin, which means that a family member (typically the mother or mother-in-law) is always available for child care. Therefore, sending children to kindergarten seems expensive and unnecessary. In addition, few Roma parents are aware that there is a kindergarten near them, while non-Roma parents seem to be more aware of available opportunities. For the few Roma parents that did recognize the value of kindergarten and wanted to send their children, the law became a binding factor: both parents must be employed for the child to get admission. “We tried to enroll our kids in kindergarten but got turned down because we are unemployed and we have no income for kindergarten. People who work have bigger earnings. But we are unemployed and can look after our children on our own. We would like our children to get in touch and make friends with other children; we would like them to play with them and acquire good manners. Their upbringing is totally different if they are at home all the time. They are not that restless when they go to kindergarten.” 68 This is corroborated by the RRS, which shows that the two most cited reasons for not sending the children to kindergarten are costs first, and then distance (Robayo-Abril and Millan 2019). 35 Roma male, 20-25 years old February 21, 2018, Kamendin, Belgrade And finally, for the rare few Roma children that did go to kindergarten, the parents complained of discrimination by parents of non-Roma children. They all felt that staff or teachers in school were typically fair and equal in their treatment of the children, but they frequently faced harassment from non-Roma parents. “When I took my children to kindergarten I heard a girl’s mom saying she wants all Romani children to be thoroughly inspected for lice. And I saw her talk to the caretakers, and I was hoping that the caretaker was telling her that every Monday morning all children are thoroughly examined. And so what if that child had lice—no big deal, it happens all the time. But I hit the ceiling when the mother turned around and said, “I want you to inspect those children over there” and pointed towards me and two more Romani mothers with their children. So I thought to myself—is that how children should be raised—to think that Romani kids are dirty and have lice, and that they should not be friends with them. The nurse came up to us and inspected the Romani children. She felt embarrassed and apologized to us.” Roma female, 40-45 years old February 11, 2018, women’s focus group, Kamendin, Belgrade The gap between Roma and non-Roma in educational opportunity becomes even more stark and really manifests itself at primary school (age 7 to 14). Roma children who have missed going to kindergarten and pre-school (age 3 to 6) are less prepared for primary school and less likely to know the Serbian language because their parents speak Romani at home. Even when compared to their non-Roma peers who did not attend kindergarten, they are less prepared because their non-Roma peers speak Serbian at home. Researcher: Why did they go to a special school? Respondent: Well, because of the language in her case. She failed a grade just because she couldn’t speak Serbian. She talked to her teacher in Romani. Roma female, 30-35 years February 21, 2018 Kamendi, Belgrade This sorting of Roma children into special schools, 69 as we will see later, has a strong impact on their job prospects and aspirations. Regarding the few Roma children that do remain in mainstream schools, their parents complained of bullying and teasing by non-Roma children in 69 Even though the qualitative data shows that most of the adult Roma respondents attended special/vocational schools, the 2017 RRS finds that the percentage of Roma students currently attending special schools is rare (only 2 percent of students aged 7–15). 36 primary school. At the secondary school level, few Roma are in regular grammar schools. Most of them are in vocational special schools such as Petar Leković as a continuation of their education in a special primary school. These are vocational three-year courses after which most girls pursue hairdressing (one of the most popular courses in the school) and boys pursue basic manual labor. Among respondents, several Roma parents and youth felt that it is enough for children to complete compulsory school (through grade 8), as they need to participate in the labor market as early as possible in order to supplement the family income. Even when there is a will to educate children, poverty and the need for additional work lead to discontinuation of school after grade 8, as evidenced in this oft-repeated sentence during interviews with respondents: “We simply did not have the conditions.” In the event of health shocks in the family (such as sickness or death of parents), children are often forced to take responsibility for taking care of the rest of the family. In cases where Roma families decide to emigrate because of their difficult social and material status (most often asylum seekers), children are withdrawn from school as well. In the event of their return to Serbia, education is almost never resumed due to the administrative barriers of transferring their degrees from EU countries to Serbia. Finally, being moved to Kamendin has resulted in an additional barrier to schooling that several Roma parents mentioned: they felt that the community was too unsafe for their children to traverse alone in order to get to school. Researcher: Do you think it used to be better? Male 5: A hundred times better. Now you can’t let your child to go to school. You just can’t. Especially in our neighborhood in Zemun. Researcher: Really? Who? Male 5: Yes, they attack them, how can I put this… Without a reason. Male 5: How can I explain, without any offence, Serbian nationality. Researcher: So, Serbian kids attack Roma children? Male 5: That’s right, they can’t go to school alone, we sometimes let them, but most often we take them to school. I put them in a car and drive them to school. And I stay for 10– 15 minutes there until they are inside. Roma males, 40-45 years old, February 11, 2018, men’s focus group, Kamendin, Belgrade Gender Gaps in Secondary School: Men should have more schooling, not women. Not ever. Girls get spoiled 37 Gender gaps become most apparent at the secondary school level 70. Families that can afford to send both their boys and girls to school prefer not to send the girls, for fear girls will lose their virginity, which is seen as a disgrace for the family. The perception that young girls’ sexuality and virginity deserves protection prevents parents from sending them to school. This perception is more prevalent among IDPs from Kosovo. They often believe that society at school has a bad influence on sexual morality, and that prolonged school attendance might jeopardize a girl’s virginity before marriage. “Men should have more schooling, not women. Not ever. Girls get spoiled. There’s one here who was taken by husband when she was 13. When I heard about that… though mine own also married at 13 too. School spoils them, I tell you. I was in the same story, but I was 15 or 16 when I fooled around with my man. My youngest daughter doesn’t have a boyfriend yet, she wouldn’t sleep here if she did.” Roma female, 40-45 years old February 22, 2018, Makis, Belgrade Gender norms also prescribe childcare as the responsibility of young girls. Therefore, girls with younger siblings typically drop out of secondary school to help with household chores and childcare. Another norm that disincentivizes girl’s education is the premise being that they will marry young and therefore do not need further education. “I think boys should be slightly better educated than girls. Girls don’t need education that much because they get married and depend on their husbands. If she marries a very rich guy, she doesn’t need anything else. It’s harder for men, they have to provide for the family.” Roma male, 20-25 years old February 21, 2018, Kamendin, Belgrade Finally, the lack of schools close to Kamendin is seen as a threat to girls as they get older. Girls who travel long distances by themselves or with their mothers to get to school are perceived to be at risk, and are more likely to be pulled out of school. Diminishing Returns on Schooling: What use is schooling if I’m going to be a cleaning lady in the end? One of the primary reasons why the Roma respondents de-prioritize education is because they see little labor market value in advanced education. This perception is made stronger by a 70 As the RRS 2017 shows, the upper secondary completion rates for Roma males and females is significantly large (16%) and significantly larger than the non-Roma gender gap (Robayo-Arbil and Millan 2019) 38 general lack of availability of decent jobs. Most Roma respondents strongly believe that the types of jobs available to them do not require much education. “It’s all the same. Here’s my husband with his 12 years of school, and it’s no use to him. He’s a cleaning man with a broom. As if he didn’t have any education. I’m a seamstress, so what.” Roma woman, 50-60 years old February 22, 2018, interview, Kamendin, Belgrade This is not trivial because it indicates a strong departure from older respondents who claimed to have prioritized learning and going to school and did not see it as a waste of time, regardless of the benefits for seeking employment. This general sense of despair and refrain of the futility of education is perhaps the only binding factor among the Roma in Serbia. Despair combined with the lack of decent jobs has resulted in a systematic erosion of human capital. Labor Markets Waste Disposal as Primary Occupation: People say that Gypsies stink, and we work for the City Sanitation—if we stopped working for two days, this entire city would stink. In principle, residents of this neighborhood are exposed to a well-developed consumer and job market in Serbia, given the extreme concentration of capital and a developed tertiary sector in Belgrade. However, despite the proximity, Roma respondents believe that decent and well-paid jobs are rarely available to them. The majority of working-age Roma of Kamendin are often unofficially self-employed (as traders of second-hand merchandise or fresh produce in open-air markets and flea markets, or in the recycling business). Men also work a range of odd jobs (in junkyards, collecting and selling paper, iron, scrap metal, or bin-diving). Of formal jobs, only a handful of men reported working in the city’s garbage disposal company, while none of the women are engaged in the formal sector. Many Roma respondents claimed that they face discrimination in the local job market in several ways. They either face direct rejection or, once selected, by differential treatment, shorter contracts, and lower wages. The respondents felt that oftentimes their names and the color of their skin are used as markers of their ethnicity and used to discriminate against them. One Roma woman for instance describes below the precise ways in which she is treated differently as a direct result of her ethnicity: Interviewee: “The problem is my name and my surname (she has an Albanian name but Serbian surname). The conditions are different, they treat us differently as soon as they 39 hear our name. I applied for a job once and there was another woman who applied too, so both of us got the job, but she signed a six-month contract and I only a one-month. You get different treatment, and that’s why I quit—whenever there’s something missing, they accuse us first. Researcher 1: Because you are Romani? Interviewee: Precisely. Researcher 2: Is that situation something you encounter all the time? Interviewee: Yes, literally all the time. If you manage to get a job, you will have to work your fingers to the bone. Once I found a job in a bakery, but I had to work from 5 am, and didn’t finish until 11 pm, 12 am, 2–3 am. You have no benefits and you work for peanuts.” Roma woman, 40-45 years old February 11, 2018, women’s focus group, Kamendin, Belgrade Roma respondents claimed that they are “crowded in” in low-paying, odd jobs that others in the city are reluctant to do such as city sanitation. A male resident of Kamendin in a focus group said: “I haven't tried to find a regular job, but I have a lot of relatives who are employed. But their incomes are rather low. And not only are their salaries small but they are also frowned upon because they are Roma, because they are black, you know. The best job for us is at the City Sanitation Service. Only Roma and Gypsies work there, no one else. We endure 12 or 13 hours of that work, in this weather, for a laughable amount of money”. Roma male, 40-45 years old February 21, 2018, Kamendin, Belgrade Documentation: I know that despite a lack of documents, I can get any job I want, but they [the Roma] can’t For a long time, lack of documents was considered a major barrier for Roma seeking jobs. While most Roma residents claimed to have documents (such as the personal identification card ) in this neighborhood, it still is an issue for several non-Roma refugees. The few Roma respondents that still do not have documents or did not in the past consider this as having been a major constraint in their ability to secure jobs, whereas the non-Roma residents claim that documentation has never been a constraint for them in the past or present. As one of the Macedonian Roma men claimed, “There are only few companies which hire you to work unregistered, you just need an identity card, some other data, and that’s it. But for most other jobs I can’t apply, I don’t have elementary school and I am not registered.” 40 Roma female, under 25 years old February 22, 2018, Kamendin, Belgrade Contrasting this to an excerpt from a non-Roma neighbor who is gainfully employed in a client- facing job in an accessories store in a shopping mall, despite lacking documents; the fact that is aware of this discrepancy makes the contrast even more stark. Labor market outcomes for Roma IDPs who were unregistered for over a decade are much worse than those of local residents and “voluntary movers” with similar characteristics. Hence, while documentation is typically seen as a resolved issue in Serbia, the labor market costs of forced displacement, driven primarily by a lack of documents for an extended period of time, are long-lasting in nature. When it comes to discussions on ideal jobs or aspirations there was a stark Roma-non-Roma gap. This was true of Kamendin and the other two locations as well. While the Roma respondents were limited in their aspirations on the one hand by the knowledge of which jobs are possibly available, and by realistic possibilities on the other hand, the non-Roma had greater aspirations. Among the non-Roma, for example, office work is perceived as ideal (for both the respondents and their children), even though respondents lacked clear insight into what kinds of office work are available. What matters is that the person who works in an office has to be clean, does not have to invest great physical effort, and is well paid for that. Among the Roma however, a great number of the respondents perceive as ideal the jobs which they already do, except on a more regular basis (such as construction work and garbage disposal for men and cleaning and hair- dressing for women). Women in Kamendin are primarily on social welfare, and only those who do not have husbands work. Some progress is observable in the cases of younger female respondents without children. One woman is actively seeking a job and has been employed more than once for a sustained period of time in the music industry, while another is currently taking a course to support her future employment efforts. However, for the most part, young girls and women are expected to stay involved in house work especially after marriage. A certain politics of adaptation to the present circumstances can be noticed in the cases of Roma women who migrated from Kosovo. Before they moved to Belgrade, several of the female respondents wanted to qualify as seamstresses; but in Kamendin and Makiš, where they are now settled, this is not a prospective profession because there is no adequate market.71 In these new circumstances, they believe that the ideal job for a woman is that of a hairdresser. They can obtain training for this job more easily because the special schools they go to organize courses for hairdressers (for bakers too, but they 71 In areas of Southern Serbia that border Kosovo (Bujanovac), the tailoring profession is still highly developed. 41 are less favored, and the women who took these courses still can’t find jobs). Women can also find employment as hairdressers, either official (in a salon) or unofficial (working in their own apartment complex/their clients’ apartments). In this way they can earn money even if they are officially unemployed. Household Decision Making On household decision making, both Roma and non-Roma men and women often first claim that there is no gap, both asserting that there is regular participation of men in household duties and bringing up children. However, further detailed discussions revealed that men rarely participate in domestic chores and majority of housework falls on women—whether employed or unemployed. The most common childcare duty they take up is “playing with children.” Older male respondents were clearer in their stance on household work and saw domestic work as the woman’s responsibility. Several younger female respondents who work, or actively seek or plan to seek jobs, complained about being assigned family duties that prevent them from pursuing full-time employment. Roma and non-Roma women are equally tied up with family and the household because of the traditional provider-protector models. However, a particular set of local circumstances in Kamendin further aggravates the gender gap among the Roma and is more likely to lead to women dropping out of the labor force and young girls dropping out of school. A significant number of Roma children are sent to primary and secondary special schools that are located outside of Kamendin, which obliges many mothers to escort children to schools and drop out of the labor force. Shifting Norms: Best thing for our children is to not have the same life we had, the Gypsy life Norms around education, labor markets, and decision making within the household are constantly in flux and being challenged and pushed on the periphery. For instance, some norms and aspirations around early marriage, time use, and education are shifting. While young parents, especially Roma returnees, themselves have had little education due to their circumstances, they want to encourage their children to stay in school longer and push the age of marriage. Parents recognize that they themselves were not best prepared for primary school because they didn’t go to kindergarten, and hence desire to send the next generation if they can afford to. The predominant attitude towards kindergarten, despite their circumstances, is positive. Its positive effects on children, such as learning the majority language and early socialization, are acknowledged. “Even though I didn’t go to kindergarten, I would like for my children to go. We tend to spoil them at home. I do, grandpa does, Dad also does sometimes. Mom too. So, the child just doesn’t turn out right. Kids should learn about culture. They should spend time and 42 make friends with other children. Fit in. Learn to write. Learn the language. And get ready for school.” Roma female, 20-25 years old February 21, 2018, interview, Kamendin, Belgrade When it comes to time use, while household work continues to be the woman’s responsibility, men are increasingly more ashamed to admit that they are not involved in household work. This becomes even more evident when comparing narratives of the older generation of men to the younger ones. The former are more willing to assert that the woman’s place is in the household while the latter are less eager to admit that they do not help out in the household. Even norms around the age of marriage are shifting. Parents recognize that early marriage has detrimental effects on education and job seeking and claim that they would dissuade the next generation from marrying early. “The most important thing for them is to finish school and university and find a job, any job they want. They should be free to choose their career paths, and they shouldn’t marry young. I didn’t have a choice, but I would give it to my children.” Roma male, 40-45 years old February 21, 2018, interview, Kamendin, Belgrade It should also be noted that for some of the respondents it is the change of residence and, consequently, the urban cultural context, that has brought progress in the area of early education, primarily for girls. This particularly applies to immigrants from Kosovo. In that environment girls' education, including kindergarten, was not supported, but immigrants to Kamendin are more inclined to send their girls to school in the city. Some norms are more resistant to change—the stickier ones concerning participation in the labor force and decision making within the household. Most Roma parents want their children to start working soon like them, doing the same types of jobs that the parents do. The notable exception is Roma returnees, who seem to have greater aspirations for their children than themselves. According to respondents, having lived in countries where their possibilities are less bounded by their identity has allowed them to envision better lives for themselves and their children. Section 6: Novi Bečej Novi Bečej, with a population of 13,000, is the urban center of the eponymous municipality in Vojvodina (autonomous province of northern Serbia). It has a significant Roma population, 43 mostly concentrated in the town. 72 Novi Bečej is a small town in the industrial-agricultural region of central Banat. By all accounts, the town did not benefit from the social and economic structural changes taking place from the early 1990s (“postsocialist transition”). The existing combined industrial-agricultural sector was mostly dismantled in the transition. Job opportunities in the town are now very scarce for the entire working-age population, and especially for marginalized Roma. Most Roma with roots in the region are unemployed, and those who are not are mostly engaged in the informal sector, doing seasonal agricultural work for instance. A rare few are employed in the almost depleted industrial sector of the municipality (and its vicinity). The Roma community resides throughout town with a significant concentration in the Begej neighborhood. There is a significant influence of Romanian Roma traditions. The Roma of Novi Bečej are a largely traditional and close-knit community when compared to Belgrade Roma, since they have been rooted in this location for at least a century (since the Austrio-Hungarian empire). Most of them speak both Serbian and Romanian, and to a somewhat lesser degree also Roma language (respondents in the research were all bilingual in Serbian). Local public services (primary schools, preschools and kindergartens, health care centers) are still functioning well and are accessible to all, including Roma. Most high schools and hospitals are located in larger nearby towns (including Bečej, a 10-minute drive, and Zrenjanin, a 30-minute drive). Respondents described a dire economic situation. Both Roma and non-Roma pointed out that even if the discrimination against the Roma were somehow to instantly cease, there would still be no employment options as there are not any for most of the locals. Compared to Belgrade, economic prospects in Novi Bečej and many small towns throughout Serbia are particularly grim. This has given rise to a shadow economy that Roma men referred to obliquely in their interviews—they are engaged in petty trade, in some cases illegal. Some respondents hoped for some kind of assistance from researchers; they offered their services or products, or inquired about potential employment or help from the team members. The older generation of local Roma (men and women) worked in industrial facilities in town and its vicinity during socialism, which is not an option any longer. Unlike Belgrade, housing did not represent a problem for most informants, given that they were mostly homeowners, or lived in rented housing with their (extended) families. More importantly, unlike Kamendin the Roma here are not a transient population. They have been living in the same neighborhood for generations with their non-Roma neighbors, which presents a safer and more stable permanent housing environment when compared to Kamendin. 72 1,295 people declared Roma ethnicity in the 2011 census in the municipality; the estimated number of Roma is higher. 44 Education In the locality of Novi Bečej among the residents interviewed, there were no cases of kindergarten attendance. Extreme poverty in the local Roma community is cited as the reason for not considering this stage of school. Another significant barrier is that schools do admit children if both parents are unemployed. More than in Belgrade, Roma families in Novi Bečej are likely to be living in extended families. As a result, parents claim that there is always someone at home to take care of the child, particularly the mother- or father-in-law. Interestingly, Roma poverty is also cited as a reason to not keep children at home for too long. By the time they are of primary school age (7 years), most children are sent to school because women claim that they cannot afford to have their children at home. Often it is necessary for women to engage in informal and occasional work (often with their husbands). This also means that women do not join the informal labor force until all children in the family have been enrolled in some type of education (discussed in more detail in the labor markets section). Also, children’s right to access social benefits is linked to enrollment and attendance in primary school. All respondents said that at least a primary level of education is necessary to obtain any job, formal or informal. However, there remain several obstacles to attending primary school in Novi Bečej. While all boys and girls enroll, they do not all finish school, and the main reason for the termination of primary education is helping a parent or parents in informal and agricultural jobs. Moreover, several younger students reported dropping out of school to leave the country with their parents, whether seeking asylum or other forms of emigration. Upon returning to the country, there is the problem of continuing education according to the Serbian curriculum. Certain classes from abroad may not be recognized, and schools ask for documents that individuals cannot produce. The same problem occurs among the so-called returnees, that is, Roma who were forced under the readmission agreement to return to Serbia, sometimes after several decades, from Western Europe. Regardless of whether our interviewees were considering girls or boys among Roma in Novi Bečej, the belief prevails that school is essentially unnecessary. According to the Roma respondents, even if they finish school there is no work, and hence completing school is seen as a waste of time and resources. I had to drop out in the fifth grade. So that I wouldn’t be stolen while going to school. 45 Compared to Kamendin, the norms around girls’ education seemed to be more stringent in Novi Bečej. Romani girls in this region are particularly vulnerable to dropping out of primary education. Almost all girls enroll in primary school but a large number of them do not finish for four reasons in addition to the ones stated above. First, girls may be required to take over household obligations, most often the care of younger siblings. Second, girls may marry at a young age, either in the form of an arranged marriage or as a voluntary leaving of school and the parental home (eloping with a boy). Third, several parents and youth alike felt that a girl’s education need only go as far as acquiring basic skills such as reading and writing. It is considered far more important that a girl prepares at home instead for her “real” roles—as housewives, wives, and mothers. As was repeatedly mentioned by young men in interviews and focus groups: “well, they [girls] don’t need it [primary school]. Because they’re getting married. They get married and what is the use of schooling? The guy that marries them won’t let them work. I’m 100 percent sure.” Roma male, 40-45 years old February 14, 2018, interview, Novi Bečej A fourth reason for dropping out of school cited by residents is very prevalent in Novi Bečej: the notion that young Roma girls once they reach puberty are at risk of eloping or losing their virginity. This kind of body policing of Roma girls from a young age, even when the distance to school is short, forces girls out of school early, but has to be understood within the context of a greater perceived risk of harassment of Roma girls (than boys or non-Roma counterparts). Much like Kamendin however, gender norms and aspirations for the next generation are undoubtedly shifting. Several Roma mothers who reported having suffered from the lack of education and early marriage wanted to push their girl child’s age of marriage and years of education. “When it comes to my children, it is very much important that they finish school. I will do my best to see that happen. My daughter is eight. I keep telling her that she must finish school and be independent. Not expect from her husband to provide this or that. She needs to be her own woman. I will do my best” Roma male, under 25 years old February 14, 2018, interview, Novi Bečej Labor Markets What other options could there be with eight grades elementary only? Apart from going abroad, cleaning or working illegally? 46 Devastated by the transition, the local population in Novi Bečej is adapting and coping with these new circumstances in innovative ways. Elderly respondents (both men and women) claimed to have worked factories that formerly existed in the town and its vicinity (some also met the criteria for pension), but this option is unavailable nowadays, except in the factories outside the town (in Bečej, Zrenjanin, and Novi Sad) or the cold storage plant in Novi Bečej itself. Several Roma respondents have resorted to car repairing and selling second-hand clothes. For the Roma, the “gray economy” 73 is considered a viable and profitable alternative when faced with limited formal jobs. The majority of male respondents deal in various forms of sale and resale of illegal goods. A smaller section of the population (both men and women) also perform seasonal agricultural work, and some women also clean houses. It is a frequent case in Novi Bečej that wives and daughters help the male family members perform their unofficial or agricultural jobs and earning money, but women’s independent work outside the house happens only sporadically. Apart from the priorities set by the traditional model (that children and the household are predominantly women’s responsibility), rigid gender norms around work are a barrier to women’s inclusion in the labor market. A woman performing an activity independently outside her house could be taken as a sign that the husband is not capable of earning enough to support the family himself. In addition, it was also mentioned that the frequent visibility and movement of the woman in public (for example, as part of the workforce) can be interpreted as wantonness and therefore discouraged. Such beliefs were more frequently expressed in Novi Bečej than in other locations. All the respondents claim that it is nowadays virtually impossible to find a job for both Roma and non-Roma. Local responses to lack of prospects still differ. The majority Novi Bečej respondents see emigration into one of the EU countries as the only possible prospect for a good job. A lot of respondents expressed a certain feeling of anticipation and temporariness regarding their current state of (un)employment. Most of them have already lived as immigrants for longer or shorter periods of time (most often in Germany, Austria, and Italy, frequently as asylum seekers), and most of them plan to move back to one of the EU countries. Many respondents also worked, often illegally, during their stay in EU countries. They have pleasant memories of these experiences, and by comparison life in Serbia is always worse. In a way, the “ideal job” for many of the respondents who are also returnees is de facto (another) emigration, that is, any kind of reasonably paid work in the countries of the EU. 73 Most of our informants, asked for the type of job they do for living, mention "trading stuff" as their primary occupation. Many of them were reluctant to talk publicly about the exact goods they are trading with. Since it is very likely the most of their trading activities are in some way illegal, it is not surprising that the interviews don't include detailed information on the topic. 47 Non-Roma respondents seem to be better positioned in the job market than their Roma neighbors. A wider range of jobs is apparently available to non-Roma than Roma of the same educational level and similar professional skills. Unlike their Roma neighbors, non-Roma respondents did not mention any barriers in the job market because of their ethnicity or social status (apart from a general lack of job opportunities). This also translates to higher job aspirations: non-Roma respondents regularly stated higher education managerial office jobs as an ideal career prospect for their offspring or in general. Household Decision Making When people get a daughter-in-law, they all like them to be literally children, to be virgins; people like children in their homes The traditional model of Balkan patriarchy, the influence of which can clearly be seen in all the surveyed communities, also precisely determines the position of women within these communities. This is reflected in the division of the household chores, household economics, as well as in the customarily required virginity of girls prior to marriage, which in turn results in the high percentage of early marriages in the surveyed Roma communities. In Novi Bečej, if a bride is found not to be a virgin, kris (common-law court) is still summoned sometimes to determine that the bride’s family should return all the money invested into the wedding to the family of the bridegroom. Such a traditional model used to be present in the Balkans in the majority community as well, but during the socialist period it went through a continual process of dissolution. The model, however, persisted more visibly in the Roma communities, although the processes of dissolution and transformation can be perceived in this case too. I always ask her opinion, but I have it my way. Traditional family relationships still pose the greatest obstacle to reducing the gender gap in decision making and sharing responsibilities within the household, particularly in Roma households. This is the case in all the surveyed communities, and especially those of Novi Bečej. In Novi Bečej, the influence of the husband’s parents is still crucial to the decisions made in the household, which still predominantly consists of extended family members living together under the same roof. In nuclear families, the husband makes all household-related decisions. The norm may be shifting slightly in that men are more likely to ask their wives’ opinion, but men claim to have ultimate decision-making power. The chores are still women’s responsibility exclusively, and the majority of male respondents from Novi Bečej were not even able to comprehend the possibility of an alternative in which they would undertake so-called “women’s jobs.” In the household, the man most frequently engages in the outdoor work, in the courtyard, or in 48 physically more demanding tasks, while inside the house he completes chores in extraordinary circumstances only (for example, when his wife is ill). Section 7: Vranje Vranje is a city in the southernmost part of Serbia, and also a subregional center (capital of the Pčinja District, Pčinjski okrug). It is home to 60,000 inhabitants (2011 census), 84,000 if we include semi-agricultural suburbs within the administrative city limits. Vranje is home to a large Roma community (4,645 people declared Roma ethnicity in the 2011 census; the estimated number of Roma is higher), which is spread throughout the city in various settlements, including the historic urban nucleus. Roma settlements can be found throughout the city of Vranje, but by far the most prominent is the neighborhood of Gornja čaršija, which is the most ancient part of the city and Vranje’s most populous Roma quarter. It is also the oldest continuously inhabited Roma quarter in all of Serbia, and local Roma take pride in being autochthonous urban dwellers, rooted in this location for centuries. All the respondents in the research were inhabitants of Gornja čaršija, and they traced their origins solely to the city (recalling that their great-grandparents were also local). Most Roma in Vranje are relatively integrated in the wider urban community (all fluent in Serbian). More transparent marginalization (including intra-Roma marginalization) is aimed at Roma settlers/“newcomers” (mostly from surrounding villages and Kosovo), who usually inhabit the fringe urban settlements (particularly Ciganski rit). Vranje is located next to the major highway connecting Serbia, North Macedonia, and Greece, and is not facing rapid depopulation characteristic of most midsized towns in the country. In the neighborhood where fieldwork was conducted most houses are one or two-story buildings constructed throughout the twentieth century (all equipped with electricity, water supply, and sewage system). Gornja čaršija has no school or childcare facility, so children attend kindergarten and school in the nearby urban districts together with non-Roma children (the primary school is located closest to Gornja čaršija). Most of the high schools are located in the city center, and Vranje also has a general hospital and a primary health care center. The Roma of Gornja čaršija represent an old, closely knit community, where socializing and marriage take place almost exclusively within the community, but they are not completely isolated from other Roma communities of the city or from non-Roma citizens. Roma in the neighborhood (and in Vranje in general) are overwhelmingly Eastern Orthodox (with a Muslim minority, mostly settlers from Kosovo), with a noticeably rising presence of different forms of Evangelical Christianity (as is the case in other Roma communities in Southern Serbia). The introduction of new Christian denominations in the community is perceived benevolently, since Christianity as a whole is 49 experienced as a “Roma religion” (romska vera). Jehovah’s Witnesses are the exception and are shunned by the rest of the Roma. Education In my opinion, here school is worth nothing In Vranje, unlike Kamendin and Novi Bečej, both parents often are employed, and it is common for them to work different shifts so that somebody can stay with the children. Childcare is distributed among all adult extended family members. Only in very rare cases of young couples separating from the husband’s family was childcare the parents’ responsibility since there was no extended family in the young couple’s home to help. While secondary education was not frequent among respondents in Novi Bečej and Kamendin, in Vranje there were more respondents who attended at least this level of education. The general attitude of the respondents is that the likelihood of finding a job is so low (not only among Roma) that even that level of education is no guarantee of finding a job. As some of our respondents claimed: “even when they finish school, there are no jobs out there. I don’t know about college, but some finished high school. And still, they sit at home. They don’t get work”. “he had finished the right school, and they wouldn`t give him a job although they needed workers. In my opinion, here school is worth nothing.” Roma males, 20-25 years old February 14, 2018, youth focus group, Vranje If Roma in Vranje finish secondary education and have qualifications for a particular job, according to one of the respondents, they will face a “glass ceiling.” Also, access to schools that are viewed as promising or high-quality is very low. If they even decide to go to high school, young Roma go almost exclusively to vocational three-year schools or attend courses in special secondary schools. This is reflected in the most common occupations for young Roma— hairdresser for girls and basic manual laborer for boys. Finding a job of any kind is very difficult, and finding a job that corresponds to one’s qualifications is virtually impossible. People with a high school degree struggle even to get jobs as unskilled workers (both men and women). Respondents say one needs to “have connections” or “know people” in order to find employment locally, or even give a considerable sum of money to someone, which is not an option for most of them. The respondents also complained about the discrimination persistent discrimination in the local branches of the National Employment Service. 50 “At the employment office they always seem to have some special measures for Roma people, and in their meetings they usually say that Roma people refuse to work. But if they show some interest, they are not given the chance. I've been there many times to ask to be given training in English or German … “we will let you know,” they say every time. They have never called me, and as that guy this morning said … he was asked if he wanted to be sent to training to be a caregiver instead” Roma female, 30-35 years old February 24, 2018, interview, Vranje Labor Markets Although the economic situation in the city is not as grim as was the case in Novi Bečej, Vranje did experience a decade-long decline from a developed industrial city to a poorer mixed economy (services and industry), with only three industrial facilities still operating. Among respondents, Roma men are sometimes employed in the radiator factory or the insolvent tobacco factory. Several dozen Roma women work in a factory of a famous multi-national company, which outsources a number of semi-skilled poorly paid jobs to officially unemployed women (including Roma). While only a handful of Roma are employed in industries, the remaining earn income trading in open-air markets and seasonal agriculture (usually further away from Vranje and its vicinity). A lot of Vranje Roma are also emigrants to EU countries, and a significant number of local families make ends meet through foreign remittances—the latter being quite uncommon in the other two field sites. The majority of Roma people from Vranje in fact have informal jobs only, although their parents’ generation was until the period of transition employed in industry, which in this town was very well developed. Men are mostly musicians, and they play at various celebrations all over the former Yugoslavia, as well as in Western Europe (depending on how famous and successful they are). Roma from this part of Serbia are famed for their musical skills, and although this does not represent steady earning, income from music surpasses other sources of income for Roma in Southern Serbia. They declare being unemployed, claiming that music business only earns them their “bread and butter.” But Roma who deal with music professionally and successfully do not actually want to find employment locally since their income, however irregular, still is considerably higher than that of the employed Roma. Women usually work informally as cleaning ladies in the Serb, non-Roma houses in downtown Vranje. It is a job that is carried from mother to daughter or some other young relative once the mother can no longer perform hard physical labor; it is thus a job acquired on recommendation, 51 and the relationship between the client and the worker becomes permanent/long-standing. The job is still less paid than factory work, so all these women would like to find other employment. A large number of women also perform commissioned work for the local shoe manufacturing factory, such as manual shoe sewing. This is a hard and extremely poorly paid job, which is occasionally done by entire families. There are also families who sustain themselves by occasional work for wages in agriculture across Serbia (there are fewer of them in the Gornja čaršija neighborhood, and more in other settlements). Agricultural work is done only by those who have no other option to earn money, because it is mostly hard labor done in extremely bad conditions. It is noteworthy that a sort of “socialist heritage” still endures in Vranje, where women are viewed (both by themselves and by others) as regular members of the work force, and are not seen in the traditional role of housewives being in charge only for household duties. The pre-transition era is fondly recollected by many of the respondents, who allege that it is presently impossible to acquire official employment unless one is “well connected” (the respondents employed in industry or the public sector started working in the 1980s or early 1990s). The Vranje narrative of declining employment opportunities and living conditions converges with similar experiences from other Roma communities that were once well integrated into the labor force and working-class lifestyle of the industrial towns of socialist Serbia (and the former Yugoslavia). Household Decision Making In Vranje, most Roma and non-Roma respondents believe that girls and boys should have the same level of education. Some Roma women thought that girls need more education than boys so that they can get jobs and be relatively independent in the future husband’s family. “If I had my way, I would have educated the girls more. Because, she would go to another home, her husband’s home. And this way she could have her own salary, and depend only on herself, and she wouldn’t suffer as much [as a daughter-in-law in her husband’s household]” (woman < 25). This attitude is correlated with a higher degree of integration of the local Roma in this city, where completion of primary education (up to ISCED 2) is common in the surrounding community. This also stems from the fact that in Vranje during the socialist period, the employment rate among Roma and especially Roma women was relatively high, given the large number of factories that traditionally employed women as textile workers, for example. Having a perspective of finding a job in factories, members of Roma community used to consider education as a valuable and meaningful asset. Even though there are far fewer job opportunities than during socialism, the importance of education is still recognized among our informants, a consequence determined by the modernizing impetus from the socialist era. 52 In Vranje, again unlike the other two cities, Roma women's participation in the labor force during the socialist period inevitably changed their role in the household. In addition to their household duties, Roma women are equally active in the public sphere. There are men who share the household chores and child care with their wives, especially if the wives are employed. A certain number of male respondents readily answered that the division of the household chores is based on equality. A more detailed look at the statements of husbands who “help” their wives complete household chores suggests that this occurs when the woman is absent from the house (for example, to go to the doctor or town council); and child care is reduced to men’s playing with them. Of special interest is women’s phrasing that “my husband will sometimes look after the children for ME,” which implies that children are the occupation and task pertinent to mothers. Nevertheless, a transformation of the ideational model can be observed in that men nowadays perceive equal distribution of the household chores as socially desirable, although few of them actually implement it in practice. This change at the level of the desirable ideal, albeit only declarative, suggests that a tendency exists toward a transformation of the traditional model, within which the man works “outdoors,” while women’s responsibilities are tied exclusively to “the house.” Major decisions in the family are claimed to be made by the mutual agreement of husband and wife. Besides husband and wife, decisions are also made by all the other adults in the cases of extended families. Some families still preserve the idea that father/husband “has the final say,” but reaching decisions on these grounds is not always feasible in reality, especially when it comes to the lives of the young people. It is an increasingly more frequent occurrence that girls get married without their parents’ approval, very often even before they finish elementary school. In Vranje in fact a certain departure from the strict observance of the ideal patriarchal model can be noticed. Formal insistence on the bride’s virginity is still present (and positive examples given of unmarried girls older than 20 who are still virgins), but in practice this is frequently counterfeited, with the knowledge and approval thereof on the part of all the actors. 53 Section 8: Conclusion and Recommendations Through its findings this report contributes to the policy dialogue in three unique ways: 1. By analyzing behavioral norms that buttress actions, we can build on the momentum of norms that are already yielding to change in order to bring about structural change. 2. By drawing attention to the variation within the Roma populations, we can focus on interventions that are tailored to location- and context-specific barriers rather than designing for Roma households as a whole. 3. By advancing a structural perspective on marginalization it calls for a multi-sectoral approach to reducing both the Roma and non-Roma gap, as well as the gender gap in outcomes. First, we need to ask what a structural marginalization perspective adds to a study of social norms from the policy perspective. Some answers are found in the following example. Roma respondents report a preference for early marriage and for withdrawing their girl child from school early. Considered in isolation, this could be seen as a regressive gender norm attributable to the community itself that possibly needs behavioral interventions to “rectify.” However, contextualizing these same findings within the constrained opportunity structures surrounding them broadens the focus for intervention beyond the Roma and prevents blaming a marginalized community and its behavioral norms for its own outcomes. The context for girls’ early marriage and school dropout includes factors such as discrimination by the wider social environment, lack of safety in schools and public spaces for Roma girls, unsafe housing communities, and widespread unemployment. Let’s take another example. Roma respondents may report limited employment aspirations, and their ideal jobs can differ vastly from the non-Roma population. Considered in isolation, these findings might invite remediation thorough a Roma-targeted skills training program in order to build aspirations and skills. However, these same findings when contextualized help explain why Roma aspire to be hairdressers and car mechanics and non- Roma aspire to be doctors and lawyers. One contextual factor is that, despite a similar developmental starting point, Roma children are gradually sorted into special and vocational schools that offer training for hairdressing, baking, and manual labor. This happens because mainstream schools are unable to accommodate children that speak Romani or lack appropriate documentation. With this context, our efforts are turned to designing inclusive institutions (in this case of schools) rather than making behavioral “nudges” at the individual or community level. Second, we need to ask what does a nested stigmatization perspective and the fact of diversity within the Roma population add to the policy dialogue? Let’s take another example. If we mistakenly regard the Roma in Kamendin as a single community, we are likely to design 54 interventions that are one size fit all. For places where housing is mixed, which is now increasingly the fate of the Roma, there need to be location-specific interventions. In a paper titled “One Camp, but Three Refugee Groups: Challenges for Local NGOs,” 74 the authors highlight the development challenges faced in refugee camps where interventions are designed from the outside for refugees as a whole, but each group (Syrians, Palestinians from Syria, and Palestinians from Lebanon) perceives themselves to be more disadvantaged than the other. The groups compete for limited resources, which increases tensions in the camp. The authors suggest labor market and education interventions that simultaneously promote social cohesion within the camps, and tackle the deep-seated discrimination faced by each refugee group from the rest of the country. Similarly, several neighborhoods of Belgrade that the government and policy makers alike consider homogenously Roma are in fact considered mixed neighborhoods by those inhabiting them. The divisions between the four different kinds of Roma—IDPs from Kosovo, refugees from other parts of the Balkans, “original” Belgrade Roma, and returnees from EU countries—are stark, leading to a neighborhood filled with internal strife and tension. In places like Kamendin and other mixed Roma neighborhoods of Belgrade and similar urban areas it is important to think of interventions that take into account the tension and diversity within, and design projects that have a strong social cohesion component as well. Examples of such interventions are detailed below. Finally, we need to ask how it matters for policy that certain norms are in flux more so than others. It matters because interventions can build on the momentum of those shifting norms in order to bring change. For instance, gender norms around financial decision making are yielding to change quicker and more easily than those around decision making on social issues; thus, the former is a stronger entry point to bring about women’s empowerment as a whole. Similarly, if Roma parents are shifting norms around sending their children to kindergarten rather than keeping them at home, then focusing on removing the structural barriers to kindergarten attendance is more likely to be successful. The more deeply entrenched social norms require a combination of both behavioral as well as large-scale cultural change that challenges traditional structures of power and patriarchy and leads to greater gender equality. The audience for this study is the Government of Serbia, the World Bank, Roma NGOs, and other international donors. Based on the findings of this study, the following short- and long-term policy and operational proposals are recommended for further investigation and consideration. 74 Mahmound and Roberts (2018). 55 Strategy #1: Build on the Momentum of Shifting Norms to Bring Structural Change R1. Vocational educational opportunities should be synced with growth sectors (such as engineering, natural sciences, education, and high-technology), so that Roma boys and girls who are more likely to attend vocational schools can transition into these sectors. Currently the types of skills taught in vocational schools (hairdressing and baking for girls and manual labor or construction work for boys) limit the capacity of Roma boys and girls to aspire compared to their non-Roma counterparts who attend mainstream schools, and gender segregates certain types of jobs. Suggested interventions include revising the structure of vocational colleges to meet demand from growth sectors, providing vocational training scholarships to girls, and creating aspirational quotas for female attendance. Such policies could (a) capitalize on the recent attitudinal shifts towards education and wanting to keep children in school longer (particularly girls), and (b) bring about a shift in the types of occupations that young Roma boys and girls aspire to. Vocational courses could also be made available to rural women on business and associated support skills (for example, accounting and legal training). Policies could also target the importance of girls finishing secondary school, as a support to transition to higher education or employment. R2. Public and promotional campaigns in combination with soft skills, training and networks can both highlight successful women entrepreneurs and adolescent girls and build their skills. Roma women (particularly the youth) unlike their non-Roma counterparts believe that there are limited role models from their community to aspire to, which makes it difficult for them to visualize career pathways. Several of them have been pulled out of schooling and left the labor force early and/or involuntarily, and they desire to see a better future for their girls. Showcasing a diverse range of successful Roma students and business women at the neighborhood, village, or municipality level would go a long way in helping Roma girls. To this end, campaigns promoting success stories can encourage girls and women to pursue education and business activities and help convince men of the benefits of women’s increased agency and employment opportunities. The global evidence on the cost effectiveness and impact of campaigns alone on changing gender norms (and any identity-based exclusion) is weak. Yet, it continues to be a favorite tool for government programs because of its short-term (sometimes political) appeal. Evidence suggests that they seem to work better as part of a bundle of interventions that address information gaps or provide awareness about a program’s goal or as part of a community-based intervention (as an accompanying intervention but not a stand-alone intervention). In order to promote Roma women entrepreneurs and normative change about Roma female students, we propose a combination of networks, soft skills training, along with campaigns. An excellent model to consider is the type of intervention piloted (and rigorously evaluated) under the Adolescent Girls Initiative (see Box 1 below on AGI). 56 Box 1: Adolescent Girls Initiative (AGI) AGI was implemented between 2008-15 by the World Bank in eight countries in partnership with the Nike Foundation and the governments of Afghanistan, Australia, Denmark, Jordan, Lao People’s Democratic Republic, Liberia, Nepal, Norway, Rwanda, South Sudan, Sweden, and the United Kingdom. The initiative was designed to better understand what works in helping adolescent girls and young women transition to productive employment. The AGI piloted and rigorously evaluated innovative interventions that could be scaled up or replicated if they were found to be successful. Interventions included business development skills training, technical and vocational training targeting skills in high demand, and life skills training. The initiative has generated some of the strongest evidence to-date on the impact of employment programs for young women. The AGI pilot implementation experience has also generated valuable operational lessons on how projects can achieve employment impacts. Operational lessons are shared in this Resource Guide. Source: See AGI Resource Guide Strategy #2: Advance Location- and Context-Specific Reform R3. Accelerated Learning Programs (ALPs) can be introduced in primary schools in order to support the integration of Roma children, particularly IDPs and returnees into “mainstream” education. Roma children seem to have a host of learning disadvantages in primary school, including not attending kindergarten like many of their non-Roma counterparts; speaking Romani (or for returnees, German or Italian) rather than Serbian (which is the language of the curriculum); and facing “other-ing” and discrimination from a very young age. For these children, an Accelerated Learning Program can help develop the requisite coping mechanisms and prepare them for mainstream education. This program can be aligned with or led by the newly introduced Roma teaching mediators in school. Box 2 presents an example of an ALP implemented throughout Lebanon in order to integrate children from Syria into mainstream education. Another example of an ALP is the Komplex Instrukciós Program in Hungary, based on a pedagogical approach developed at Stanford University in the 1970s that has been adopted worldwide. Komplex Instrukciós aims to create a more equitable classroom atmosphere in which status differences among pupils are not recognized and do not hinder learning. The methodology is especially conducive to high-quality education in diverse classrooms in which learning abilities and social backgrounds vary. It encourages children to become active partners in their own learning. The program builds on three methodological pillars: multiple ability assignments, group work, and status treatment. Given that Roma children tend to live in lower socioeconomic status 57 households and may face learning challenges because of language barriers and family background, such a methodology may be especially effective in helping Roma children catch up to their non-Roma peers (Robayo-Abril and Millan 2019). Box 2: Accelerated Learning Programs (ALPs) for Syrian children in Lebanon Accelerated Learning Programs have educational and psychosocial aims. They are designed to help children catch up on missed education, build skills to study independently, and develop coping mechanisms to help them deal with bullying they might encounter because they are behind or new to the school. ALPs have focused on developing the language skills of children from Syria so that they can participate in schools in Lebanon, where the curriculum is often taught in English and French. Elements of the Syrian ALPs help children to deal with trauma, and support is provided to parents to help them understand how their children have been affected by their experiences and how they can support them. The holistic approach has been praised by the children taking part, by parents, and by the implementing staff. Source: Mahmoud and Roberts (2018) R4. Increase awareness of existing scholarships and training opportunities for Roma at the municipality and village level. Of the five communities interviewed across the country, not a single Roma parent was aware of scholarships or affirmative action for Roma children in schools. Of the Roma returnees and IDPs, none were aware either. Better financial aid and academic supports targeting Roma students, especially as led by the Roma Education Fund, have paved the way for change, but increased awareness is needed among Roma communities on available programs in their specific locations. Strategy #3: Take a Multi-Sectoral Approach to Roma interventions R5. Support a women’s self-help group (SHG) program with Community Driven Development (CDD) elements that cuts across Roma and non-Roma. Roma-focused interventions in Serbia have either been targeted towards a specific sector such as education or employment, or towards the Roma alone. Instead what is essential is to design interventions that take into account intersectionality. Gender, ethnicity, and class create specific constraints for Roma women. Isolated gender equality or anti-racist interventions do not always benefit women, and sometimes are even have adverse effects.75 Bringing together into the same self-help group women from diverse ethnic communities but a similar socioeconomic group can help build 75 Kóczé and Raluca (2009). 58 solidarity among SHG members towards a common goal – increasing savings and access to credit for consumption smoothing at the very least, and create entrepreneurial skills and small businesses at the most. SHG programs have been successful not only in giving access to credit, but also in building social capital by giving economically and socially disadvantaged women access to a well-defined network of people (by cutting across identities). They also cultivate new capabilities that have long-term impacts on livelihoods, voice, and bargaining power within households. Such capabilities could defy deeply entrenched conventions of gender, and more importantly increase notions of self-worth among Roma women. See Box 3 for an example of a World Bank funded self-help group intervention in India, which cuts across caste in bringing together poor women. Box 3: Jeevika, a women’s livelihood program in Bihar, India Jeevika, or the Bihar Rural Livelihoods Project (funded partially by the World Bank) began over a decade ago in 2006 in a very challenging environment. At that time, the Bihar’s rural poverty ratio was 44.6 percent: the absolute number of people living in poverty in the state was at 36 million out of the total population of 82 million. Poor households were unable to invest in self- employment opportunities, both because of limited access to credit and assets, and limited aspirations. Moreover, gender and caste hierarchies were the most oppressive in Bihar, compared to the rest of India. Under these challenging circumstances, Jeevika was introduced and women from the poorest households were mobilized into community institutions, that is, self-help groups. Through these institutions, rural households could access low-cost credit, leverage services from the formal financial system, and gain information and training on improved livelihoods in the farm and non-farm sectors. The project’s rationale was that with greater social capital—gained from joining self-help groups—the poor would have a greater say in the quality of services they received. As of 2018, Jeevika has mobilized 8.2 million women. The self-help groups have collectively saved US$65.6 million and leveraged more than US$550 million of formal credit from banks. Because cheaper credit was channeled into the household through the woman, it gradually increased women’s bargaining power in the household. More women were emboldened to enter spheres of activity outside the household. Their increased participation in civic, political, and financial institutions began to break down long-standing normative restrictions on what lower caste women in particular could achieve, thereby de-stigmatizing them. More important, by first acknowledging deep-rooted gender norms specific to lower caste women, the project took actions at multiple levels in order to undo them (targeting individuals, husbands, local banks, public institutions holistically). This not only broke down long-standing normative restrictions on what lower caste women in particular could achieve and aspire to, thereby de- 59 stigmatizing them, but more critically, by creating a new identity anchored squarely in the SHG and poverty (rather than caste or kinship) it created social capital where there was none before. Sources: Sanyal, Rao, and Majumdar (2015). R6. Build in social cohesion and social inclusion components, especially for heterogeneous communities and neighborhoods. In Kamendin and Vranje a significant discrepancy has been observed between the “original residents” and “newcomers”, who are for the most part IDPs from Kosovo and Roma returnees. This has led to considerable tension and conflict among the Roma communities to the extent that it has spillover effects on perceptions of safety in public spaces, and on physical mobility of women and girls. Targeted interventions are necessary to respond to this. An example of such an intervention is the Social Inclusion Cash Grants within the Citizens’ Charter Project in Afghanistan, 76 which is a key part of the government’s response to the displacement (IDP/returnee) crisis in the country. These are grants that target poor and very poor households (female-headed or disabled households, old persons, drug addicts, and so forth) in order to mitigate poverty. A specific focus is on collective action: communities are expected to contribute to a goods/food bank collectively to help the most vulnerable households, and the government provides matching funds equaling the value of the community’s in-kind contribution. Without an intervention of this nature, where transfers are contingent on communities coming together, the conflict among the Roma would only be further exacerbated, especially with the arrival of new communities. It is important therefore to take this diversity into account and address it directly through such interventions. R7. Finally, generating more evidence on existing Roma interventions is key. In particular, evidence should be gathered on (a) whether or not social categorical targeting on Roma specifically is working; (b) the impacts of existing Roma-specific interventions (the Roma Education Fund for instance has done a lot of work on education but with few impact evaluations); (c) the potential for scale-up of existing interventions that are successful in certain municipalities; and (d) tracking new interventions, including with strong M&E and feedback loops. Data collection and M&E will be key for ensuring that barriers from social norms are relieved and not reinforced due to World Bank and/or government interventions. 76 In recent years, Afghanistan has faced a displacement crisis due to high numbers of returnees and IDPs from Pakistan and Iran in 2016. Even larger numbers are expected from these and other places over the next few years. In 2017, the World Bank and Afghanistan Reconstruction Trust Fund (ARTF) jointly responded with a total of $US140 million to Afghanistan’s Ministry of Rural Rehabilitation and Development (MRRD), providing 14 districts within the country with Social Inclusion Grants to eligible communities, specifically targeting the poorest/most vulnerable households in all communities. 60 The kinds of interventions that are most likely to succeed will create multiple, repeated instances of situations where women and Roma participate (more) equally and are acknowledged as equally competent to similar men and non-Roma counterparts at socially valued tasks. Beliefs about the status of a group of people play a crucial role in the process through which inequality across different dimensions is generated and justified. Interventions that involve, in the course of jointly working toward accomplishing a common goal, between individuals who belong to separate nominally different categories (in this case women and men, or Roma and non-Roma) who are also different in their resource position – play a vital role in the genesis and diffusion of norms and beliefs, until they become consensually-held beliefs about the status value of the nominal categories. 61 References Allen, Richard. 2016. “IDPs in Serbia: Summary Report and Proposals.” Report, United Nations High Commissioner for Refugees (UNHCR). Belgrade, Serbia. Appadurai, A. 2004. “The Capacity to Aspire: Culture and the Terms of Recognition.” In M. Walton and V. Rao, editors, Culture and Public Action: A Cross-Disciplinary Dialogue on Development Policy (pp. 59–84). Washington, DC: World Bank Publications. Bodewig, C., and A. Sethi. 2005. “Poverty, Social Exclusion and Ethnicity in Serbia and Montenegro: The Case of the Roma.” World Bank's Programmatic Poverty Assessment for Serbia and Montenegro. Belgrade: World Bank. Butler, Judith. 2004. Undoing Gender. Psychology Press. CARE International. 2011. “Situational Analysis of Education and Social Inclusion of Roma Girls in Serbia.” Report, CARE International, Serbia, Belgrade. Council of Europe (CoE). 2012. “Protecting the Rights of Roma.” Link: http://www.strassburg- europarat.diplo.de/contentblob/3392612/Daten/1862070/CoE_publication__Roma.pdf. Corradi, L. 2017. Gypsy Feminism: Intersectional Politics, Alliances, Gender and Queer Activism. Routledge. De Laat, J., and C. Bodewig. 2011. “Roma Inclusion is Smart Economics.” Knowledge Brief, Europe and Central Asia 39 (April 2011). World Bank, Washington, DC. Drezgić, R. 2010. “Religion, Politics and Gender in the Context of Nation-State Formation: The Case of Serbia.” Third World Quarterly 31(6): 955–70. Durst, Judit. 2002. “Fertility and Childbearing Practices among Poor Gypsy Women in Hungary: The Intersections of Class, Race and Gender.” Communist and Post-Communist Studies 35(4): 457–74. European Roma Rights Centre (ERRC). 2017. “Written Comments of the European Roma Rights Centre Concerning Serbia. For Consideration by the Human Rights Council, Working Group on 62 the Universal Periodic Review, of the 29th Session (January–February 2018).” Available at: http://www.errc.org/cms/upload/file/serbia-hrc-submission-30-june-2017.pdf. EU COM 2008. Communication from the Commission to the European Parliament and the Council Report on the Evaluation of the EU Framework for National Roma Integration Strategies up to 2020. See link: https://eur-lex.europa.eu/legal- content/EN/TXT/?qid=1544112037077&uri=CELEX:52018DC0785 Gatti, R., S. Karacsony, K. Anan, C. Ferré, and C. de Paz Nieves (eds.). 2016. Being Fair, Faring Better: Promoting Equality of Opportunity for Marginalized Roma. Washington, DC: World Bank. Government of the Republic of Serbia (GoRS). 2010. “Strategy for Improvement of the Status of Roma in the Republic of Serbia.” GoRS, Belgrade. Available at: http://www.seio.gov.rs/upload/documents/ekspertske%20.misije/protection_of_minorities/str ategy_for_roma.pdf. Kabeer, N. 2016. “Gender equality, economic growth, and women’s agency: the “endless variety” and “monotonous similarity” of patriarchal constraints”. Feminist Economics, 22(1), 295-321. Kagin, J. D. 2013. “Norms, Values and the Effect of Churches on an Excluded Community: The Case of the Roma in Eastern Europe.” PhD Thesis, University of California, Davis. Kolin, M. 2010. “Gender Inequalities and the Status of Women in the Labor Market in Transitional Serbia.” Anthropology of East Europe Review 28(1): 119–37. Kóczé, A., and M. P. Raluca. 2009. “Missing Intersectionality: Race/Ethnicity, Gender and Class in Current Research and Policies on Roma Women in Europe.” Center for Policy Studies, Central European University. Lorber, Judith. 1994. Paradoxes of Gender. Yale University Press. Mahmound, Olfat, and Rebecca Roberts. 2018. “One Camp, Three Refugee Groups: Challenges for Local NGOs.” Forced Migration Review 57 (February): 7–9. Marjanovic, Dragana. 2015. “Assessment of the Structural Barriers that Hinder the Employment of Roma Women.” UN Women. Accessed at: 63 http://www.europeanprogres.org/dokumenti/48_646015_asesment-of-the-structural-bariers- for-roma-women-employment.pdf. Mihajlović, M. 2004. “Needs Assessment Study for the Roma Education Fund–Serbia (without Kosovo).” Roma Education Fund, Budapest. Muñoz Boudet, Ana María, Patti Petesch, Carolyn Turk, and Angélica Thumala (eds.). 2013. On Norms and Agency: Conversations about Gender Equality with Women and Men in 20 Countries. Directions in Development: Human Development. Washington, DC: World Bank. © World Bank. https://openknowledge.worldbank.org/handle/10986/13818 License: CC BY 3.0 IGO. Jovanovic, Zeljko, and Mensur Haliti. 2012. “Roma Feel Less Fear and More Hope After Census.” Open Society Foundations, December 12. Available at: https://www.opensocietyfoundations.org/voices/roma-feel-less-fear-and-more-hope-after- census. Organisation for Economic Co-operation and Development (OECD). 2010. OECD Employment Outlook 2010: Moving Beyond the Job Crisis. Paris: OECD. 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. Oxfam and Argument. 2001. “The Roma Livelihood in Belgrade Settlements.” Survey, Oxfam/Argument, Belgrade Pešić, J. 2006. “Persistence of Traditionalist Value Orientations in Serbia.” Sociologija 48(4): 289–307. Powell, John A. 2013. "Deepening Our Understanding of Structural Marginalization." Poverty and Race 22 (5): 3–5. Praxis. 2011. “Analysis of the Main Problems and Obstacles in Access of Roma in Serbia to the Right to Education.” Praxis, Belgrade. Available at: https://www.regikonyvek.hu/kiadas/analysis-of-the-main-problems-and-obstacles-in-access-of- roma-in-serbia-to-the-right-to-education-2011-praxis. Praxis, Institute on Statelessness and Inclusion (ISI), European Network on Statelessness, and European Roma Rights Centre (ERRC). 2017. “Joint Submission to the Human Rights Council at 64 the 29th session of the Universal Periodic Review (Third cycle, 15–26 January 2018).” Serbia, 29 June 2017. Available at: http://www.institutesi.org/UPR29_Serbia.pdf. Praxis and UNHCR. 2017. “Stop to Child Marriage.” Belgrade. Available at: http://praxis.org.rs/images/praxis_downloads/Stop_Decijim_Brakovima.pdf. Reva, A. 2012. “Gender Inequality in the Labor Market in Serbia.” World Bank Policy Research Working Paper No. 6008. World Bank, Washington DC. Ridgeway, Cecilia L. 1991. “The Social Construction of Status Value: Gender and Other Nominal Characteristics.” Social Forces 70(2): 367–86. Risman, Barbara J. 2004. “Gender as a Social Structure: Theory Wrestling with Activism.” Gender & Society 18(4): 429-50. Robayo-Abril, M., and N. Millan. 2019. Breaking the Cycle of Roma Exclusion in the Western Balkans. Washington, DC: World Bank Group. Sadiq, K. 2008. Paper Citizens: How Illegal Immigrants Acquire Citizenship in Developing Countries. Oxford University Press. Sanyal, P., V. Rao, and S. Majumdar. 2015. “Recasting Culture to Undo Gender: A Sociological Analysis of Jeevika in Rural Bihar, India.” Policy Research Working Paper WPS 7411. World Bank, Washington, DC. Available at: http://documents.worldbank.org/curated/en/346831467995637047/Recasting-culture-to- undo-gender-a-sociological-analysis-of-Jeevika-in-rural-Bihar-India. 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): 37–43. Surdu, M. 2016. Those Who Count. Central European University Press. United Nations Development Programme (UNDP), World Bank, and European Commission (EC). 2011. Regional Roma Survey (RRS) 2011. https://cps.ceu.edu/article/2012-10-16/undpworld- bankec-regional-roma-survey-2011. United Nations Development Programme (UNDP), World Bank (WB), and European Commission (EC) (UNDP-WB-EC). 2017. Regional Roma Survey (RRS) 2017. Country fact sheets. 65 http://www.eurasia.undp.org/content/rbec/en/home/library/roma/regional-roma-survey- 2017-country-fact-sheets.html. United Nations Children's Fund (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 Surveys in Roma Settlements in the Three Countries.” UNICEF, Geneva, July 2015. https://www.unicef.org/eca/21352_UNICEF_Roma_Woman_web_V2.pdf United Nations Children's Fund (UNICEF). 2017. “Child Marriage Among the Roma Population in Serbia: Ethnographic Research.” UNICEF, Belgrade, November 2017. https://www.unicef.org/serbia/Child_marriage_among_the_Roma_population_in_Serbia_web. pdf. United Nations High Commissioner for Refugees (UNHCR). 2016. “Persons at Risk of Statelessness in Serbia: Progress Report 2010 – 2015.” Edited by Slobodan Cvejić. UNHCR, Belgrade, June 2016. http://www.refworld.org/docid/57bd436b4.html. UNHCR/UN OCHA. 2004. “Analysis of the Situation of Internally Displaced Persons from Kosovo in Serbia and Montenegro: Law and Practice.” A Legal Analysis prepared by the IDP Interagency Working Group, June 2014. West, Candace, and Don H. Zimmerman. 1987. “Doing Gender.” Gender & Society 1(2): 125– 151. World Bank. 2010. “Roma Inclusion: An Economic Opportunity for Bulgaria, Czech Republic, Romania and Serbia.” Policy Note. World Bank, Washington, DC. ———. 2013. “Roma Inclusion in the Western Balkans: Country Briefs.” Policy Note. World Bank, Washington DC. ———. 2015a. “Republic of Serbia: Country Partnership Framework 2016 – 2020.” Report No. 100464-YF. World Bank, Washington DC. http://documents.worldbank.org/curated/en/650701467993479160/Serbia-Country- partnership-framework-for-the-period-FY2016-20 66 ———. 2015b. “Serbia Systematic Country Diagnostic.” Report No. 100471-YF. World Bank, Washington DC. http://pubdocs.worldbank.org/en/63111446462343770/Serbia-SCD-web.pdf ———. 2017. “Serbia—Inclusive Early Childhood Education and Care Project (English).” Washington, D.C.: World Bank Group. http://documents.worldbank.org/curated/en/918021487905271830/Serbia-Inclusive-Early- Childhood-Education-and-Care-Project. ———. 2019. World Development Report 2019: The Changing Nature of Work. Washington, DC: World Bank. ———. Forthcoming. “Serbia Policy Note: Social Protection and Labor.” World Bank, Belgrade 67 Annex 1: Sampling Strategy for Qualitative Study Regions for sampling were selected based on criteria including: national geographical coverage, high regional poverty rate and high percentage Roma population. The three regions selected were Belgrade, Vranje in the south, and Novi Becej in Vojvodina in the north, and these three represent one urban, one peri-urban (located less than 10-15 km from urban area) and one semi rural region respectively. Remote rural settlements i.e. more than 40 km from urban area are not typical of Roma settlements, hence these were not selected as part of the sample. Within each region, one municipality each was selected which was a subset of the quantitative sample of the RRS 2017. All three municipalities have high poverty rates and high percentage Roma. Within each of the three municipalities one Roma settlement each was selected that is the most representative of the type of Roma settlement in that region (with the exception of Belgrade where two different types of settlements were selected). be selected. Note: again, ideally these are representative of the nature of Roma settlements in the municipality i.e. select an urban Roma settlement for Central / Belgrade if that is the most representative of the type of Roma settlement in that municipality. Within each area, 5 FGDs (of five respondents each) and 10 key informant interviews were conducted in marginalized Roma settlements (i.e. with more than 40% Roma). The Bibija team worked closely with community leaders and key informants in order to shortlist from the quantitative sample, settlements that represent sufficient diversity in terms of housing type, home ownership status and age of settlement. Once identified, the field team took the assistance of key stakeholders in the community to screen and enlist the men and women of these households for FGDs. Three kinds of FGDs were conducted and these will be further split into three – Roma men, Roma women, and Roma youth. a) Roma women – working and non-working: i.e. a mix of Roma women that do not have income and women starting to make an income or have been working for over 5 years (FGDs). Additional criteria apart from working history include the following: marriage status (never married, married, single parent) and educational level (no education above the compulsory level / some education above the compulsory level). a) Roma men – working and non-working: i.e. a mix of Roma men that do not have income and men starting to make an income or have been working for over 5 years (FGDs). Additional criteria apart from working history include the following: marriage status (never married, married, single parent) and educational level (no education above the compulsory level / some education above the compulsory level). 68 a) Roma youth – working and non-working: i.e. a mix of Roma youth that do not have income and youth starting to make an income or have been working for over 5 years (FGDs). Additional criteria apart from working history include the following: gender (both boys and girls), marriage status (never married, married, single parent) and educational level (no education above the compulsory level / some education above the compulsory level). For the 10 key informant interviews in each area, we selected at least two Roma women (working and non-working), two Roma men (working and non-working), two Roma youth (same), and four non-Roma respondents (men, women and youth) in the vicinity of the Roma settlement. Selection of Regions, Municipalities, Settlements and Respondents (Total number of respondents, 94) Serbia Region 1 in South Region 2 in North Region 3 in Center (high pov, high % (high pov, high % (high pov, high % Roma) Roma) Roma) Municipality 1 (high Municipality 2 (high Municipality 3 (high pov, high % Roma) pov, high % Roma) pov, high % Roma) Semi-Urban Proximate Rural (24 Urban (24 respondents) respondents) (24 respondents) 10 interviews 5 FGDs (Roma men, (Roma men, Roma Roma women, women, Roma Roma youth) youth, non-Roma) 69 Both interviews and FGDs were conducted by two interviewers at a time from the Institute of Ethnography (one facilitating the conversation, and the other taking notes) in the homes of neighbors or local cafes in complete privacy. During pilot stage, several locations were considered for interviews and FGDs, and the homes of neighbors and local cafes were found to be most comfortable for the respondents (as opposed to the Bibija office or homes of community leaders – both of which skewed their responses). They felt at ease, and at the same time distance enough from their own household members to be able to talk openly and privately. Finally, only those who self-identified as Roma were selected for the Roma interviews regardless of how they identify in the census. In addition, note was taken of how they identify with this category. Annex 2: In-depth Interview Discussion Guide - Roma Women and Men Moderator Note This particular document is addressing a Roma primary female caregiver or male head of household ages between 25-65. The objective of this discussion guide is to present a suggested way of tackling the topics of interest for this qualitative study. To that end—and as expert qualitative researchers— the moderators should start by establishing strong rapport and building trust with the respondents as well as allaying any concerns they may have as a segue to conduct naturally flowing interviews. As need be, moderators should have the freedom to reword questions or ask them in a different order to capture relevant information in case the respondents bring them up independently. In order to facilitate the discussion, there is a brief description at the beginning of each section that lists the general topics to be covered, as well as suggested probing questions that would help address these topics more comprehensively. Before starting the discussion, a moderator should stress that there are no right or wrong answers, and that they are interested in everything the respondent has to say. Introduction Introduction to the Interview and Purpose of the Research: First of all, I want to thank you for taking the time to meet with me today. My name is XXXX, and I am part of a research team at the SASA Institute of Ethnography working in collaboration with the World Bank to conduct a 70 research with Roma regarding drivers and barriers to education and labor force participation, particularly for females. Purpose of the interview: Today, I would like to talk to you about your education background, work inside and outside the house, your role in making decisions related to yourself, as well as your opinions towards education, employment and early marriage. Interview Length and Incentive: The interview should take no longer than 90 minutes. In return I would like to thank you by offering a voucher for Lily, the neighboring pharmacy and retailer. Confidentiality and anonymity: All responses will be kept confidential. This means that your interview responses will only be shared with project team members and we will ensure that any information we include in our report does not identify you as the respondent. No names, addresses or any identity information will be shared with anyone else - including other project team members. This information will be kept completely separate from the other information that you give me. The recordings and notes from this interview will be stored safely as well to ensure that it is not shared with anyone else. Right to decline: You don’t have to talk about anything you don’t want to and you may end the interview at any time Questions and Concerns: Are there any questions about what I have just explained? Do you have any concerns that we may allay before starting the interview? Consent to conduct the Interview: Are you willing to participate in this interview? Yes/No Recording: We would like to record the session because we don’t want to miss any of your comments. Although we will be taking some notes during the session, we can’t possibly write fast enough to get it all down. Consent to Recording: Are you willing to let us record it? Yes/No A. Background ● Collect brief personal details: age, marital status, number and age of children. RECORD ON FACT SHEET. FACT SHEET Question Answers Comments Gender 1. Male 2. Female 71 Age 1. 25-35 2. 36-45 3. 46-55 4. 56-65 Place of Birth 1. Country 2. Capital 3. District center / City – municipal center/city 4. Town 5. Village 6. Unregulated area Place of Residence 1. Country 2. Capital 3. District center / City – municipal center 4. Town 5. Village 6. Unregulated area Marital Status 1. Married – traditionally 2. Married – officially 3. Married – both traditionally and officially 4. Divorced 5. Separated 6. Widowed/ partner passed away 7. Cohabitation 8. Never married Age at marriage 1. Less than 18 2. 18 and above Number of household members Number and age of children Current/ Highest Level 1. No formal education of Education of 2. Incomplete primary respondent 3. Completed primary 4. Completed primary special school 5. Incomplete secondary school 72 6. Incomplete Second chance education 7. Completed secondary school 8. Completed secondary special school 9. Completed Second chance education 10. Post-secondary education other than college/university 11. Associate (2-3 years) College 12. Incomplete university 13. Bachelor 14. Masters 15. PhD / Specialist Economic Activity 1. Working – full time 2. Working – part time 3. Working – ad hoc jobs 4. Self-employed 5. Full time homemaker (looking after the home/children/relatives) 6. He/she is on paid parental leave 7. Doing unpaid work in family business (enterprise/farm/other) 8. Doing other unpaid or voluntary work 9. Not working 10. Asking for money [INT: Begging] 11. Student/pupil/ 13. In vocational training / apprenticeship 14. Retired 15. Too old to work 16. Unable to work due to long-term illness / disability 17. In compulsory military / community service 95. Other, specify 73 RF/DK (Refused/Don’t Know – do not read) Current/ Highest Level 1. No formal education of education of spouse 2. Incomplete primary 3. Completed primary 4. Completed primary special school 5. Incomplete secondary school 6. Incomplete Second chance education 7. Completed secondary school 8. Completed secondary special school 9. Completed Second chance education 10. Post-secondary education other than college/university 11. Associate (2-3 years) College 12. Incomplete university 13. Bachelor 14. Masters PhD / Specialist Current/ Highest Level 1. Preschool/ kindergarten/ crèche of education of each 2. Primary child <18 years old 3. Primary special school (specify age and gender 4. Secondary of child) 5. Secondary special school 6. Second chance education 7. Other (specify) Religious Affiliation 1. Orthodox 2. Catholic 3. Protestant 4. Muslim 5. No religion Other (specify) 74 Regular Income for 1. Amount respondent and spouse 2. Frequency (hourly, weekly, (from paid work) monthly, yearly) Other regular income 1. Amount for respondent and 2. Frequency (hourly, weekly, spouse (other than paid monthly, yearly) work) A. Time Use The objective of this section is to explore the division of labor among household members, including work inside and outside the house, household chores, and other social activities and engagements. It is important to understand who decides the allocation of such daily activities, the amount of time spent on each activity, as well as the respondent’s opinion regarding this arrangement—particularly of their own activities. Understanding gender roles are of ultimate importance during the discussion. I would like to start by asking you about the daily life of your family members in the household. On a normal day, could you describe what kind of work and chores inside and outside the house does each family member do? (e.g. formal paid work, informal paid work, household chores, playing, taking care of family, looking for work, accessing basic services, engaging in social networks/ organizations/associations, etc.)? Probes a. Who decides this allocation? b. How many hours spent on each of the chores/ activities? c. How do you feel about this arrangement? Would you wish to change it one way or another? Why or why not? What would be the ideal arrangement for them? B. Household Decision-making Still focusing on the family, this section explores the intra-household dynamics and decision- making regarding important aspects of the livelihood of family members. It is preferable that the respondent gives an example or two of decisions that were made in the past (including the process of decision-making and the outcome of such decisions), whether the impact of the decision concerns the respondent him/herself or another family member. The important thing is to understand to what extent are the opinions of different family members solicited including the person whom the decision will impact—all while focusing on 75 gender and age nuances. Now I would like to ask you about how important decisions (that concern you or any family member) are made in your household. Can you give me an example or two in more details? Probes a. The moderator can suggest topics such as: migration and relocation, spending (esp. allocation of funds to children education and health versus other expenses), saving, buying/ selling assets, borrowing/ getting into debt, having control over one’s own money and over the entire earnings of the household, reproductive matters, or matters related to work, marriage, and education despite they might be brought up in later sections. b. What was the role of the decision-maker versus the person whom the decision was supposed to impact? Does age or gender of the person affect the extent to which their opinion matters/ is solicited? c. What was the short- and long-term outcomes of the decision? How did the decision impact the concerned person? d. How do you feel about the process of making that decision? Should the power of decision have been more vested in one family member versus the other? Why or why not? C. Education This section shifts the focus on the experience of the respondent in schools, decisions around their education, and the perceived barriers they have faced in accessing/ continuing education. Moderators should also explore how respondents go about decisions regarding their children education. Now, let me ask you more questions about your experience and opinions with regard to the topic of education. 1. As a child and young adult, what has been your experience in schools so far? Probes a. Classroom composition: Roma vs non-Roma b. Perceived discrimination: from teachers/ admin staff, bullying from other students, discrimination for being a girl, etc. c. Explore ethnic (Roma and non-Roma) and gender (boys and girls) relations as well as experience in special schools 76 If Respondent is currently not (or have never been) in education: 2. Why did you not complete your highest level of education (if incomplete) or transition to the next education level? Probes a. Dynamics of the decision-making: Who made the decision? Was the opinion of the respondent solicited? Did they accept or resist? b. Reasons driving the decision: e.g. issues with documentation, migration and relocation, discrimination incidents, low performance/ failure to graduate, early marriage, need to work and support family, need to stay home to take care of household chores or family, tradition within family to quit education at a certain level, distance to school and transportation issues, lack of encouragement from home, active discouragement from home, etc. c. Anything you can say about how you compare to your spouse in terms of education attainment and barriers to continuing education. 3. How do you think decisions around continuing (or discontinuing/ dropping out) your education has affected your life and employment opportunities? Probe fully for disadvantages and advantages of either outcome. 4. What about your children? How do you and your spouse think about education decisions for your young children? Probes a. Dynamics of the decision-making: Who made/ would make education decisions for a child? Was/ would the opinion of the child be solicited? Did they accept or resist? b. How important is education for your children? What level of education do you think is sufficient for them? Why? Note: Focus on gender nuances. c. In general, how important do you think education is for men? What about for women? What level of education do you think is sufficient for men? What about for women? Why? D. Early Marriage The conversation would now shed light on the respondent’s experience and opinions around the topic of early marriage. The moderator would solicit information on the process of decision-making and drivers of early marriage for the respondent as well as their opinions on getting married young for their own children and in general for young males and females. 77 1. Now, let us talk a bit about your experience and opinions around the topic of marriage. Tell me more about the story of your marriage. Probes a. Dynamics of the decision-making: Who made the decision? Was the opinion of the respondent solicited? Did they accept or resist? b. Reasons driving the decision, with emphasis on social and gender norms (in the household/ community) around early marriage c. When you got married, did you continue your education or not? If not, then why? d. If not currently married: plans for getting married, at what age? Why? e. In general, what age do you think is appropriate for men to get married? What about for women? Why? 2. How do you and your spouse think about marriage decisions for your children? Probes a. Dynamics of the decision-making: Who made/ would make the decision? Was/ would the opinion of the child be solicited? Did they accept or resist? b. Reasons driving the decision, with emphasis on social and gender norms around early marriage in the household/ community. Note: Focus on gender nuances. c. What age do you think is appropriate for your children to get married? Why? Note: Focus on gender nuances. 3. At what age do you think a woman should start having children and become a mother? When do you think a man should start having children and become a father? Why? Probes a. Explore opinions about the right of a man and of a woman to decide having children at a later age or not having them at all, b. Explore the right of a man and of a woman to decide how many children he/she can have (i.e., is family planning acceptable and if so, who decides?) E. Labor Force Participation In this section, the moderator would now focus on the topic of employment and paid work outside the house. Like the topic of early marriage, the moderator would solicit information on the process of decision-making and drivers of employment as well as the respondents’ 78 opinions on that front. They should also try to explore how respondents would go about employment decisions for their own children. 1. Ok, now I would like to ask you about your experience and opinions around finding employment outside the house and your experience working. Tell me more about your experience in searching for and finding paid work outside the house. Probes a. Dynamics of the decision-making: Who made the decision? Was the opinion of the respondent solicited? Did they accept or resist? b. Reasons driving the decision, with emphasis on social and gender norms (in the household/ community) around working c. Anything you can say about how you compare to your spouse in terms of age of starting work outside the house, reasons and decisions? d. If ever worked/ currently working: age at first work, type of job, challenges in finding a job e. If not currently working: Why not? plans for searching for/getting a job? What would make them look for work? f. If never looked for a job: reasons why you may not have looked for a job? What difficulties do you expect to encounter when looking for a job? g. What kind of work would you ideally like to do? h. What types of jobs are available for you in reality? Why? Note to moderator: explore perceived/ anticipated challenges driven by ethnicity, gender and other structural barriers. i. In general, what age do you think is appropriate to start looking for a job/ getting employed for men? What about for women? 2. What about your children? How do you and your spouse think about employment decisions for your young children? Probes a. Dynamics of the decision-making: Who made/ would make the decision for a child to work? Was/ would the opinion of the child be solicited? Did they accept or resist? b. What age do you think is appropriate for your kids to start looking for work outside the house? Note: Focus on gender nuances. c. In general, what age do you think is appropriate to for men to start looking for work outside the house? What about for women? Why? 79 F. Early Childhood Development As a last section, we would like to tackle a few questions relating to Early Childhood Development, particularly usage and preferences for formal childcare arrangements versus informal ones. 1. Finally, I would like to ask you about your opinions and preferences regarding childcare arrangements for your children. But let me first ask you about the types of care arrangements you already have. Probes Explore types of care arrangements they already have (e.g. daycare, crèche, kindergarten, nursery, nanny, grandparent, left alone, etc.), and how many hours a day are spent in either of these care arrangements. 2. Now let me talk about your opinions and preferences. In general, what do you think of the idea of sending children to formal day care? What about having them cared by family, friends or neighbors? Probes a. If using/ preferring formal childcare arrangements, ask why (e.g. lack of informal childcare options, time spent at formal childcare program/ facility is useful for the child, enabling the caregiver to look for a job or work outside the house, etc.). Probe fully about childcare as an enabler for women to work. b. If not using/ preferring formal childcare arrangements, then ask why (e.g. plans to send next year, too expensive, no program nearby, nearby program is full, child(ren) not treated well, lack of trust in teachers/ schools, lack of documentation, no benefit perceived from sending to formal childcare, no need for formal childcare, etc.) c. If not using/ preferring formal childcare arrangements, then what would it take to change their minds and opt for formal children options? (e.g. caregiver has to work outside the house, Roma teacher / mediator in school, No school fees (attendance for free), Food coupons for family, Free transport, etc.) 3. What do you think is the appropriate age to start sending children to day care/pre- school/school? When do you think it is acceptable to leave them alone at home? 80 G. Closing Interview ● Is there anything else they would like to say about any of the topics discussed? ● Discuss whether they felt uncomfortable speaking about any of the issues covered, and what reassurances they and others like them would like 81