Gender Dimensions in the Educational Sector in Romania Background Study for the Romania Gender Assessment Monica Robayo-Abril and Britta Rude The World Bank © 2023 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.orgpubrights@worldbank.org. Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank, therefore, does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to reuse a component of the work, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. 2 Summary Addressing gender inequalities in educational outcomes is crucial from a human rights and development perspective. Building human capital early in life has crucial implications for developmental and labor market outcomes later in life. In this background note, prepared to inform the Romania Gender Assessment 2023, we rely on a variety of data sources to descriptively study gender inequalities in educational outcomes, such as enrollment rates and test scores, in Romania. We analyze these inequalities for the total population as well as for different income groups and regions. Our evidence shows that gender equality in aggregate estimates often masks important inequalities between subgroups. While in some cases, boys outperform girls, there are also cases in which girls outperform boys. These patterns differ across income groups, regions, and educational levels. Based on this evidence, the Romanian government should take a nuanced approach to achieving gender equality in the educational sector. Moreover, we find that—in the case of all indicators—both Romanian girls and boys perform significantly below the European average and that there are some negative trends over time, especially with respect to enrollment rates in secondary schools and school performance. Reversing these trends is crucial to ensure the full development of both boys and girls. Moreover, although girls outperform boys in several educational outcomes, these advantages do not translate into the labor market. We identify several constraints that could drive (reversed) gender gaps in educational outcomes: social norms and gender stereotypes (at home and within the schooling system), relatively low public spending on education at nearly all levels, marginalization and discrimination, teenage pregnancy, and school-based violence. We generate evidence on intergenerational educational persistence, which affects girls more. Moreover, we identify a lack of systematic evidence on what works best to close these gaps, and several important data limitations, such as a lack of indicators that clearly identify Roma children. Lastly, we identify ten high-level policy areas and recommend tailored policy interventions to address gender inequalities in education in Romania. 3 Table of Contents Summary ................................................................................................................................................. 3 Table of Contents .................................................................................................................................... 4 Figures ..................................................................................................................................................... 5 Boxes ....................................................................................................................................................... 6 Tables ...................................................................................................................................................... 6 1. Gender Inequalities in Education in Romania................................................................................. 7 2. Gender Equality in Education According to Gender Indices ......................................................... 13 3. Gender Inequalities in Educational Outcomes ............................................................................. 15 3.1. Gender Inequalities in Access and Enrollment Rates ........................................................... 15 Early Childhood and Preprimary Education .................................................................................. 15 Primary Education ......................................................................................................................... 17 Secondary Education..................................................................................................................... 18 Tertiary Education ......................................................................................................................... 19 Adult Learning ............................................................................................................................... 21 3.2. Educational Attainment Differs across Groups. .................................................................... 21 3.3. Gender Segregation and Drops in Educational Performance ............................................... 23 3.4. Additional Educational Indicators Confirm Reversed Gender Gaps ..................................... 24 4. Potential Drivers of Educational Gender Inequalities................................................................... 26 4.1. Contextual Supply-Side Factors ............................................................................................ 26 Government Expenditure on Education Is Relatively Low ............................................................ 26 Available Infrastructure Is Decreasing .......................................................................................... 27 Deficiencies in the Quality of Education ....................................................................................... 30 Constraints around Affordability of Childcare and Education ...................................................... 31 4.2. Other Contextual Factors ...................................................................................................... 33 Gendered Social Norms Could Restrict Boys’ and Girls’ Education .............................................. 33 Marginalization and Discrimination Play a Role ........................................................................... 37 School-Based Violence Could Push Children out of School .......................................................... 38 4.3. Individual/Demand-Side Factors........................................................................................... 38 Poverty and Child Labor Could Restrict the Demand for Education ............................................. 39 Evidence of Intergenerational Education Attainment and the Family Context ............................ 40 4.4. High Levels of Teenage Pregnancy and School Dropout....................................................... 40 Sexuality Education ....................................................................................................................... 41 Abortion Rights ............................................................................................................................. 41 Access to Birth Control.................................................................................................................. 41 5. Current Political Landscape........................................................................................................... 42 4 6. Policy Recommendations .............................................................................................................. 43 7. Conclusion ..................................................................................................................................... 52 References ............................................................................................................................................ 53 Appendix 1. Detailed Table on Diagnostic, Constraints, Interventions, and Recommendations ......... 61 Appendix 2. Additional Regression Tables ............................................................................................ 62 Figures Figure 1. Knowledge dimension index points of EIGE Index 2022, Romania vs. EU member states.... 13 Figure 2. HCI gap, girls vs. boys, selected EU and non-EU countries in ECA, 2020 ............................... 14 Figure 3. Reversed gender gaps in Romania according to selected measures ..................................... 14 Figure 4. Net ECE enrollment rates of zero- to two-year- olds, 2020 ................................................... 16 Figure 5. Gross preprimary education enrollment rates by gender, 2013–19 ..................................... 16 Figure 6. Net preprimary education enrollment rates of three- to six-year-old children by gender and area, 2020 ............................................................................................................................................. 17 Figure 7. Net preprimary education enrollment rates of three- to six-year-old children by gender and income quintiles, 2020 .......................................................................................................................... 17 Figure 8. Gross primary education enrollment rates by gender, 2013–19........................................... 18 Figure 9. Net primary education enrollment rates, 6–11 years old, by gender, rural and urban areas, and income group, 2020 ....................................................................................................................... 18 Figure 10. Gross secondary education enrollment rates by gender, 2013–19..................................... 18 Figure 11. Net secondary education enrollment rates, 16- to 18-years old, by gender, rural and urban areas, and income group, 2020 ............................................................................................................ 18 Figure 12. Gross tertiary education enrollment rates by gender, 2013–19 ......................................... 20 Figure 13. Net tertiary education enrollment rates by gender, 2020 .................................................. 20 Figure 14. Female share of graduates in STEM-related fields, Romania vs. selected Europe and Central Asia countries, 2016–17 ........................................................................................................... 20 Figure 15. Female share of graduates in Agriculture, Forestry, Fisheries, and Veterinary, Romania vs. selected Europe and Central Asia countries, 2016–17 ......................................................................... 20 Figure 16. Highest level of educational attainment by gender, 2019................................................... 21 Figure 17. Highest level of educational attainment by gender and rural or urban area, 2019 ............ 21 Figure 18. Highest level of educational attainment by gender and income quintiles, 2020 ................ 22 Figure 19. Highest educational attainment by gender and age groups, 2019...................................... 23 Figure 20. PISA test scores by gender, 2015 vs. 2018 ........................................................................... 24 Figure 21. Learning poverty rate by gender, 2011 (%) ......................................................................... 25 Figure 22. Learning-adjusted years of schooling, Romania vs. selected Europe and Central Asia countries, 2020 ..................................................................................................................................... 25 Figure 23. Early-school-leaving rate, 18–24 years of age, by gender and rural and urban areas, 2020 .............................................................................................................................................................. 25 Figure 24. Early-school-leaving rate, 18–24 years of age, by gender and income group, 2020 ........... 25 Figure 25. Government expenditure per primary student (% of GDP per capita), Romania vs. EU average, 2012–16.................................................................................................................................. 27 Figure 26. Government expenditure per secondary student (% of GDP per capita), Romania vs. EU average, 2012–16.................................................................................................................................. 27 Figure 27. Available creches and nurseries by rural and urban areas, 2014–21 .................................. 28 Figure 28. Available kindergartens by rural and urban areas, 1996–2021 ........................................... 28 5 Figure 29. Available primary and lower secondary schools by rural and urban areas, 1996–2021 ..... 29 Figure 30. Available primary and lower secondary schools by public or private ownership, 1996–2021 .............................................................................................................................................................. 29 Figure 31. Childcare usage by type for households with at least one child below six years old by rural and urban areas, 2020 .......................................................................................................................... 30 Figure 32. Childcare usage by type for households with at least one child below six years old by income quintiles, 2020 .......................................................................................................................... 30 Figure 33. Primary education pupil-teacher ratio, 2012–17................................................................. 31 Figure 34. Secondary education pupil-teacher ratio, 2012–17 ............................................................ 31 Figure 35. Available kindergartens by type of ownership, 2014–21 ................................................... 32 Figure 36. Available creches and nurseries by type of ownership, 2014–21 ....................................... 32 Figure 37. Educational expenditure by households with at least one child below six years old by category and gender, 2019 ................................................................................................................... 33 Figure 38. Educational expenditure by households with at least one child between 6 and 11 years old by category and gender, 2019 .............................................................................................................. 33 Figure 39. Distribution of answers to the statement “University is more important for a boy than for a girl” by gender, 2018 .......................................................................................................................... 34 Figure 40. Distribution of answers to the statement “University is more important for a boy than for a girl” by income level, 2018 ................................................................................................................. 34 Figure 41. Distribution of responses to the statement “Men should have more rights to a job when jobs are scarce” by rural and urban areas, 2018 .................................................................................. 35 Figure 42. Distribution of responses to the statement “It is a problem if women have more income than their husbands” by rural and urban areas, 2018 .......................................................................... 35 Figure 43. Distribution of responses to the statement “University is more important for a boy than for a girl” by education level, 2018 ....................................................................................................... 36 Figure 44. Distribution of responses to the statement “Men should have more rights to a job when jobs are scarce” by education level, 2018 ............................................................................................ 36 Figure 45. Abortion rates per 1,000 women of reproductive age, Romania vs. the EU, 2013–20)...... 42 Figure 46. Share of women of reproductive age using contraception, any vs. any modern method (%) .............................................................................................................................................................. 42 Boxes Box 1. Main findings on educational gender inequalities from the diagnostic that guided this background note ..................................................................................................................................... 9 Box 2. Summary of diagnostic around factors that potentially drive gender inequalities in education .............................................................................................................................................................. 10 Box 3. Framework for the study of the drivers of educational gender inequalities ............................. 26 Tables Table 1. Summary of main findings and policy recommendations....................................................... 11 Table 2. Individual and demand-side determinants of the probability of attending (pre)school, 2010– 20 .......................................................................................................................................................... 39 Table 3. High-level policy recommendations and potential entry points of detailed interventions .... 43 Table A.1. Diagnostic and Recommendations ...................................................................................... 61 Table A.2. Results with interaction term B40 and gender .................................................................... 62 Table A.3. Intergenerational persistence of education by gender ....................................................... 63 6 1. Gender Inequalities in Education in Romania Romania ranks the second lowest among the European Union member states on the Knowledge dimension of the EIGE Index 2022, pointing to important shortfalls in gender equality in Romania in the education subdomain. Romania only scores 52.2 index points—the second-lowest score—on the Knowledge dimension of the EIGE Index, which measures attainment in tertiary education, participation in formal or nonformal education and training, and segregation, measured as the share of tertiary students in the fields of Education, Health and Welfare, Humanities, and Art.1 This value is 10.3 index points below the EU average and 22.4 index points below the best-performing country, Sweden. Compared to the EIGE 2013, Romania improved by 5 index points, but lost ground (saw a decline of 0.6 index points) when compared to the EIGE 2021. Addressing gender gaps in the educational sector is crucial for multiple reasons; even though many people still associate the educational “gender gap” with girls being left behind, gender parity also means identifying and addressing when boys fall behind. There are several reasons why educational gender parity is important. First, according to UNICEF (2023), educated girls are less likely to marry young and they live healthier and more-productive lives and create more-inclusive and resilient societies. Given that more-educated girls also have increased earnings (UNICEF 2023), educating girls is also beneficial from an economic perspective. At the same time, addressing the underachievement of boys and men is equally important in an agenda that targets gender equality, as gender equality is about closing both positive and reversed gender gaps. Moreover, educating boys and men has significant positive effects on gender equality. Previous studies show that education is associated with men having attitudes that are more favorable toward gender equality (Levtov et al. 2014). However, the role of boys and men in achieving gender equality has traditionally been neglected (Farré 2012). In this background study, we first analyze the gender dimensions of the educational sector in Romania and investigate multiple vulnerabilities, as the bigger picture obscures considerable disparities across urban/rural areas and income levels. We analyze potential gender gaps in several dimensions, including preprimary, primary, secondary, and tertiary school enrollment; educational attainment; and performance. Previous research shows significant variation in educational gaps across regions and social categories in Romania (Cismaru and Corbu 2019). Therefore, when possible, we also investigate the multidimensional vulnerabilities of female and male groups by intersecting our analysis by income groups and rural and urban populations. In addition, we analyze in detail access to childcare, given its importance for female labor force participation (Loshkin and Fong 2006; Robayo- Abril and Rude forthcoming). We next investigate potential drivers behind the observed (reversed) gender gaps in educational outcomes. To this end, we follow frameworks from the previous literature. According to research by Rumberger (2001), the explanations for student dropout adopt one of the following two approaches: individual factors and contextual factors. The former pertain to characteristics such as demographics, attitudes, and educational backgrounds. The latter, on the other hand, include elements like students’ families, schools, communities, and peers. Overall, the study of the drivers behind gender inequalities in educational outcomes is complex, multifaceted, and situated (Aikman and Rao 2012). It is often difficult to identify linear barriers and make direct connections between potential barriers and observed gaps in educational outcomes. Based on the literature, we look at several contextual factors related to supply-side constraints (available funding, infrastructure, and quality), other contextual 1 EIGE, 2022 Gender Equality Index, Romania | 2022 | Gender Equality Index | European Institute for Gender Equality (europa.eu). 7 factors (gender stereotypes, social norms, violence), and individual/demand-side constraints (awareness of the importance of education, parental support, poverty). We put a special focus on potential drivers behind the high teenage pregnancy rate, because it could drive girls and young women out of education. In addition, we shed light on recent developments in sexuality education, contraceptive prevalence rates, abortion rights, and the potential connection with the adolescent pregnancy rate in Romania, which is the highest in the European Union (EU). Studies conducted in Romania have investigated the reasons behind the high rate of teenage pregnancy and their impact on education. Radu et al. (2022) and Diaconescu et al. (2015) both found that teen pregnancies frequently resulted in students’ dropping out of school. Additionally, Iorga et al. (2021) discovered that 7 out of 10 mothers of girls who experienced teenage pregnancies identified school dropout as the most prevalent consequence, followed by psychological trauma. We generate descriptive evidence based on several data sources and a review of the existing literature. First, we rely on administrative data from the Ministry of Education to get a better understanding of the overall enrollment rates and infrastructure in place in the educational sector in Romania. However, while administrative data are often available by gender, such data are not available by gender and further disaggregated (for example, by gender-income groups or gender-area groups). From an equity perspective, it is crucial to understand the potential gaps between different groups of the child population in Romania. Therefore, we rely on household surveys to shed light on differences between children across income groups and rural and urban areas. We mainly use data from the EU-SILC 2020, the latest available survey round at the time we conducted this study. Importantly, these data are subject to limitations when it comes to measuring the outcomes of children (for a detailed overview of these limitations, see Greulich and Dasré 2018). The estimates presented in this report should therefore be taken with some caution and might deviate from estimates based on administrative data. We rely on several other data sources to generate a descriptive picture of gender gaps in the educational sector in Romania, namely the World Bank’s World Development Indicators (WDI) database and the Human Capital Index (HCI). We also use data from the Gender Data Portal of the World Bank, Eurostat, the World Contraceptive Use 2022 data set constructed by the Population Division of the United Nations, and PISA test scores published by the OECD. Moreover, we create estimates of household spending on education by relying on data from the Household Budget Survey conducted by the National Institute of Statistics (INS). We also rely on findings from studies on gender equality in education in Romania conducted to date. We find that there are gender gaps in enrollment rates, but also some reversed gender gaps, depending on the population group and educational level. Our analysis shows that zero- to two-year- old girls lag in net enrollment rates in rural and urban areas and in the high-income group. For three- to five-year-olds, net enrollment rates are lower for girls, independent of area or income group. However, boys fall behind girls in enrollment rates in primary education. When it comes to secondary and tertiary enrollment rates, girls in urban areas and high-income groups lag behind boys, while there are reversed gender gaps in rural and low-income groups. This evidence is in line with data on dropout rates. Box 1 summarizes the main evidence on gender inequalities in educational outcomes. In addition, we document gender segregation in terms of the field studied and some negative trends in educational outcomes, especially among boys. We find evidence of girls’ underperforming on evaluations of STEM- and ICT-related skills and of gender segregation in the field of study. Nevertheless, Table 1 shows that there are reversed gender gaps in several educational outcome variables that are particularly associated with children in rural areas and low-income groups. Moreover, boys lag behind girls in several of the performance indicators and some negative trends in educational outcomes, such as secondary enrollment rates and school performance, are especially 8 pronounced for boys. These trends are worrisome, given that more-educated men and women are less likely to believe in harmful gender norms. Box 1. Main findings on educational gender inequalities from the diagnostic that guided this background note • Very low net enrollment rates of zero- to two-year-old children, independent of gender, but even lower among girls2. • Slight gender gaps in net enrollment rates of three- to six-year-old children. • Reversed gender gaps in net primary enrollment rates. • Reversed gender gaps in secondary and tertiary enrollment rates for rural and high-income groups but positive gender gaps for urban and low-income groups. • Negative trends in secondary enrollment rates and school performance over time. • Net enrollment rates are almost always lower in rural areas and for low-income groups. Men and women in urban areas/high-income groups are more educated than those in rural areas/low-income groups. • Gross enrollment rates are significantly below the EU average at all education levels, independent of gender. • Gender segregation exists in terms of field of study; girls perform better (worse) than boys on reading (math) tests, and girls score lower on tests of digital skills. • Boys underperform girls in other educational indicators (for example, learning poverty). Source: Elaboration of authors based on the findings of this assessment We identify several potential supply-side drivers behind the (reversed) gender gaps: lack of infrastructure, public childcare, public spending, and quality. We show that supply-side factors might play a role in driving gender inequalities. We document a significant decrease in the number of (early) childhood education centers in rural and urban areas in the last decade and it is not clear what has driven this decline. Furthermore, we find that there is a limited availability of public childcare in Romania, with most families relying on the support of other family members. Second, government expenditure on education is significantly below the EU average, and there are some negative trends. The relatively low quality the educational system could also play a role. These factors could partly explain the recent decline in school performance and secondary enrollment rates and also the low performance of Romanian children in many educational indicators in comparison to the averages for children in EU member states. There is evidence of additional contextual factors: gendered social norms, school-based violence, and marginalization. We demonstrate that both men and women see themselves confronted with gender norms, which could drive some of the (reversed) gender gaps. These social norms are manifested both within the family and the school system. High levels of school-based violence and bullying are a problem in the educational system in Romania, a potential driver of cycles of abuse. In connection with this finding, certain groups of students witness and experience discrimination, especially those with a Roma background. In general, Roma women and girls seem to face additional barriers to education. We also find some suggestive evidence of individual, demand-side factors’ playing a role in educational outcomes. First, teenage pregnancy (the rate of which is high in Romania) causes girls and young women to drop out of school. Second, research has shown that poverty and the need to generate income can drive children, especially boys and young men, out of education. Third, parents and their children might not be aware of the value of investing in education. Finally, we generate new 2 These estimates are based on EU-SILC data and might deviate from estimates based on administrative data. 9 evidence showing that educational attainment is subject to intergenerational persistence, especially among girls. Box 2. Summary of diagnostic around factors that potentially drive gender inequalities in education Supply-side contextual factors: - Limited infrastructure, especially in rural areas, could restrict children’s access to education at all levels. - Low levels of sanitation and bad conditions in learning facilities, especially in rural areas, might drive girls out of school. - Relatively low government spending might generate a situation of scarce resources that in turn creates gender inequalities. - Educational quality might be low and drive children out of school. - Affordability constraints might contribute to larger gender gaps among the poor. Other contextual factors: - Gendered norms around the role of men and women in Romanina society might limit access to and demand for education, especially for rural and low-income groups. - Marginalization of and discrimination against certain population groups, such as the poor and Roma. - School-based violence might drive both boys and girls out of school. - Low support for parentless children. Individual/Demand-side factors: - Poverty and the need to generate household income might drive children out of school, especially boys. - Low awareness of the importance of education and low parental support could further limit access to education, especially for boys. - Evidence of intergenerational poverty, especially for girls. Teenage pregnancy: - Romania reports one of the highest teenage pregnancy rates in the world. Evidence shows that teenage pregnancy leads to school dropout, stigmatization, and trauma. - Teenage pregnancy might be related to deficiencies with regard to contraceptive usage, sexuality education, and access to abortions. Source: Elaboration of authors based on the findings of this assessment Lastly, we identify important data and knowledge gaps concerning some of the identified gaps and developments around gender equality in education. We identify several data and knowledge gaps, especially around boys’ and girls’ access to reproductive health education and services, with respect to drivers of declining educational outcomes and the educational realities of minorities, such as Roma children. Based on the diagnostic, we propose 10 high-level policy recommendations: 1. Conduct comprehensive studies to identify and eliminate the barriers affecting access to early childhood education (ECE) for children ages zero to two years. This entails examining and addressing constraints pertaining to both the demand and supply sides of ECE provision. 2. Implement gender-sensitive policies that specifically target the enhancement of access, availability, and utilization of ECE services. These policies should encompass a holistic approach that considers the unique needs and challenges faced by different genders. 3. Develop interventions that specifically target the reduction of school dropout rates and the improvement of enrollment rates, particularly among genders who are disproportionately affected. 10 4. Encourage and invest in the development of girls’ skills and interests in fields such as science, technology, engineering, and mathematics (STEM), as well as information and communication technology (ICT). Similarly, promote the acquisition of reading skills by boys and foster their engagement in academic disciplines traditionally dominated by girls. 5. Recognize gender equality as a multidimensional concept that extends beyond mere access to education. This comprehensive understanding should form the basis of policy formulation and implementation. 6. Formulate policies that address and correct specific types of reversed gender gaps, wherein certain genders may encounter disadvantages or underrepresentation within educational contexts or disciplines. 7. Allocate funding for research projects and monitoring and evaluation initiatives that prioritize gender sensitivity. This financial support will enable the undertaking of studies aimed at better understanding effective strategies and facilitating evidence-based decision-making. 8. Generate systematic information and data that facilitate the comparative analysis of the educational experiences and outcomes of Roma children in relation to their non-Roma peers by gender; generate updated data on child labor and time use of children. 9. Increase public spending on education at all levels. 10. Develop gender-responsive education sector planning (GRESP) and an operationalized gender-sensitive assessment tool. Table 1. Summary of main findings and policy recommendations High-level policy Evidence on gap Evidence on constraints recommendation - Availability: Potentially limited availability of formal public childcare centers as legal entities and increasing role of childcare provided by private institutions; negative trends in the number of available creches/nurseries and kindergartens as legal entities, in both rural and urban areas. (1) Study and address Very low net enrollment rates - Affordability constraints among the poor, demand- and supply-side in ECE for both boys and girls with gross and net childcare costs (as % of constraints relating to (zero- to two-year-olds) women’s median full-time earnings) below access to ECE for zero- to the EU average of 14% in Romania but two-year-olds higher among the poor; - Acceptability: parental workforce participation and willingness to use institutional care; the share of children ages less than three years cared for only by their parents is 76.8%, the largest in the EU Based on international evidence, when Gender gaps in net enrollment faced with limited resources families might (2) Gender-sensitive rates of three- to five-year-olds prefer to put boys to school (MEB Primary policies targeting access, for all groups Education General Directorate and UNICEF supply, and usage of ECE Turkey 2011) Negative trends for boys and A report by Apostu (2014) identifies three girls: gross enrollment rates factors behind children’s being out of school are below the EU average for and at risk of dropping out in Romania: (a) (3) Interventions that boys and girls and have fallen sociocultural demand-side factors, (b) target school dropout and over time in the case of economic demand-side factors, and (c) low enrollment rates primary and secondary supply-side barriers. We argue that each of education; school dropout these factors affects boys and girls rates are relatively high; boys’ 11 and girls’ performance on the asymmetrically, leading to (reversed) gender PISA has declined gaps (4) Address gender segregation in education: Educational segregation in the incentivize and invest in field of study and gender gaps the development of girls’ in digital literacy; girls Lack of role models and gender stereotypes skills and interest in STEM underperform boys in math and ICT and of boys' skills while boys underperform girls and interest in reading in reading and female-dominated fields of study The severity of constraints differs across Gender gaps are groups: the prevalence of social norms that (5) Understand gender multidimensional: they differ could limit boys’ and girls’ access to equality as a across income groups and education differs across groups, as do the multidimensional subject between rural and urban areas available infrastructure and economic conditions There are complex and multiple constraints: there is evidence of gender norms that could lead to children’s leaving school early; Boys underperform in several international evidence also points to the role (6) Policies that address educational outcomes, such as of labor market demands, household need reversed gender gaps learning-poverty rates to generate income, and boys’ being asymmetrically affected by school-based violence and bullying (UNESCO 2022). (7) Fund gender-sensitive Knowledge gaps: lack of No systematic knowledge of what works research projects and systematic, accessible, updated best to address the multidimensional monitoring and studies on what works best (reversed) gender gaps evaluation (M&E) initiatives Data gaps: Roma children Studies on the intersection of (8) Produce systematic perform worse in terms of marginalization, poverty, and educational information on Roma educational outcomes gender inequalities are restricted by a lack children compared to compared to non-Roma of updated, representative, systematic non-Roma children by children, but gender-sensitive information on educational outcomes of gender; generate information is missing; no Roma children, child labor, and time use of information on child labor information on child labor and children and time use of children time use of children is available Relatively low public spending Relatively low public spending in the on education (as a percent of educational sector could generate increased (9) Increase public GDP) when compared to the competition among subgroups over scarce spending on education EU average resources (10) Develop gender- responsive education Overall, gender inequalities in No gender-sensitive strategies developed or sector planning (GRESP) the educational sector in implemented to date and an operationalized Romania exist gender-sensitive assessment tool Source: Elaboration of authors based on the findings of this assessment We also provide a menu of interventions that can serve as potential entry points for tackling the identified constraints responsible for gender inequalities in the educational sector in Romania in section 6 and in table A.1. 12 2. Gender Equality in Education According to Gender Indices An important gender equality index in the EU is the EIGE Index; Romania took the second-lowest place on the Knowledge dimension of the index in 2022. The EIGE Index, the gender equality index of the European Institute for Gender Equality, measures countries’ progress toward gender equality. Each country is assigned from 0 to 100 index points across a range of dimensions, with 100 indicating that a country has fully closed all gender gaps. In the 2022 index, Romania occupied the second-lowest place in the Knowledge dimension ranking with 52.2 index points, substantially below the EU average of 62.5 index points (figure 1).3 Romania’s low performance on this indicator was mainly driven by gender gaps in educational attainment and participation and, to a lesser extent, by educational segregation. Figure 1. Knowledge dimension index points of EIGE Index 2022, Romania vs. EU member states 80.0 70.0 60.0 50.0 40.0 30.0 20.0 10.0 0.0 LU LV EU27 LT ES EE HU EL RO PL PT BG HR BE IE IT CZ CY DK MT DE SE NL FR AT FI SI SK Source: EIGE, 2022 Gender Equality Index, Romania | 2022 | Gender Equality Index | European Institute for Gender Equality (europa.eu). Note: The dark yellow and light yellow bars represent the index points of EU-27 average and Romania, respectively. Girls perform better than boys on all dimensions of the World Bank’s Human Capital Index (HCI), but in general, Romania’s HCI scores for both boys and girls are among the lowest in the EU and in the Europe and Central Asia region, even when considering GDP per capita levels. For example, a girl born in Romania today will be 61 percent as productive when she grows up as she could be if she attained complete education and enjoyed full health, whereas for a boy the figure 56 percent) (figure 2). This is explained by higher adult survival rates and learning-adjusted school years among girls. For example, the number of years of education a girl born today in Romania can expect to achieve at 18 is slightly higher than it is for boys (11.9 vs. 11.8 years). Still, after factoring in what children learn, more-significant differences appear, favoring girls (8.5 vs. 8.2 years), because girls have better learning outcomes than boys as reflected in harmonized test scores (figure 3). It is worth noting that the levels of HCI for both boys and girls are among the lowest in the EU and the whole Europe and Central Asia region, even when considering GDP per capita levels. 3 EIGE, 2022 Gender Equality Index, Romania | 2022 | Gender Equality Index | European Institute for Gender Equality (europa.eu). 13 Figure 2. HCI gap, girls vs. boys, selected EU and non-EU countries in ECA, 2020 0.12 0.10 HCI Gap (Girls -Boys) 0.08 0.06 0.05 0.04 0.02 0.00 Source: World Bank HCI, https://databank.worldbank.org/source/human-capital-index. Note: Estimates based on the World Bank HCI. A positive gap means girls are better off. Figure 3. Reversed gender gaps in Romania according to selected measures Human Capital Index 2020 0.05 Probability of Survival to Age 5 0 Expected Years of School 0.16 Harmonized Test Scores 9.67 Learning-Adjusted Years of School 0.3 Adult Survival Rate 0.1 0 1 2 3 4 5 6 7 8 9 10 Source: World Bank HCI, https://databank.worldbank.org/source/human-capital-index. Note: Estimates based on the World Bank HCI. A positive gap means girls are better off. According to the Global Gender Gap Index, Romania has nearly closed gender gaps in the educational sector. Romania is assigned a value of 100 percent on the educational dimension of the Global Gender Gap Index from 2022 (World Economic Forum 2022). However, the index only considers the following four indicators: the literacy rate and enrollment rates in primary, secondary, and tertiary education. However, important facets of human capital, hidden or not measured by the HCI or the Global Gender Gap Index, and some data show that girls are at a disadvantage compared with boys. The Education components of the HCI importantly distinguish years spent in school vs. learning. However, the HCI does not present the fields of study that boys and girls stream into and the underrepresentation of girls in STEM, which we investigate later in this report in more detail. Importantly, the favorable educational outcomes for girls do not translate into better outcomes in later stages in life, especially when considering labor market outcomes, which makes the case for 14 studying gender equality in education to a greater extent. Although girls outperform boys when acquiring their education in several dimensions, these positive margins do not translate into better labor market outcomes for women. The female-to-male labor force participation (LFP) ratio was only 66 in 2021, indicative of significant gender gaps in LFP rates.4 These gaps are much more marked than are documented in the EU in general, across which the average female-to-male LFP ratio is 81. In addition, there are marked differences by gender in the share of youth not in education, employment, or training (NEET). In 2021, the NEET was 14.7 percent for men and 21.5 percent for women.5 Furthermore, the female NEET was significantly higher than the EU average of 11.1 percent. In contrast, the wage gap is relatively low in Romania, standing at 5.8 percent (OECD 2023), one of the lowest values across OECD countries. Given these puzzling results, a more detailed analysis of gender equality in the educational sector in Romania could help to understand the disconnect. Moreover, these indexes might not capture gender gaps in the educational outcomes of subgroups, such as low- versus high-income groups. 3. Gender Inequalities in Educational Outcomes The educational system is tuition free and egalitarian and follows the model of the French education system (International Trade Administration 2023). Compulsory education in Romania includes preschool education, primary education, middle school education and high school education.6 Primary education consists of one preparatory course and four compulsory classes. In 2012, Romania introduced a year of compulsory preprimary education (grade 0). Evidence shows that this preparatory year reduced dropout rates and disparities among students (World Bank 2019). The preparatory year takes place during the ages of six to seven and at primary schools, and the Ministry of Education, therefore, considers it as part of primary and not preprimary education (K12 Academics 2023). All children who turn six by August of each year must be enrolled in preparatory courses.7 It is not possible for children to repeat classes after the preparatory or the first grade of primary education.8 Preparatory classes only take place in the morning. The academic year starts in September and ends in June (International Trade Administration 2023). We next study whether there are gender inequalities in access to each of the different educational levels in Romania by analyzing enrollment rates. 3.1. Gender Inequalities in Access and Enrollment Rates Early Childhood and Preprimary Education Romania has a well-established early childhood education (ECE) system. ECE includes all children between zero and six years old in Romania.9 The zero- to three-year-olds attend the so-called preschool level, while the three- to six-year-olds attend preschool education. The latter is divided into small, middle, and large groups and takes place in kindergartens and schools. The former takes place in nurseries, kindergartens, and daycare centers. There are private and public establishments in Romania, all of which follow the national standards and the same national curriculum. 4 World Bank, World Bank Data, https://data.worldbank.org/. 5 World Bank, World Bank Data, https://data.worldbank.org/. 6 Law 198/2023 on pre-university education, art.13. para. (1). 7 Romania, Ministry of Education, “Enrollment in Primary Education,” https://www.edu.ro/inscriere- invatamant-primar. 8 Romania, Ministry of Education, “Primary Education,” https://www.edu.ro/invatamant-primar. 9 Romania, Ministry of Education, “Preschool Education,” https://www.edu.ro/invatamant-prescolar. 15 Enrollment in ECE is very low among Romanian families, affecting both boys and girls. Romania has low net enrollment rates in ECE and care services (zero- to two-year-olds). Overall, according to administrative data, only 5.7 percent of children between zero and two years of age were enrolled in early childhood education in 2020.10 There are sizable gender differences in net enrollment rates disfavoring girls, particularly in rural areas and the upper-income group. There are gender differences in both urban and rural areas, but the gaps are particularly striking in urban areas, where, according to data from the EU-SILC 2020, only 7.5 percent of girls were enrolled in ECE services (ISCED 2011 level 0) compared to 17.0 percent of boys (figure 4). Figure 4. Net ECE enrollment rates of zero- to two-year- Figure 5. Gross preprimary education enrollment rates by olds, 2020 gender, 2013–19 Sources: Own estimates based on 2020 EU-SILC, https://ec.europa.eu/eurostat/web/microdata/european-union-statistics- on-income-and-living-conditions (figure 4), and World Bank, World Development Indicators, https://databank.worldbank.org/source/world-development-indicators (figure 5). Note: Net enrollment in ECE is calculated as the number of children ages zero to two who receive at least one hour of education per usual week in an ECE facility, expressed as a share of the population in this same age group. B40 is defined as the households located in the bottom 40 percent group in terms of income. We do not report results for T60, the top 60 percent group in terms of income, due to sample size limitations. The sample size is small, with 117 children, and accordingly estimates might be biased. The gross enrollment ratio for preprimary school is calculated by dividing the number of students enrolled in preprimary education regardless of age by the population of the age group who are officially eligible for preprimary education and multiplying by 100. The gender differences in gross preprimary enrollment are small; still, access to preprimary education remains low for boys and girls in Romania compared to the averages of access levels in EU member states generally. Overall, there are barely any gender gaps in gross preprimary enrollment rates (figure 5). Nevertheless, the levels of access to preprimary education for both male and female children are significantly lower than the EU averages (figure 5). According to administrative data, net enrollment rates are overall similar for boys and girls in the age group eligible for preprimary education, with about three out of four children enrolled.11 When looking at net preprimary enrollment rates, a slightly different picture emerges. Compared to the net preprimary enrollment rate of zero- to two-year-old children, a higher share of children ages three to six attend preprimary schools. This pattern of results is visible for both male and female children. As with zero- to two-year-old children, there are gender gaps in rural and urban areas, but 10 INS, Baze de date statistice, http://statistici.insse.ro:8077/tempo-online/#/pages/tables/insse-table. 11 INS, Baze de date statistice, http://statistici.insse.ro:8077/tempo-online/#/pages/tables/insse-table. 16 the gaps are less marked (figure 6). In general, enrollment rates are lower in rural areas than in urban areas. This might be related to underfunding, short schedules, and low availability in rural areas (K12 Academics 2023). When analyzing net enrollment rates of three- to six-year-old children by income groups, we find negligible gender gaps in the T60 (figure 7). Among the B40, however, girls report lower enrollment shares than boys. Net enrollment rates are higher for three- to six-year-old children than for zero- to two-year-old children for both genders and across all income groups. These patterns could be related to either the compulsory preprimary school year for six- to seven-year-old children or increased demand among families to put older children into ECE. Figure 6. Net preprimary education enrollment rates of Figure 7. Net preprimary education enrollment rates of three- to six-year-old children by gender and area, 2020 three- to six-year-old children by gender and income quintiles, 2020 0.7 0.65 0.62 0.7 0.63 0.64 0.6 0.6 0.48 0.52 0.5 0.45 0.48 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.1 0.0 0.0 rural urban b40 t60 female male female male Source: Own estimates based on 2020 EU-SILC, https://ec.europa.eu/eurostat/web/microdata/european-union-statistics- on-income-and-living-conditions. Note: Net enrollment in ECE is calculated as the number of children ages three to six who receive at least one hour of education per usual week in an ECE facility, expressed as a share of the total population of this age group. The small sample size (257 children) means estimates might be biased. We do not consider six-year-old children participating in compulsory schooling. Primary Education There are gender gaps in primary gross enrollment rates disfavoring females but reversed gender gaps in net enrollment rates. Figure 8 presents gross enrollment rates of male and female children of primary school age from 2013 to 2019. Like the trends observed for preprimary gross enrollment rates, Romania lags behind the EU on this indicator. In addition, gross enrollment rates are lower for female than male students. A different picture emerges, however, when analyzing net enrollment rates. We find reversed gender gaps in this case across both rural and urban areas (figure 9). The reversed gender gaps in primary net enrollment rates are mainly driven by the B40. The differences between the gross enrollment rates and the net enrollment rates show the incidence of overage and underage enrollment, as the net enrollment rates exclude overage and underage students and thus more accurately captures the system’s coverage and internal efficiency. 17 Figure 8. Gross primary education enrollment rates by Figure 9. Net primary education enrollment rates, 6–11 gender, 2013–19 years old, by gender, rural and urban areas, and income group, 2020 Sources: World Bank, World Bank Data, https://data.worldbank.org/indicator/SE.PRM.ENRR, and own estimates based on 2020 EU-SILC, https://ec.europa.eu/eurostat/web/microdata/european-union-statistics-on-income-and-living-conditions. Note: Gross enrollment ratio is the ratio of total enrollment, regardless of age, to the population of the age group that officially corresponds to the level of education shown. Net enrollment is calculated as the number of children ages 6–11 who receive at least one hour of education per usual week in a compulsory education facility, expressed as a share of the total population of this age group. The small sample size of 611 children means the estimates might be biased. Secondary Education Romania achieved gender parity in secondary enrollment as of 2019, though levels are extremely low compared to the EU and have decreased over time. Although secondary education is compulsory in Romania for grades five to eight, enrollment rates for male and female students are below the EU average and have decreased over time. Figure 10 depicts gross enrollment rates for male and female students in secondary education in Romania and the EU from 2013 to 2019. As can be seen in the figure, the gender gaps are negligible, both in Romania and the EU on average. In a worrisome development, secondary enrollment rates have decreased over time in Romania, widening the gap between the rates in the country and the average rates across the EU. Figure 10. Gross secondary education enrollment rates by Figure 11. Net secondary education enrollment rates, gender, 2013–19 16- to 18-years old, by gender, rural and urban areas, and income group, 2020 Sources: World Bank, World Bank Data, https://data.worldbank.org/ (figure 10) and own estimates based on 2020 EU-SILC, https://ec.europa.eu/eurostat/web/microdata/european-union-statistics-on-income-and-living-conditions (figure 11). Note: We restrict the sample to this age group, given that data for the 13-15-year-olds is not included in the EU-SILC. 18 For 2020, we find reversed gender gaps in net secondary enrollment rates of children between 16 and 18 years old for the rural and high-income groups, but positive gender gaps for those living in urban areas or belonging to the B40. After the compulsory first phase of secondary education (grades five to eight), children in Romania have the possibility to attend high school.12 During high school, children can specialize and develop key skills related to their preferred field and specialization.13 Though we find reversed gender gaps in the net enrollment rates of children between 16 and 18 years old in rural areas (figure 11) as of 2020, the gaps were negligible. In urban areas, however, a different picture emerges. As can be seen in figure 11, net enrollment rates were slightly higher for boys than for girls in this age group. Moreover, patterns differed across income groups. While boys reported higher net enrollment rates than girls in the B40, the reverse was true of the T60 group. Tertiary Education Among older cohorts, there are reversed gender gaps in tertiary, with gross enrollment rates favoring girls; however, Romania lags behind the EU average. Figure 12 presents gross enrollment rates for female and male students in tertiary education between 2013 and 2019 in Romania and the EU. The graph reveals that, while there were reversed gender gaps both in the EU and Romania, male and female students in Romania lagged in gross enrollment rates. Still, the gross tertiary enrollment rate increased over time. The gender ratio (male to female) in students enrolled in tertiary education is close to one, which indicates that gender gaps were nearly closed.14 However, a closer look reveals that the gender ratio was below one for the lowest three income quintiles and above one for the two highest income quintiles. For the period 2013–19, we find reversed gender gaps in net tertiary enrollment rates for the rural and high-income groups, but positive gender gaps for those living in urban areas or belonging to the B40. Net enrollment rates in tertiary education were higher for girls than boys in rural areas from 2013 to 2019. In contrast, boys’ net enrollment rates were higher than the net enrollment rates of girls in urban areas in the same period (figure 13). For both girls and boys, net enrollment rates were significantly higher in urban areas. In addition, as of 2020 there were marked differences in net enrollment rates across income groups: the B40 reported positive gender gaps, while the T60 reported reversed gender gaps.15 12 Romania, Ministry of Education, “High School,” https://www.edu.ro/invatamant-liceal. 13 Romania, Ministry of Education, “High School,” https://www.edu.ro/invatamant-liceal. 14 The 2020 EU-SILC, https://ec.europa.eu/eurostat/web/microdata/european-union-statistics-on-income-and- living-conditions. 15 The 2020 EU-SILC, https://ec.europa.eu/eurostat/web/microdata/european-union-statistics-on-income-and- living-conditions. 19 Figure 12. Gross tertiary education enrollment rates by Figure 13. Net tertiary education enrollment rates by gender, 2013–19 gender, 2020 Sources: World Bank, World Bank Data, https://data.worldbank.org/ and Own estimates based on 2020 EU-SILC, https://ec.europa.eu/eurostat/web/microdata/european-union-statistics-on-income-and-living-conditions Note: Tertiary education includes short-cycle, bachelor’s, master’s, and doctoral programs. There is significant gender segregation in terms of the field studied. Just 4 out of 10 graduates in STEM-related fields in Romania were female in 2016–17, so that the country’s female STEM graduation rate was comparable to the rates in other countries in the region (figure 14).16 Another group of fields in which women are underrepresented in terms of graduation rate is Agriculture, Forestry, Fisheries, and Veterinary. As with STEM-related fields, 4 out of 10 graduates in these fields were female in 2016–17; however, here Romania’s graduation rate was considerably below the rates in other countries in the region (figure 15). On the other hand, the share of female graduates was significantly higher in fields that are traditionally considered to be more feminine, such as education (91 percent), health and welfare (72 percent), and social sciences (75 percent), indicating that as of 2016 there was considerable gender segregation in terms of the fields studied by women and men. Given that STEM jobs are often higher paying (Fry, Kennedy, and Funk 2021), these patterns are an important barrier to closing the gender pay gap. Figure 14. Female share of graduates in STEM-related fields, Figure 15. Female share of graduates in Agriculture, Romania vs. selected Europe and Central Asia countries, Forestry, Fisheries, and Veterinary, Romania vs. 2016–17 selected Europe and Central Asia countries, 2016–17 43.4 60.0 56.3 56.8 41.2 50.0 46.2 40.7 38.9 40.0 36.9 30.0 20.0 10.0 0.0 Bulgaria Croatia Poland Romania Bulgaria Croatia Poland Romania Female share of graduates from Science, Female share of graduates in Agriculture, Technology, Engineering and Mathematics (STEM) Forestry, Fisheries and Veterinary programmes, programmes, tertiary (%) tertiary (%) Source: World Bank, World Bank Gender Data Portal, https://genderdata.worldbank.org/indicators/. 16 World Bank, World Bank Gender Data Portal, https://genderdata.worldbank.org/indicators/. 20 Adult Learning Adult learning does not seem to play a significant role in Romania, independent of a person’s gender. Less than 1 percent of adults between 24 and 64 years old are enrolled in some form of education.17 Most of the 24- to 64-year-olds who are enrolled participate in short-cycle tertiary education and postsecondary nontertiary education. The share is slightly larger for women than for men (1.0 versus 0.006 percent), but female net enrollment rates in this age group are entirely driven by women living in rural areas. There are no significant differences in adult enrollment rates across income quintiles, neither for men nor women.18 3.2. Educational Attainment Differs across Groups. Educational attainment is similar for adult (15+) women and men in Romania. As shown in figure 16, overall the distribution of educational levels was similar for men and women in Romania as of 2019. While a lower share of men than women had no schooling or only primary education, women were slightly more likely to have a tertiary education degree. More than 7 out of 10 men and women had completed secondary education. Lastly, having a Ph.D. degree was very unlikely for both Romanian men and women. Figure 16. Highest level of educational attainment by Figure 17. Highest level of educational attainment by gender gender, 2019 and rural or urban area, 2019 100% PhD 90% 80% 70% Tertiary educ. 60% 50% Post-secondary, non- 40% tertiary 30% 20% 10% Secondary educ. 0% rural rural urban urban Primary educ. male female male female No schooling PhD Tertiary educ. 0 0.2 0.4 0.6 0.8 1 Post-secondary, non-tertiary Secondary educ. female male Primary educ. No schooling Source: Own estimates based on EU-LSF 2019. Note: Educational attainment is the highest ISCED education level attained at the moment the survey was conducted. Both men and women are more educated in urban areas than in rural ones, and in both areas, women are slightly less educated than men. Figure 17 depicts the educational attainment of men and women above the age of 15 years in rural and urban areas as of 2019. The figure shows that a higher share of men and women in rural areas reported having only completed primary education or not having any school degree. In addition, men in rural areas reported being slightly more educated than women. While the shares of men and women with tertiary educational degrees were close in both 17 The 2020 EU-SILC, https://ec.europa.eu/eurostat/web/microdata/european-union-statistics-on-income-and- living-conditions. 18 The 2020 EU-SILC, https://ec.europa.eu/eurostat/web/microdata/european-union-statistics-on-income-and- living-conditions. 21 rural and urban areas, there are some significant differences in the lower educational groups. For example, only 1.5 percent of men in rural areas reported no schooling, whereas 2.3 percent of women did. Similarly, 8.3 percent of men in rural areas reported having only completed primary education, but 15.4 percent women reported this. The differences were less marked in urban areas. While differences in the share of people with tertiary education are marked across income quintiles for both men and women, women report a larger share in nearly all income quintiles. While the share of women above 15 years old who have a tertiary education degree is only 1.9 percent in the lowest income quintile, more than 4 out of 10 women in the highest income quintile have completed tertiary education (figure 18). These differences demonstrate that educational attainment across income quintiles is marked for women. Similar observations arise for men. Less than 1 percent of men in the lowest income quintile have completed tertiary education, compared to 38.9 percent of men in the highest income quintile. A higher share of women than men report having completed tertiary education among all income quintiles but the second highest. In the case of secondary education, men report higher shares than women across the board. Lastly, the rate of those without schooling is similar among men and women for all income quintiles, while the share of those who only completed primary education is higher for women than men, independently of the income quintile. Figure 18. Highest level of educational attainment by gender and income quintiles, 2020 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 female male female male female male female male female male 1 2 3 4 5 PhD completed Tertiary educ. Completed Other degree Secondary educ. completed Primary educ. completed No school degree Source: Own estimates based on 2020 EU-SILC, https://ec.europa.eu/eurostat/web/microdata/european-union-statistics- on-income-and-living-conditions. Note: The income quintiles are presented along the bottom of the figure: 1 is the lowest and 5 the highest. We rely on data from the EU-SILC here, given that data on household income are not available in the EU-LSF. Younger age groups are more educated than older ones and differences are more marked for women. Figure 19 plots the educational attainment distribution for men and women across age groups, from which it can be seen that younger cohorts are more educated than older cohorts. The share of people reporting tertiary education degrees decreases with age. In line with these results, a higher share of older cohorts reported only having completed primary education or not having any school degrees. These differences are more visible for women than for men. For example, while 3 out of 10 women between 25 and 34 years of age reported having a tertiary education degree, this was only true of 4.6 percent of women above 64. Similarly, the share of women who reported only having completed primary education was 3.1 percent in the age group 25–34, but was 26.4 percent among 22 those older than 64. These results could hint toward an improvement in human capital accumulation among younger cohorts, especially among women. Figure 19. Highest educational attainment by gender and age groups, 2019 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 16-24 16-24 25-34 25-34 35-44 35-44 45-54 45-54 55-64 55-64 >64 >64 male female male female male female male female male female male female PhD Tertiary educ. Post-secondary, non-tertiary Secondary educ. Primary educ. No schooling Source: Own estimates based on EU-LSF 2019. 3.3. Gender Segregation and Drops in Educational Performance Although access to affordable compulsory education at the primary and secondary levels is fundamental to Romanian children’s opportunities, the quality of instruction is a critical determinant of how much students learn. Gender differences in test scores have been widely used as indicators of gender inequality at early stages of life (Goldin 1994; Hausmann, Tyson, and Zahidi 2008). The quality of education is, therefore, another crucial factor that could both drive gender gaps in the educational sector and also be a symptom of gender inequalities in educational attainment. Recent PISA evidence reveals that, on average, Romanian female students underperform (outperform) males in math (reading) test scores. Looking at 2018 PISA math scores, girls performed slightly worse than boys, with the former having an average score of 427 and the latter 432. These differences are not statistically significant, however. In the case of PISA reading scores, boys performed significantly worse than girls, with an average score of 411 compared to an average of 445 for girls. These gaps might be related to boys’ enjoying reading less than girls.19 The difference was negligible in the case of PISA science scores (averages of 426 for boys versus 425 for girls). Performance decreased between 2015 and 2018 among boys and girls, with the drop larger for girls in the case of science and math scores and larger for boys in the case of reading scores. Compared to PISA reading scores in 2015, in 2018 the reversed gender gap in reading capacities worsened (figure 20). While girls slightly improved their performance (from 442 to 445), boys experienced declining scores (425 to 411). There was no gender gap in PISA math scores in 2015, with both male and female students scoring 444. Thus, while both boys and girls performed worse in 2018 than in 2015, girls’ performance dropped by more. The reversed gender gap in the PISA science score decreased between 2015 and 2018, which again was driven by performance drops, especially among girls: In 2015, the 19 On the PISA 2018 index “enjoyment of reading,” boys scored an average of –0.25 points compared to 0.47 points scored by girls. 23 average score for girls was 438 and for boys it was 432, compared to 426 and 425 points, respectively, in 2018. The performance drops among girls in both math and science are worrisome, given that girls were already underrepresented in STEM-related fields. Figure 20. PISA test scores by gender, 2015 vs. 2018 444 444 445 442 438 432 432 427 425 426 425 411 2015 2018 2015 2018 2015 2018 PISA Math Score PISA Reading Score PISA Science Score Female Male Source: OECD, PISA database, https://www.oecd.org/pisa/data/. Digital skills are crucial for being able to compete in the global labor market nowadays, but Romanian students on average report having only a minimum level of functional skills, with girls reporting an even lower level of skill. Overall, Romanian students report low levels of digital skills, a worrisome finding given that digital skills are crucial to compete in the labor market nowadays (Romania Journal 2022a). Only one-fourth of all students report functional digital skills and boys are 5 percent more digitally literate than girls (Romania Journal 2022a). Given that these gaps could create gender gaps in the labor market, they should be addressed early on. 3.4. Additional Educational Indicators Confirm Reversed Gender Gaps Though we have largely presented evidence on educational gender gaps to this point, reversed gender gaps do become visible when we look at other educational indicators, such as the learning poverty rate and the learning-adjusted years of schooling. Figure 21 presents the share of children affected by learning poverty, an indicator of educational deprivation, by gender in 2011, the latest available data point. There is a clear gap between male and female students. Similarly, figure 22 demonstrates reversed gender gaps in the learning-adjusted years of schooling as of 2020. While the differences are small, both male and female students lag behind the selected countries in the region on this indicator. 24 Figure 21. Learning poverty rate by gender, 2011 (%) Figure 22. Learning-adjusted years of schooling, Romania vs. selected Europe and Central Asia countries, 2020 21.6 11.5 11.2 10.7 10.2 8.8 8.6 8.5 8.2 18.2 Bulgaria Croatia Poland Romania Learning-Adjusted Years of School, Female Female Male Learning-Adjusted Years of School, Male Source: World Bank, World Bank Gender Data Portal, https://genderdata.worldbank.org/indicators/. Note: The data presented in figure 21 are from the latest available year. Learning poverty is the share of (female) children at the end-of-primary age below minimum reading proficiency adjusted by out-of-school children. Learning-adjusted years of school are calculated by multiplying the estimates of expected years of school by the ratio of the most-recent harmonized test scores to 625. Men are more likely than women to leave school early in rural areas but less likely in urban areas; differences between men and women are negligible within both the B40 and T60 income groups and marked between the income groups. The rate of early school leavers, measured as those ages 18–24 with at most lower secondary education and who were not in further education or training, is higher for men than women in rural areas (figure 23). The same pattern of results does not hold in urban areas, where a larger share of women than men leave school early. In contrast to these findings, the differences in early school leaving rates between men and women are negligible for the B40 and T60, although the rate is much higher among poorer men and women (figure 24). More than one- quarter of men and women in the B40 leave school early, compared to approximately 5 percent in the T60. Figure 23. Early-school-leaving rate, 18–24 years of age, by Figure 24. Early-school-leaving rate, 18–24 years of age, gender and rural and urban areas, 2020 by gender and income group, 2020 0.3 0.3 0.25 0.25 0.2 0.2 0.15 0.15 0.1 0.1 0.05 0.05 0 0 rural urban b40 t60 female male female male Source: Own estimates based on the 2020 EU-SILC, https://ec.europa.eu/eurostat/web/microdata/european-union- statistics-on-income-and-living-conditions. 25 Note: The share of 18- to 24-year-olds who leave school early, defined as the number of people in this age group with at most lower secondary education and who were not in further education or training over the total number of 18- to 24- year-olds. 4. Potential Drivers of Educational Gender Inequalities To investigate the causes of the (reversed) gender gaps in educational enrollment and achievement, we examine previous studies and analyze both individual and contextual factors. Rumberger’s early research (2001) provides a comprehensive overview of the theoretical and empirical investigations into why students leave school. The study identifies two main approaches: one focusing on individual factors and the other on contextual factors. Individual factors encompass demographics, attitudes, and educational backgrounds, while contextual factors include students’ families, schools, communities, and peers. Although gender is an individual factor, we propose that additional individual and contextual factors may impact boys and girls differently, contributing to the observed gender gaps. Previous studies confirm that these factors can have gendered effects.20 Based on previous studies, we look at contextual supply-side factors, additional contextual factors, and individual demand-side factors. A report by Apostu (2014) indicates that three types of factors contribute to children being out of school and at risk of dropping out in Romania: sociocultural demand-side factors, economic demand-side factors, and supply-side barriers. We build upon this approach and argue that there are contextual supply-side factors, additional contextual factors, and individual demand-side factors driving gender inequalities in education in Romania. We look at the following supply-side factors: available infrastructure, resources, educational quality, and affordability of education. In addition, we analyze additional contextual factors: gendered social norms, marginalization and discrimination, school-based violence, and lack of support for parentless children. Lastly, we analyze individual factors, such as poverty and low awareness of the importance of education. Box 3 summarizes this framework. We also dedicate a special section to the role of teenage pregnancy. Box 3. Framework for the study of the drivers of educational gender inequalities Gender inequalities in educational outcomes Contextual supply-side Additional contextual Individual demand- factors factors side factors Source: Own elaboration based on previous literature reviews by Rumberger 2001 and Apostu 2014. 4.1. Contextual Supply-Side Factors Government Expenditure on Education Is Relatively Low Government expenditure on education is below the EU average. In 2020, the government spent 8.8 percent of its overall expenditure on education, a slight drop of 0.2 percentage points compared to 20 For example, Oyvat and Onaran (2022) find that higher social infrastructure expenditure—expenditure on education, childcare, healthcare, and social care—increases female employment more than male employment. 26 2013.21 This proportion is below the EU average of 10.4 percent and is extremely low compared to high-income countries and the global average (World Bank forthcoming). According to a study by the OECD (2020), Romania is among the countries that spend the lowest amount on ECE and childcare compared to OECD and EU member countries. In 2020, Romania spent less than 0.6 percent of its GDP on early childhood education and care (OECD 2020). In addition, within this spending, expenditure on childcare is negligible. With the EU and OECD average at just above 0.8 percent, it is clear that Romania lagged behind most countries as of 2020. Overall, government expenditure per student was significantly below the EU average for primary and secondary students for 2012–16 and even dropped during that period, though there was a positive development for tertiary students. From 2012–16, the relative government expenditure in the case of primary students saw a significant decrease (from 11.1 percent in 2012 to 7.8 percent in 2016) (figure 25). Furthermore, the gap between what Romania spends for primary students and the EU average is large (12.5 percentage points). On a positive note, in the same period the share per secondary student first increased, but then dropped during 2014–16 (from 16.3 to 15.2 percent of GDP per capita) (figure 26). These spending shares were significantly below the EU average for secondary students for the same years (23.0 percent in 2015–16). However, there was an increase in the share in the case of tertiary students from 2013 to 2016 (from 19.3 to 26.1 percent), with the result that the gap between what Romania spends and the EU average was nearly closed. Nevertheless, the developments for primary and secondary students are worrisome. They could also at least partly explain some of the worrisome trends in enrollment rates and school performance. In addition, the government of Romania currently has no gender budgeting in place, though recent consultations point to efforts in that direction under the leadership of the European Commission. Figure 25. Government expenditure per primary student Figure 26. Government expenditure per secondary (% of GDP per capita), Romania vs. EU average, 2012–16 student (% of GDP per capita), Romania vs. EU average, 2012–16 25.0 22.1 21.2 20.1 20.8 20.3 30.0 20.0 24.4 23.7 23.7 23.0 23.0 25.0 15.0 11.1 20.0 16.3 16.2 10.6 15.2 9.0 8.7 12.5 14.0 10.0 7.8 15.0 5.0 10.0 5.0 0.0 2012 2013 2014 2015 2016 0.0 2012 2013 2014 2015 2016 European Union Romania European Union Romania Source: World Bank, World Bank Data, https://data.worldbank.org/. Available Infrastructure Is Decreasing Low availability of institutions might make attending school difficult and create barriers to girls’ education as well as inconvenience for working mothers. A report by UNICEF (2023) points out that mobility restrictions might influence educational participation: for example, girls might be especially affected due to safety concerns. Moreover, if ECE facilities are far away and require substantial travel to reach, mothers might be less likely to take advantage of childcare provided by these institutions. In addition, under a scenario of limited resources, the competition for access to these scarce resources 21 World Bank, World Bank Data, https://data.worldbank.org/. 27 might be greater and create gender disparities. Based on these arguments, we investigate the number of available ECE facilities over time. According to administrative data, the number of creches and nurseries stagnated in rural areas at one facility and fell in urban areas. Data published by the INS show that between 2015 and 2021, the number of creches and nurseries with legal status remained extremely low in rural areas.22 In urban areas, there was a negative development, as the number of creches and nurseries fell from 31 in 2014 to 20 in 2021 (figure 27). Importantly, there are other ECE units, but data for them are only available for 2021. According to data published by INS, there were 344 such facilities in urban areas and 11 in rural areas.23 The number of kindergartens declined in both rural and urban areas from 1996 to 2021. There are also worrisome trends in the case of kindergartens. In rural areas, the number of kindergartens decreased from more than 9,000 in 1996 to only 81 in 2021 (figure 28). In urban areas, the number was three times lower in 2021 than the number in 1996 (figure 28). Other preprimary (preschool) education units are available in addition to kindergartens, but again, data are only available for 2021. There were 7,253 such facilities in rural areas and 2,192 in urban ones. Figure 27. Available creches and nurseries by rural and Figure 28. Available kindergartens by rural and urban urban areas, 2014–21 areas, 1996–2021 40 12000 10000 30 8000 20 6000 4000 10 2000 0 0 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 2014 2015 2016 2017 2018 2019 2020 2021 Rural Urban Rural Urban Source: INS, Baze de date statistice, http://statistici.insse.ro:8077/tempo-online/#/pages/tables/insse-table. Note: The data presented in figures 27 and 28 only consist of creches, nurseries, and kindergartens with legal status. The drop in available ECE facilities is unrelated to falling birth rates or urbanization. According to data from the World Bank,24 fertility rates increased in Romania between 1996 and 2021 (from 1.3 to 1.8 births per woman). Consequently, the drop is not related to a lower number of children. Urbanization was also constant during this period, standing at 54 percent.25 Consequently, more evidence is needed to understand the developments in available infrastructure in more detail. Similarly, the number of primary and lower secondary schools decreased significantly in both rural and urban areas. Like the developments observed for ECE facilities, there was a steep decline in the number of available primary and lower secondary educational facilities in rural areas (figure 29). The number of these institutions in urban areas also halved during that time. These developments raise concerns about the quality and availability of education in Romania. Still, a more granular analysis could help to understand further the underlying drivers behind the decline in available institutions. 22 INS, Baze de date statistice, http://statistici.insse.ro:8077/tempo-online/#/pages/tables/insse-table. 23 INS, Baze de date statistice, http://statistici.insse.ro:8077/tempo-online/#/pages/tables/insse-table. 24 World Bank, World Bank Data, https://data.worldbank.org/. 25 World Bank, World Bank Data, https://data.worldbank.org/. 28 Some privately run primary and lower secondary schools have opened since 2012, but the number as of 2021 was still negligible (figure 30). Figure 29. Available primary and lower secondary schools Figure 30. Available primary and lower secondary schools by rural and urban areas, 1996–2021 by public or private ownership, 1996–2021 12000 16000 10000 14000 12000 8000 10000 6000 8000 4000 6000 2000 4000 0 2000 0 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 Rural Urban Private ownership Public ownership Source: INS, Baze de date statistice, http://statistici.insse.ro:8077/tempo-online/#/pages/tables/insse-table. Additional evidence points toward worse conditions in rural schools, which might explain why educational outcomes are often worse among rural boys and girls when compared to those living in urban areas. According to Hamby (2018), rural schools suffer from inadequate infrastructure, lack of running water and health permits, broken desks, and limited resources for transportation. These conditions, again, could affect girls more than boys, for example, as the former might lack menstruation-hygiene measures or face a greater risk of violence when traveling to schools. In general, sanitary conditions of educational facilities are poor in comparison to what has been observed other countries (Gyaraki 2022). These differences in the quality of infrastructure could explain some of the gaps in educational outcomes observed between rural and urban areas. Studying the availability and affordability of childcare is important, given that it is a driving factor behind female labor force participation and child development. Loshkin and Fong (2006) document an important interaction between the cost of childcare and female labor force participation. In line with this evidence, Robayo-Abril and Rude (forthcoming) show that an increase in the availability of public-provided childcare provided by the public sector leads to higher female labor force participation rates. Consequently, it is crucial to study the availability and affordability of childcare. Participation in ECE also improves child development outcomes (Burger 2015). There is evidence of the limited availability of formal public childcare centers in Romania. Kindergarten education in Romania takes place between ages three to six and is divided into three stages. Children first enter the “small group” (grupa mică) when they are three to four years old, the “middle group” (grupa mijlocie) when they are four to five years old, and the “big group” (grupa mare) when they are five to six years old. Both public and private kindergartens often have 1 to 2 teachers and groups of 10 to 15 children (K12 Academics 2023). Kindergartens in Romania offer at least one meal and might offer more depending on the schedule (K12 Academics 2023). While public kindergartens do not charge tuition fees, tuition fees in private kindergartens are relatively high (K12 Academics 2023). Generally, the number of kindergartens is low, leading to long waiting lists (K12 Academics 2023). This evidence could point toward the inability of the public sector to meet the demand for public childcare in Romania. Data from the 2020 EU-SILC show that childcare provided by other household members played a significant role in Romania, while that provided by formal daycare centers played a negligible role 29 and furthermore rural households lagged behind urban ones in all types of childcare usage. Figure 31 depicts four different forms of childcare usage in 2020 by households with at least one child below six years old for rural and urban areas: childcare provided by another household member, childcare provided by a professional, childcare provided by a childcare center or daycare center, and lastly, childcare provided in preschools. As can be seen in the figure, in general the share of households with at least one child below six years old reporting some form of participation in childcare was higher among urban than rural households, independent of the type of childcare provided. Differences were especially marked in the case of preschools. Childcare provided by childcare centers or daycare centers was negligible in both rural and urban areas. Figure 31. Childcare usage by type for households with Figure 32. Childcare usage by type for households with at at least one child below six years old by rural and urban least one child below six years old by income quintiles, areas, 2020 2020 1 HH with child (<6) and child- 0.8 care by hh member 0.6 0.4 HH with child (<6) and child- 0.2 care (by professional) 0 Q1 Q2 Q3 Q4 Q5 HH with child (<6) and child- care (center-based or day- care center) HH with child (<6) attending pre-school HH with child (<6) attending HH with child (<6) and child-care (center-based or day-care pre-school center) HH with child (<6) and child-care (by professional) 0 0.2 0.4 0.6 0.8 HH with child (<6) and child-care by hh member urban rural Source: Own estimates based on 2020 EU-SILC, https://ec.europa.eu/eurostat/web/microdata/european-union-statistics- on-income-and-living-conditions. Note: Q1 refers to the bottom quintile (poorest) and Q5 refers to the top quintile (richest). In 2020, households with at least one child below six years old in the lowest income quintile reported lower access to formal childcare than those in the highest income quintiles. A loer share of households with children under six years old in the lowest income quintile relied on formal childcare provided by preschool facilities, daycare centers and professionals, 47 percent, compared to 67 percent in the highest income quintile (figure 32). In addition, richer households seemed to substitute for childcare by employing a professional, although the share was still low (below 6 percent. Not only richer households use more formal childcare, but they also relied more on family members to provide care26. Deficiencies in the Quality of Education Some evidence points to deficiencies in the quality of education in Romania, as for example the number of pupils per teacher increased in primary education, but it decreased in preprimary and tertiary education. According to the “Best Educational Systems” Index, which measures the perception of educational systems in 78 countries based on three indicators,27 Romania ranked in the 26 These care categories are not mutually exclusive in the household survey. 27 The three indicators are “has a well-developed public education system,” “would consider attending university there,” and “provides top-quality education”; a country’s ranking depends on survey participants’ responses to questions targeting the indicators. 30 lower half of countries in 2021 (World Population Review 2023). In comparison to the previous year, Romania fell by seven places. One indicator often used to approximate educational quality is the pupil- teacher ratio. The pupil-teacher ratio in primary education did increase in Romania from 2012 to 2017, going from 17.6 to 19.3 (figure 33). Moreover, as of 2017 it was still well above the average in the EU, which stood at 13.3 pupils per teacher (figure 33). In contrast, in secondary education the ratio slightly dropped during that same period (from 12.8 pupils per teacher to 12.1), ending up at a level close to the EU average (figure 34). Consequently, the drop in gross secondary enrollment rates is unlikely related to this indicator. In tertiary education, there was also improvement in the pupil-teacher ratio in 2012-2017 (from 25 to 20 pupils per teacher). Furthermore, slightly decreasing pupil-teacher ratios suggest some positive trends around the supply of teachers in preprimary education. In concrete terms, the pupil-teacher ratio decreased from 18 to 15 between 2012 and 201728. It should be noted that the ratio was still above the EU average of 13 pupils per teacher. Given the already declining performance of both boys and girls on PISA test scores, the increase in the pupil-teacher ratio in primary education is worrisome. The fact that Romania performs worse than the EU average on this indicator and on enrollment rates could hint toward a potential connection, but more-detailed studies are needed to confirm a potential causal link. Figure 33. Primary education pupil-teacher ratio, 2012–17 Figure 34. Secondary education pupil-teacher ratio, 2012–17 25.0 13.0 12.8 12.8 18.9 19.3 19.3 20.0 17.6 12.5 12.3 13.3 13.6 13.5 13.4 13.3 12.0 12.0 12.1 12.1 12.1 15.0 13.2 11.9 12.0 11.7 11.8 11.6 10.0 11.5 5.0 11.0 0.0 10.5 2012 2013 2014 2015 2016 2017 2012 2013 2014 2015 2016 2017 European Union Romania European Union Romania Source: World Bank, World Bank Data, https://data.worldbank.org/. Note: Data are not available for 2013 and 2014 in the case of Romania in primary education. The latest available data for Romania are from 2017. Constraints around Affordability of Childcare and Education The net cost of childcare is below the EU average for Romanians as a whole, but not for the B20. According to a study by the OECD (2022), gross and net costs of childcare,29 as a proportion of women’s median full-time earnings, are below the EU average of 14 percent in Romania. This pattern of results holds for single mothers and couples. In addition, gross and net costs are equal in Romania, because the care itself is provided free of charge (OECD 2022). The cost of childcare is, however, above the EU average for the B20, which raises equity concerns. Moreover, the report finds that the participation tax rates for mothers entering full-time employment, a measure of a mother’s cost to reintegrate into the labor market due to increased costs of childcare, higher taxes, and withdrawn benefits, are higher than the EU average when childcare costs are factored in. 28 World Bank, World Bank Data, https://data.worldbank.org/. 29 The OECD net childcare cost (NCC) indicator reflects the net reduction in family budgets resulting from the use of center-based childcare. 31 Studying the availability and affordability of private preprimary education is important, because poorer households might have less access to them; around 5 percent of government-authorized kindergartens are private institutions. Studying the availability and affordability of private preprimary education is important, given that the high rates of households accessing this type of ECE could hint toward low trust in public provision of ECE or the limited availability or low quality of publicly provided ECE. These factors in turn might create long-lasting, intergenerational disadvantages for poorer households. There are currently 1,023 private authorized and accredited kindergartens in Romania and 17,415 public ones.30 These numbers coincide with estimates from the Household Budget Survey from 2019, showing that around 3 percent of households with children under six years old pay fees for preprimary education (INS 2019). In combination with the previous evidence, it appears that some households responded to the limited availability of public-sector childcare by accessing private institutions. Over time, private institutions have become more and more important in this area: in 2021 they accounted for nearly 4 out of 10 kindergartens (figure 35). Private provision of childcare also plays a significant role in ECE in Romania. In 2021, there were 13 creches and nurseries under private ownership and 8 owned by the government. It should be noted, though, that in 2016 there were 31 privately owned creches and nurseries, indicating that there was a significant decrease in the number of available creches and nurseries from 2016 to 2021 (figure 36). Figure 35. Available kindergartens by type of ownership, Figure 36. Available creches and nurseries by type of 2014–21 ownership, 2014–21 14000 50 12000 40 10000 8000 30 6000 4000 20 2000 10 0 0 2014 2015 2016 2017 2018 2019 2020 2021 Private ownership Public ownership Private ownership Public ownership INS, Baze de date statistice, http://statistici.insse.ro:8077/tempo-online/#/pages/tables/insse-table. Note: The data presented in figures 35 and 36 only consist of creches, nurseries, and kindergartens with legal status. Spending on schoolbooks and supplies accounts for the largest component of educational expenditure at the household level. High costs for education and educational supplies could potentially also hinder educational access, especially among low-income countries. To investigate this point in more detail, we analyze household expenditure on educational goods and fees using data from the 2019 Household Budget Survey conducted by the INS. We find that households with at least one child below 6 years old (and those with at least 1 child between 6 and 11 years old) report the highest absolute expenditure on schoolbooks and supplies (figures 37 and 38). Households’ spending on education fees and childcare services is negligible, confirming the limited role of private institutions in providing education in Romania. Households with children in both age groups spend only a negligible amount on educational fees or childcare services (figures 37 and 38). Private education and childcare services might therefore only play a limited role in Romania, probably due to a low share of the population’s being able to pay for these. In fact, some evidence indicates 30 Romania, Ministry of Education, “Preschool Education,” https://www.edu.ro/invatamant-prescolar. 32 that income might be an important driving factor between those accessing private and public schooling in Romania, generating additional barriers between the poor and the non-poor (Hamby 2018). These barriers are especially problematic, because public schools suffer from inadequate funding and underachievement (Hamby 2018). Consequently, the educational divide between students who attend public schools and those who attend private schools could foster intergenerational poverty. Spending on educational fees and schoolbooks plus supplies differs slightly by the child’s gender. There are slight differences by the child’s gender, with parents spending larger amounts on schoolbooks and supplies when their child is female. These patterns of results are visible for children of preprimary and primary school age. For preschool educational fees, we observe slightly higher spending in the case of boys than girls (figure 37). This evidence is in line with the evidence on enrollment rates during preprimary education, showing that rates are significantly higher for boys than girls, especially among the youngest children (zero- to two-year-olds). For primary educational fees, we observe higher absolute spending for households with at least one female child (figure 38), which is also in line with evidence on enrollment rates for this age group. Based on this evidence, income could potentially be one of the driving forces behind (reversed) gender gaps in enrollment rates. Figure 37. Educational expenditure by households with at Figure 38. Educational expenditure by households with at least one child below six years old by category and least one child between 6 and 11 years old by category and gender, 2019 gender, 2019 School books and School books and supplies supplies Child care services Child care services School accommodation School accommodation Primary education fee Primary education fee Preschool education fee Preschool education fee 0 20 40 60 80 100 0 50 100 150 HH with girl (<6) HH with boy (<6) HH with girl (6-11) HH with boy (6-11) Source: Own estimates based on INS 2019. 4.2. Other Contextual Factors Gendered Social Norms Could Restrict Boys’ and Girls’ Education Research has shown that social norms are important drivers of gender gaps in education. Research has produced empirical evidence showing that gender norms can influence the persistence of gender gaps in educational outcomes generally speaking. Both gender norms within the family and within the schooling system play a role (Breda et al. 2020; Chisamya et al. 2012; Gray and Leith 2004). Accordingly, we undertake a descriptive analyis to determine whether there are harmful gender norms present in Romanian society. Harmful gender norms could explain educational gender gaps in Romania, especially among lower- income groups. According to data from the World Values Survey (2017–2022), about a fifth of men in Romania still believed that university is more important for a boy than for a girl (figure 39). Women in 33 Romania are not exempt from these types of harmful gender norms, as 16.5 percent (strongly) agreed with this statement, only a slight decrease from the 17.1 percent of women who agreed with this statement during the previous survey wave (2010–2014). The share of men who agreed with this statement dropped more significantly, with one-fourth of men agreeing during the previous survey wave. Moreover, the share of women who disagreed increases significantly by income and educational level. For men, however, there are no clear patterns across income groups and less-marked differences across educational levels. Overall, a higher share of the low-income group than of the medium- and high-income groups tended to (strongly) agree with the statement that university is more important for boys than for girls (figure 40). A study by Brătucu et al. (2018) on gender stereotypes among the young generation in Romania concludes that they are marked. Figure 39. Distribution of answers to the statement Figure 40. Distribution of answers to the statement “University is more important for a boy than for a girl” by “University is more important for a boy than for a girl” gender, 2018 by income level, 2018 60 49.1 48.7 47.4 50 44.9 42.2 40 33.6 33.2 32.8 28.3 30 23.3 20 12.712 13.9 1211.2 8.4 8.3 10 5.5 6.7 7.3 5.1 4.1 2.7 3 0.30.9 0.8 0 Agree Agree Disagree Don't No Strongly Agree Agree Disagree Strongly Don't strongly know answer disagree strongly disagree know Men Women Low Medium High Source: World Values Survey (2017–2022). Men are equally confronted with social norms around masculinity, which could explain why they perform worse on some of the educational indicators and have lower enrollment rates in some cases. When surveyed, nearly 4 out of 10 Romanians believed that when jobs are scarce, men should have more rights to a job than women (World Values Survey 2017–2022), and the proportion of women who agreed with this statement was just slightly smaller than the proportion of men. Respondents from lower-income groups were more prone to (strongly) agree with the statement than respondents from higher-income groups. In line with these results, more than 2 out of 10 Romanians thought that there is a problem if women have more income than their husbands (World Values Survey 2017–2022). While there are barely any gender gaps in the shares of men and women who agreed with this statement, there were relatively strong patterns across income groups. These results demonstrate that men see themselves confronted with social norms in line with traditional concepts of masculinity.31 These norms might pressure them into entering the labor market early and might disincentivize them from investing in higher education and skills. People in rural areas are more prone to believe in harmful gender norms around the role of men and women in society. As shown in figure 41, a higher share of those living in rural areas (strongly) 31 Researchers have increasingly studied harmful masculinity and its relevance for gender inequality; for an introduction see Edström et al. (2015). 34 agreed with the statement that when jobs are scarce, men should have more rights to a job than women (World Values Survey 2017–2022). The same was true with regard to the statement that it is a problem if women earn more than their husbands (figure 42). On the other hand, the population in rural areas was more likely to be in agreement with harmful gender norms around girls’ education. For example, more than 2 out of 10 Romanians in rural areas (strongly) agreed with the statement that university is more important for boys than for girls, while 16.4 percent of those in urban areas believed so (World Values Survey 2017–2022). These patterns of results demonstrate that rural populations see themselves confronted with a higher degree of harmful gender norms, which could explain why (reversed) gender gaps are more persistent in some cases. Figure 41. Distribution of responses to the statement “Men Figure 42. Distribution of responses to the statement should have more rights to a job when jobs are scarce ” by “It is a problem if women have more income than their rural and urban areas, 2018 husbands” by rural and urban areas, 2018 30 40 36.4 25 30.9 22.2 23.1 35 25 20.7 25.9 19.6 19.9 30 18.5 21.8 20 16.3 17 25 19 13.4 20 16.5 14.4 14.8 15 15 8 10 10 5.7 3.6 2.8 2.4 5 0.1 0.1 5 1.6 0.4 0 0 Urban Rural Urban Rural Source: World Values Survey (2017–2022). As more-educated men are less likely to believe in harmful gender norms, the drop in educational performance and secondary enrollment rates among boys is worrisome. For example, 14.7 percent of highly educated men (strongly) agreed with the statement that university is more important for a boy than for a girl, compared to 20.2 percent of lower-educated men (figure 43). Similarly, 14.4 percent of highly educated men thought that men should have more rights to a job than women when jobs are scarce, while this was true of 43.8 percent of lower-educated men (figure 44). Consequently, high-educated men are less likely to believe in harmful gender norms. Based on this evidence, the negative trends in several of the educational outcomes studied are worrisome. 35 Figure 43. Distribution of responses to the statement Figure 44. Distribution of responses to the statement “University is more important for a boy than for a girl” by “Men should have more rights to a job when jobs are education level, 2018 scarce” by education level, 2018 60 30 50 25 40 20 30 15 20 10 5 10 0 0 Low Middle High Low Medium High Source: World Values Survey (2017–2022). Research shows that boys are already engaging in masculine behavior during school, which calls for school-based interventions to address harmful gender norms. Results from the PISA show that as of 2019 boys in Romania were more competitive than girls32 and less likely to cooperate at school.33 Verdonk, Seesing, and De Rijk (2010), who studied the interaction between the concept of masculinity and educational attainment, show that these behavioral patterns might be related to traditional concepts of masculinity. This evidence indicates that interventions targeting harmful gender norms should already be implemented during school and not only during adulthood. Implementing gender-neutral education systems is about both granting girls access to education and providing them with the education they need to thrive; eliminating gender stereotypes in education is crucial. Gender-responsive education is about granting girls access to education and creating a learning environment where they feel safe and supported (UNICEF 2023). If teaching practices are not gender responsive, they might reinforce gender stereotypes. Related to this point, many studies show that school textbooks often suffer from pro-male biases.34 Along these lines, Concordă (2018) conducts a review of the curriculum material in Romania and concludes that there is discrimination against women. There is thus an urgent need to address these biases and replace current schoolbooks with updated, bias-free textbooks that could contribute to gender equality in Romania. Evidence on the role of parents, teachers, and parent-teacher interaction in driving gender equality in Romania is lacking, but studies from other countries show the importance of all three. Parents’ gender stereotypes can be crucial to gender inequality (UNESCO 2020). Similarly, teachers’ 32 Boys in Romania scored 0.34 on the PISA competitiveness index, which ranges from -3 to 3, while girls only scored 0.27. 33 Boys in Romania scored 0.04 on the perception of cooperation at school (wle) PISA index, which ranges from -3 to 3, while girls scored 0.17. 34 Islam and Asadullah (2018), for example, demonstrate a strong gender bias in Malaysian, Indonesian, Pakistani, and Bangladeshi school textbooks. Similarly, Papadakis (2018) detects gender stereotypes in an analysis of Greek computer science textbooks and Sovič and Hus (2015) show that men are overrepresented in young learners’ books and assume more active roles, while women ass ume quiet, socially approved roles. 36 stereotypes can create critical barriers to gender equality (see, for example, Carlana 2019 and Alan, Ertac, and Mumcu 2018). On a positive note, both parents and teachers can contribute to alleviating harmful gender norms and the latter can play a crucial role in raising awareness about the topic through gender-responsive parent-teacher interactions (Warner and Barrera 2005). Role models are crucial to achieving gender equality. Several studies show the significant impact of role models on educational choices, both within (Dryler 1998) and outside the household (Wolbrecht and Campbell 2007). The literature on role models has mainly focused on the importance of idols in increasing the gender ratio in areas where women are traditionally underrepresented. More recently, scholars have started to analyze the influence of male role models in order to rethink masculinity (Connell 2003). In Romania, the lack of role models for women in STEM is one area previously discussed as important (RDCC 2022). Marginalization and Discrimination Play a Role Updated, reliable numbers on enrollment rates of Roma children are not publicly available, but previous reports estimated that enrollment rates of Roma children are significantly lower. Previous reports have estimated that enrollment rates of Roma children in early childhood services could be half of the rates reported for non-Roma children (Bennett 2012). Previous research has also suggested that Roma children might face segregation and might intentionally be placed in special schools for children with cognitive disabilities (Bennett 2012). These observations raise questions about the practical implementation of the regulatory framework from 2007 that forbids school segregation of Roma children.35 Some studies have concluded that the reform has not been implemented as intended (Gyaraki 2022). While the evidence on the causes of decreasing enrollment rates in Romania is limited, some attention has been given to explaining dropout by underprivileged groups, such as the Roma. Rotaru (2019), on the basis of qualitative evidence, points to several underlying factors driving the dropout of Roma students: antigypsyism, marginalization in classrooms, difficulties in combining motherhood and education in the case of young mothers, and discrimination more broadly speaking. In general, Roma women face multiple challenges that affect their educational outcomes, such as isolation, poverty, discrimination, and often their embracing of masculine social norms (Andrei, Tkadlecova, and Martinidis 2015). In Romania, the inclusion of Roma individuals in the education system continues to be a significant challenge, as indicated by recent evidence. A study conducted by Patache, Ghencea, and Negurita (2022) reveals that the Roma community faces multiple barriers, including lower education levels, higher rates of illiteracy, increased early school dropout rates, reduced academic achievement, and lower educational attainment. The study attributes these disparities primarily to factors such as stigmatization, poverty, limited expectations from Roma parents regarding their children’s education, a lack of role models, early marriages, and the need to support the family by taking on both household responsibilities and income-generating activities. Both Roma boys and girls face discrimination in the educational system in Romania, but the situation for Roma girls is even worse. While Roma children receive fewer years of schooling compared to their non-Roma counterparts, the situation is worse for Roma girls. Roma girls receive 5.66 years of education, on average, compared to 6.71 years in the case of Roma boys and 10.7 years for non-Roma girls (Selander and Walter 2020). Nearly 3 out of 10 Roma women have never received 35 Romania, Ministry of Education, “Access to Education for Disadvantaged Groups,” https://www.edu.ro/acces-la-educatie-pentru-grupurile-dezavantajate. 37 any form of formal education, whereas this is true of 18 percent of Roma men (Selander and Walter 2020). The factors behind these gaps are manifold and include racial discrimination, placement in special education schools, cultural traditions, and neighborhood segregation; cultural factors, such as early marriage, might also play a role, especially for girls (Selander and Walter 2020). School-Based Violence Could Push Children out of School Many children in Romania are exposed to violence, both at school and at home. Several studies show that bullying and school-based violence are persistent problems in Romania (Ciucă 2019; Gavrilescu and Merloiu 2021). In fact, Romania ranks third highest in this regard among the EU member states (Romania Journal 2019). According to data collected by Save the Children Romania (Romania-Insider 2020), 2 out of 10 children experience verbal abuse, and 3 out of 4 are bullied at school (Romania Journals 2022b). In addition, half of parents in Romania state that they punish their children if they make a mistake (Romania Journals 2022b). Other studies show that nearly 3 out of 10 students have witnessed discrimination at school, mainly based on gender, ethnicity, and socioeconomic status (Pratama 2022). Sexual harassment and abuse are also problems in Romanian schools and universities (Gyaraki 2022). The violent school environment could induce and encourage cycles of abuse and gender-based violence later in life. The high exposure of children to violence and bullying is problematic, because being subject to violence can create cycles of abuse (Plummer and Cossins 2018). In fact, boys often become abusers and start demonstrating violent behavior during school (Drob 2020). The high levels of school-based violence are also in contrast to expert recommendations about schools having the potential to be safe havens and presenting readily available settings for violence-prevention programs (Nadel et al. 1996). Boys could drop out of school because of bullying and violence (UNESCO 2022). Parentless children might be neglected; some previous evidence gathered about children without parents raises concerns. Some studies raise the issue of children without parents facing a devastating situation in Romania (for example, Mihaileanu 2018). Approximately 10,000 children enter the Romanian welfare system each year, and many stay until they are adults (Mihaileanu 2018). Adoption rates are low and parentless children are faced with a system that requires them to move out of foster care into state institutions. There are no minimum standards in place for unpaid foster families or birth relatives, and cooperation with NGOs is challenging (Mihaileanu 2018). A gender-sensitive study of parentless children could help to shed light on the different realities of boys and girls. 4.3. Individual/Demand-Side Factors We next analyze individual and demand-side factors that could limit gender inequalities in educational outcomes by analyzing the interaction of household and children’s characteristics with the probability of attending (pre)school. We analyze the potential existence of individual factors by running a simple linear regression on the probability of children younger than 16 years old attending (pre)school. To this end, we use data from the EU-SILC 2010 to 2020 to construct a variable that measures children’s school attendance as outlined in Robayo-Abril and Rude (forthcoming). We then run a simple linear probability regression on the probability of entering preschool as follows: = + + where Y is an indicator variable that takes the value of one if children attend (pre)school for at least one hour per week and F and X are individual and household characteristics, respectively. Standard errors are robust to account for potential heteroskedasticity. The regression reveals a significant gender gap in the probability of attending (pre)school. The results presented in Table 2 reveal that there is a clear and significant gender gap in the probability of 38 attending (pre)school. The coefficient in row 1 is negative and significant at the 1-percent significance level. In addition, older children are less likely to attend (pre)school than younger children, as the coefficient in row 2 is negative. Table 2. Individual and demand-side determinants of the probability of attending (pre)school, 2010–20 (1) VARIABLES Enrolled Female -0.226*** (0.00518) Age -0.0394*** (0.000828) Partner in HH 0.0321** (0.0128) No. of HH members with max. primary schooling -0.0193*** (0.00617) No. of HH members with secondary schooling -0.00604*** (0.00216) No. of HH members with tertiary education 0.0106* (0.00613) HH members 0.000273 (0.00226) No. of working HH members -0.0152*** (0.00388) No. of unemployed HH members -0.0418*** (0.0100) Bottom 40 -0.00928 (0.00709) Urban -0.00492 (0.00665) Crime, violence, or vandalism in the area -0.00895 (0.00902) Constant 0.926*** (0.0152) Observations 22,482 R-squared 0.199 Note: The table presents results from running a linear probability model on the probability of children younger than 16 years old attending (pre)school, using data from the EU-SILC 2010–2020 (https://ec.europa.eu/eurostat/web/microdata/european-union-statistics-on-income-and-living-conditions). Robust standard errors are in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. Poverty and Child Labor Could Restrict the Demand for Education Individual factors that drive educational gender inequalities could be related to poverty and the need to generate household income. Previous studies have shown that low-income families might have their children engage in income-generating activities, which might drive children out of school (ILO and UNICEF 2021). Child labor affects boys more than girls, is more prevalent in rural than urban areas, and especially affects low-income families (ILO and UNICEF 2021). These patterns of results would appear to be in line with our observations in the case of Romania. However, studying the possibility of child labor as a driver of low net enrollment rates, especially among boys, would require updated and disaggregated data on child labor and time use among children. A study by Ioana et al. 39 (2015) in Romania relates school dropout to deepening poverty, especially in rural areas. While we do not find evidence in this direction (in table 2 the coefficient of belonging to a household in the B40 is negative but insignificant at the standard significance levels), more-detailed analysis is needed to confirm this channel. There is some evidence showing that poverty could affect boys’ school attendance more negatively than girls’. Previous research from other countries also shows that when families face limited resources, they prefer that boys attend school (MEB Primary Education General Directorate and UNICEF Turkey 2011). We investigate this possibility by interacting the indicator of belonging to a B40 household with children’s gender. Indeed, we find that the indicator variable is significant and positive, which means that there is a gendered impact of poverty on school attendance. While the impact of boys is negative (although insignificant), the interaction term (which is significant at the 1-percent level) is positive, showing that girls’ school attendance is not negatively impacted by poverty (for the detailed results, see the appendix). These patterns of results could be related to the traditional roles of men and women in society and the pressure on boys to generate household income. Evidence of Intergenerational Education Attainment and the Family Context Children’s families play a crucial role in explaining school dropout in Romania. In a study conducted by Ciolca (2020), respondents indicated that a lack of awareness of the importance of education and a lack of parental support for children’s education, learning difficulties, and low development levels of personal skills drive school dropout in Romania, along with the involvement of children in household chores and income-generating activities. The evidence in table 2 confirms the crucial role of family characteristics in determining children’s probability of attending school, especially intergenerational educational attainment. Several of the coefficients measuring the correlation between children’s school participation and household characteristics in table 2 are significant at the standard levels, demonstrating the crucial influence of the family system on children’s human capital accumulation. Moreover, there is evidence of intergenerational transmission of educational attainment, in line with previous evidence from other countries (for example, Europe: Checchi, Fiorio, and Leonardi 2013; Austria: Fessler, Mooslechner, and Schürz 2012; several African countries: Azomahou and Yitbarek 2016). The intergenerational transmission of educational attainment can be measured via the correlation between the parents’ and the child’s level of schooling (Huang 2013). The coefficients in table 2 reveal that the larger the number of household members with lower than tertiary schooling, the lower a child’s probability of attending school. In contrast, the higher the number of household members with tertiary education, the higher a child’s probability of attending school. The intergenerational transmission of educational attainment is more persistent for girls than for boys. When interacting with these variables by children’s gender, we find that the impact of intergenerational transmission of educational attainment is larger for female children. Consequently, it is crucial to introduce mechanisms that target these educational traps, especially among girls and their families. The fact that girls in Romania are more affected by intergenerational education persistence is in line with previous evidence from Austria (Schneebaum, Rumplmaier, and Altzinger 2014) and Turkey (Demirel and Okten 2020). The appendix presents the results with the interaction terms. 4.4. High Levels of Teenage Pregnancy and School Dropout Romania reports high rates of child pregnancy and underage abortions, which could potentially lead to school dropout. Romania had the highest rates of teenage pregnancy in the EU in 2015 (Eurostat, 40 2017) and underage abortions (Eurostat, 2023)36, probably due to inadequate policies addressing child pregnancy or reproductive health services, but further evidence is needed to confirm this channel. In 2020, Romania ranked seventh in the EU in terms of abortion rates.37 Moreover, in 2015, among the EU member states Romania has the highest proportion of teenage mothers, with a rate of 12.3 percent compared to the EU average of 4 percent (Eurostat 2017). The rate was higher in rural than urban areas, probably driven by lower socioeconomic levels and access to education in these areas (Iorga, Socolov, and Socolov 2016). Lack of education and teenage pregnancy might create vicious circles limiting educational attainment among girls and young women. Many recent studies relate the high rates of school dropout to teenage pregnancy. Several studies on the drivers and consequences of the high levels of teen pregnancy in Romania show that they lead to school dropout (Radu et al. 2022; Diaconescu et al. 2015). Iorga et al. (2021), for example, interviewed mothers of girls who had become pregnant: 7 out of 10 mothers named school dropout as the most frequent risk related to teenage pregnancy, followed by psychological trauma, and 3 out of 4 supported school-based sexual education. Across the studies there is the general conclusion that Romania urgently needs a national intervention that combines medical access, sexual education, and family involvement to tackle the high level of teenage pregnancy (Diaconescu et al. 2015). Sexuality Education Sexuality education is a controversial topic, and several groups argue against the relevant interventions, though they are urgently needed. There is extensive evidence showing that sexuality education results in lower adolescent birth rates (UNESCO 2023). Sexuality education is still a highly controversial topic in Romania and not obligatory, according to consultations and qualitative evidence (The Decree Chronicles 2021). Since 2022, sex education can only be taught from the eighth grade onward and only with parents’ written content (Gyaraki 2022). This development could negatively impact boys’ and girls’ access to important information about safe sex, sexually transmitted diseases, contraceptive use, and consent. Abortion Rights Hard-fought-for abortion rights are being threatened by hardening attitudes toward abortion. While abortion was legalized by the end of the 1980s, country consultations and soft evidence38 suggest that women still face significant barriers to abortions in practice, and some even report malpractices. The reluctance to have abortions could explain why abortion rates have fallen over time and are now even slightly below the EU average (figure 45). There could be other drivers behind these developments, such as increased contraceptive usage. Unsafe abortions and overall maternal mortality rates rose sharply when abortion was illegal in Romania from 1966 to 1989 (Horga, Gerdts, and Potts 2013), so facilitating access to abortions could avoid health risks. Access to Birth Control At the same time, the share of women using contraception has increased over time, although updated information is not available for Romania. Figure 46 plots the data available for the years 36 Romania reported 153 legally induced abortions for mothers below 15 years old in 2021, only outnumbered by Germany and Spain (Eurostat, 2023 - indicator demo_fabort). There was a significant decline over time (from 664 in 2013). 37 The abortion rate is the number of abortions per 1,000 women of reproductive age in a given year. We rely on the “abortrt” indicator published by Eurostat. 38 A study conducted by Vdovîi and Bird in 2019 found that over 30 percent of hospitals (60 out of 190 contacted hospitals) refused to grant abortions to women (Vdovîi and Bird 2019). 41 1978, 1993, 1999, 2004, and 2005. As the figure shows, the share of women using contraception has increased during this time, although a significant share of women abstained from using modern methods. In 2020, Romania ranked in the middle ground of the Contraception Atlas, and the European Parliament Forum for Sexual and Reproductive Rights estimates its contraceptive prevalence rate to be 55 percent (EPF 2023). Updated information on contraception usage in Romania and disaggregated statistics by age and income groups could help to design tailored and effective policies to address the relatively high levels of teenage pregnancy. Figure 45. Abortion rates per 1,000 women of reproductive Figure 46. Share of women of reproductive age using age, Romania vs. the EU, 2013–20) contraception, any vs. any modern method (%) 16. 80.0 14. 70.0 12. 60.0 10. 50.0 8. 40.0 6. 30.0 4. 20.0 2. 10.0 0. 0.0 2013 2014 2015 2016 2017 2018 2019 2020 1970 1980 1990 2000 2010 Romania EU average Any method Any modern method Sources: Eurostat (https://ec.europa.eu/eurostat/data/databasehttps://ec.europa.eu/eurostat/data/database), indicator “abort,” (figure 45) and World Contraceptive Use 2022 (figure 46). Note: We calculate the EU average by taking the average abortion rates of all countries in the EU that report data. In figure 46, “any method” indicates methods such as withdrawal or rhythm, while “any modern method” indicates methods such as the pill, condoms, and sterilization. 5. Current Political Landscape The educational system in Romania is overseen at the national level by the Ministry of Education, which cooperates with other ministries and subordinate institutions. The basic principles of the educational system are formulated in the Romanian Constitution, specifically in Article 32 of Chapter II of Fundamental Rights and Freedoms in the “Right to Learning” section (European Commission 2022). All citizens of Romania, other EU member countries, EEA countries, and the Swiss Confederation have equal rights of access to school education, higher education, and lifelong learning. While the educational principles do not specifically mention the equality of boys and girls, they enshrine the principle of equity and nondiscrimination, as well as ensuring the equality of opportunity and social inclusion (European Commission 2022). There have been significant efforts to improve educational outcomes in Romania, and although these efforts have lacked a gender-sensitive perspective per se, they could have beneficial effects on gender equality. Over time, several local and national programs have been/are being implemented, to improve school and university infrastructure, including programs such as: School Rehabilitation Program, School Infrastructure Rehabilitation Project, Early Education Reform Project, Program Safer, Inclusive and Sustainable Schools, the National Program for Local Development, operational programs financed from non-reimbursable European funds (such as the Regional Operational Programme), projects financed from the National Recovery and Resilience Plan (NRRP), and other programs financed from the national and local budgets. One prominent 42 example is “Educated Romania,” a national project initiated by the President of Romania to rethink the educational culture in Romania. This initiative was supported by an allocation of 3.6 billion euros (International Trade Administration 2023). Other reforms that were implemented during the current programming period that ends in 2023 are curriculum reforms, professionalization of teaching, and supporting at-risk children (International Trade Administration 2023). As more-educated boys and girls are more likely to engage in behaviors that promote gender equality, generally speaking all investments and efforts around educational outcomes in Romania could potentially have beneficial effects on gender equality. Still, incorporating a gender-sensitive perspective into these initiatives could make them more effective in contributing to gender equality in Romania. The Ministry of Education has presented two draft education bills, namely the Preuniversity Education bill and the Law on Higher Education bill, to the Romanian Parliament. These draft bills form part of the overarching Educated Romania Policy. Romania has implemented national strategies to achieve gender equality that also made a connection to the educational sector. Between 2018 and 2021, Romania had the National Strategy for the Promotion of Equal Opportunity and Treatment between Women and Men in place and education was one domain for interventions undertaken as part of the strategy. These consisted of three main activities: gender mainstreaming in policies and programs, creating networks of experts, and increasing NGO capacities and gender budgeting (EIGE 2022). For 2022, the government approved an updated version of this strategy: the National Strategy for Promoting Equal Opportunities between Women and Men and Preventing Domestic Violence 2021–2027. Romania currently has a conditional cash transfer program in place that is linked to preschool attendance. The program, introduced in 2015, was inspired by the “Every Child in Preschool” initiative by a local association, Asociatia OvidiuRo (Gheorghiu et al. 2020). The main goal of the initiative was to close gaps in access to kindergartens among Roma and other marginalized groups. It is a social benefit for children ages three to six from disadvantaged families. The social benefit is paid as a shopping voucher (social ticket) and assumes a value of lei 119.6 per month. The social ticket can be used for purchasing education-related goods or food. The child is entitled to receive this financial aid for the duration of kindergarten on the condition that the child regularly attend kindergarten. Importantly, only impoverished children are eligible for the social ticket. 6. Policy Recommendations We next provide a menu of tailored interventions from the previous literature and our newly generated evidence, as well as best practices and impact evaluations from the international literature, that can help to address each of the 10 high-level recommendations we derived from our diagnostic. In table 3, we connect high-level policy recommendations and potential entry points to address these via detailed interventions. These interventions are derived from our diagnostic, previous studies in the Romanian context, and best practices/impact evaluations from the literature. For the full overview of the diagnostic and recommendations, see table A.1. Table 3. High-level policy recommendations and potential entry points of detailed interventions High-level policy recommendation Potential entry points of detailed interventions (a) Detailed and updated gender-sensitive demand and (1) Study and address demand- and supply-side supply assessment of formal childcare and ECE services constraints on access to ECE for zero- to two- to design tailored interventions to increase access; year-olds (b) provide financial incentives to support enrollment among the poor; 43 (c) increase public spending, quality, and available infrastructure in the area of ECE (interventions to tackle acceptability) (a) Alleviate financial constraints through targeted (2) Implement gender-sensitive policies intervention strategies; targeting access, supply, and usage of ECE (b) Study of girls facing additional constraints to ECE a) Norm-based interventions tackling social norms that constrain boys’ and girls’ educational participation; (b) Interventions that target the awareness of the importance of educational attainment among parents and children (Apostu 2014); (c) Teacher training on gender stereotypes and marginalization of other subgroups of the population (for example, the poor or Roma) (Apostu 2014); (d) School-based, violence-preventing interventions and eliminating violence at schools (Anderson 2011); (e) Increase public spending and invest in educational quality and attractiveness (Apostu 2014); (f) Reentry programs and childcare services for young (3) Design and implement interventions that mothers (Robayo-Abril and Rude forthcoming) and those target school dropout and low enrollment rates that tackle child pregnancy; (g) Address the marginalization and discrimination faced by Roma boys and girls that play a crucial role in school dropout in Romania (Apostu 2014); doing so requires bottom-up approaches (Andrei, Tkadlecova, and Martinidis 2015). Teacher training on the sensitivities around the marginalization of and discrimination against Roma students is also crucial (Apostu 2014). (h) Previous studies have shown that (conditional) cash transfer programs lead to improved educational outcomes (Baird et al. 2014). (i) Tackle intergenerational transmission of education, which is more persistent for girls than for boys. (a) Facilitate interactions with matched-background mentors (Kricorian et al. 2020); (b) Norm-based interventions that target stereotypes; (c) Partnering with the private sector, addressing gender (4) Address gender segregation in education: biases in learning materials and among parents, and incentivize and invest in the development of encouraging girls’ participation in extracurricular girls’ skills and interests in STEM and ICT and of activities (Hammond et al. 2020); boys’ skills in reading and female-dominated (d) Strategies that address the enforcement of gender fields of study stereotypes in the educational system, such as the introduction of bias-free schoolbooks, instrumentalizing the media, gender-responsive teacher training, and parental workshops (Bruck and Cater 2016; Concordă 2018; Unterhalter et al. 2014; UNICEF 2023; Farré 2012). Targeted interventions that factor in the multidimensionality of (reversed) gender gaps are better (5) Understand gender equality as a suited; teachers should focus on addressing the multidimensional subject marginalization and discrimination that impede Roma boys and girls from fulfilling their potential in the schooling system (Rotaru 2019). Follow international recommendations around the (6) Design and implement policies that address underachievement and low enrollment of boys and roll reversed gender gaps out interventions in the macro-, meso-, and microenvironment (UNESCO 2022). 44 Invest in research to identify what works best, for (7) Fund gender-sensitive research projects and example via impact evaluations and rigorous monitoring M&E initiatives and evaluation strategies, in collaboration with the academic community and technical specialists. (8) Generate systematic information on Roma Include ethnicity parameters in surveys; generate children compared to non-Roma children by administrative data on subgroups and make information gender; generate information on child labor accessible to the public; generate data on child labor and and time use of children time use. Conduct a gender-sensitive public expenditure review of (9) Increase public spending on education the educational sector and invest money where it is most needed, applying gender-budgeting tools. Gender-responsive education sector planning (GRESP) is an efficient way to engrave gender equality into the heart (10) Develop gender-responsive education of national education systems and consequently address sector planning (GRESP) and an operationalized gender equality by design (UNGEI 2023). An gender-sensitive assessment tool operationalized gender-sensitive assessment tool could help with tracking educational gender gaps more closely. 1) Study and address demand- and supply-side constraints behind access to ECE for zero- to two-year-olds Our findings reveal significantly low enrollment rates of children ages zero to two years in Romania compared to the EU average. In light of this evidence, we propose implementing a comprehensive study that investigates and addresses the constraints related to both the demand and supply sides of access to ECE for this age group. Our research indicates that these constraints primarily revolve around the availability, affordability, and acceptability of ECE services. Availability constraints manifest through a limited number of formal public childcare centers and an increasing reliance on private institutions for childcare provision. Additionally, there are negative trends in the number of available creches, nurseries, and kindergartens with legal status that affect both rural and urban areas. Affordability poses a challenge, particularly for the economically disadvantaged population, as the gross and net childcare costs (as a proportion of women’s median full-time earnings) in Romania are below the EU average of 14 percent but are higher for the poor. Acceptability constraints arise from the prevalent parental workforce participation and a significant preference for parental care of children under three. In fact, a substantial proportion of children (76.8 percent) in Romania receive care exclusively from their parents, which is the highest rate in the EU.39 Considering these findings, we put forward the following recommendations: a) Conduct a detailed and up-to-date gender-sensitive assessment of demand-side and supply- side factors concerning formal childcare and ECE services. This assessment will serve as the foundation for designing tailored interventions to enhance access for children ages zero to two years. b) Implement financial incentives to support enrollment among families facing economic challenges, with a particular focus on the poor segment of the population. c) Increase public spending to improve the quality of ECE services and expand the available infrastructure. Such interventions will contribute to addressing the acceptability barriers related to parental preferences and workforce participation. 39 Eurostat, https://ec.europa.eu/eurostat/databrowser/view/ILC_CAPARENTS__custom_6629577/default/table?lang=en. 45 2) Implement gender-sensitive policies targeting access, supply, and usage of ECE To address gender disparities in ECE among children ages three to five years, we propose the implementation of gender-sensitive policies that specifically target access, supply, and usage of ECE services. International evidence such as a report by MEB Primary Education General Directorate and UNICEF Turkey (2011) suggests that families, when faced with limited resources, may prioritize sending boys to school over girls. Based on these findings, we recommend the following actions: a) Implement targeted intervention strategies to alleviate financial constraints that hinder access to ECE services. These interventions should be designed to specifically address the financial barriers that families face, particularly those that disproportionately affect girls’ enrollment rates. By addressing these constraints, access to and gender equality in ECE participation will be enhanced. b) Conduct a comprehensive study to identify and understand the additional constraints that girls may encounter in accessing ECE. This study should focus on exploring the specific challenges and barriers that hinder girls’ enrollment and participation in ECE programs. By gaining a deeper understanding of these constraints, policy makers can develop targeted interventions that effectively address the unique needs and circumstances of girls. By implementing gender-sensitive policies that target access to and supply and usage of ECE services for children ages three to five years and by addressing the financial constraints and additional barriers that girls may face, greater gender equality and inclusivity can be fostered in ECE. Moreover, increasing enrollment rates among children of this age group, in general terms, would benefit Romania in the long run; the rates are currently below the EU average. 3) Implement interventions that target school dropout and low enrollment rates Reversing recent trends of performance drops and school dropout among both boys and girls would be beneficial to gender equality. This assessment shows that educated men and women are less likely to believe in harmful gender norms than are uneducated men and women. At the same time, it shows that Romania has recently experienced some adverse trends in its human capital accumulation, negatively affecting both boys’ and girls’ educational outcomes in several dimensions and at several educational stages. Consequently, there is an urgent need to reverse the negative trends in educational outcomes in Romania to avoid vicious circles around gender equality. Educating both men and women in Romania is beneficial for sustainable development in a broader context and it also fosters gender equality. A detailed updated study on the underlying drivers of negative trends in educational outcomes would help to identify the most effective policies to address them. The government currently has a program in place that targets high dropout rates, the National Programme for Reducing School Dropout (PNRAS), and has the goal of reducing school dropout in at least 25 percent of the educational units participating in the program.40 However, a detailed updated study that analyzes the underlying drivers of school dropout and worsening test results in Romania could help to shed light on which interventions could tackle the underlying mechanisms behind these negative trends. The analysis at hand shows that the underlying drivers could be manyfold. Declines in educational indicators could be related to a lack of investment in the necessary infrastructure, for which the remedy is an increase 40 This information was shared with a World Bank Team in January, 2023, during technical consultations in Bucharest, Romania. 46 in government expenditure on education. They could also relate to the need of families to have members enter the labor market early on to generate additional household income or to households’ lacking the financial means to pay for a high-quality education of their children. In either case social protection programs and cash transfer programs might be the most appropriate. On the contrary, if the trends are related to gender norms, norm-based interventions might be most effective. A report by Apostu (2014) indicates that three factors contribute to children’s being out of school and at risk of dropping out in Romania: sociocultural demand-side factors, economic demand-side factors, and supply-side barriers. It has been argued that these factors have asymmetric effects on boys and girls, resulting in both gender gaps and reversed gender gaps. Considering this evidence, the following recommendations are proposed: firstly, the implementation of norm-based interventions is recommended to address social norms that restrict educational participation for both boys and girls (OECD 2022). Secondly, interventions should focus on raising awareness among parents and children about the significance of educational attainment (Apostu 2014). Additionally, teacher training on gender stereotypes is essential to challenge and counteract biases in the educational context (Apostu 2014). School-based interventions aimed at preventing violence, such as teacher training and the inclusion of curricular-based interventions on topics like history or health, as well as teaching conflict resolution skills, are recommended (Anderson 2011). Increasing public spending and investing in educational quality and appeal are vital for improving educational outcomes (Apostu 2014). Reentry programs and the provision of childcare services for young mothers are also important measures (Robayo-Abril and Rude forthcoming). Addressing the marginalization and discrimination faced by Roma boys and girls is crucial in tackling school dropout in Romania; doing so requires approaches that engage communities from the bottom up (Andrei, Tkadlecova, and Martinidis 2015). Additionally, teacher training focusing on sensitivities surrounding the marginalization and discrimination of Roma students is imperative (Apostu 2014). Lastly, previous studies have demonstrated that conditional cash transfer programs have the potential to enhance educational outcomes (Baird et al. 2014). We also generate evidence showing the significant role of the family context and intergenerational transmission of education, especially among girls; interventions that target intergenerational education persistence are crucial. Researchers have explored several strategies to address the intergenerational persistence of education. For example, Demirel and Okten (2020) show a positive effect of compulsory schooling in Turkey. Akresh, Halim, and Kleemans (2023) show that the construction of more schools have positive intergenerational effects on educational outcomes. Several other studies from Europe show that the intergenerational persistence of education is stronger in highly tracked educational systems (Lange and von Werder 2017; Burger 2016; Hanushek and Wößmann 2006). Working with families is also crucial (Oliva, Parra, and Sánchez-Queija 2019). To tackle school dropout by teenage mothers, Romania needs tailored reentry policies and support measures. Evidence cited in this assessment highlights the facts that, on the one hand, Romania suffers from high levels of teenage pregnancy, and on the other faces increasing levels of school dropout. As several studies show a connection between the two issues, Romania should invest in tailored reentry programs for teenage mothers. Providing access to childcare for young mothers could facilitate their access to education. Giving young mothers access to childcare could facilitate their reintegration into the educational system. Investing in childcare also has many other benefits, such as better development outcomes among children, increased family welfare, and ultimately greater productivity and economic growth (World Bank 2022). Childcare should be of high quality and accessible in financial terms as well as close to the mother’s home or school. 47 Sexuality education should be obligatory and its design should follow best practices. Given the evidence that the provision of sexuality education results in lower adolescent birth rates (Dongarwar and Salihu 2019), Romania should follow other countries with a long-standing history of obligatory sex education as part of the school curriculum, such as the Scandinavian countries (UNESCO 2023; Goldfarb and Lieberman 2021). Sexuality education should follow best practices and be continuous, multisectoral, context oriented, gender responsive, and implemented in cooperation with youth, parents, and the broader community (WHO Regional Office for Europe and BZgA 2010). If sexuality education does not follow these standards, it might perpetuate gender stereotypes or remain ineffective in addressing the underlying drivers of gender-based violence, sexually transmitted diseases, and teenage pregnancy (Hall et al. 2019). Increasing access to and information about contraception is crucial in Romania, given the high adolescent pregnancy rate, and education around reproductive health needs to target men and women. The use of contraception should be destigmatized in Romania among the broad public, for example via a communication campaign (EPF 2023). Moreover, the cost of contraception should be reimbursed, especially for low-income and vulnerable groups. A study by Rada (2014) finds that certain groups in Romania are at higher risk of unwanted pregnancy, such as those in rural areas, those with lower educational levels, and those with low income. Based on these findings, these groups should be prioritized when designing intervention strategies. Overall, the inclusion of boys and men in reproductive health education is crucial (Farré 2012). Past evidence shows that fertility and family planning programs that only focus on females meet with limited success (Farré 2012). 4) Address educational segregation in the field of study and gender gaps in digital literacy; girls underperform boys in math while boys underperform girls in reading Addressing gender segregation in the field of study and gaps in digital literacy could decrease economic gender inequalities later in life. We find evidence of gender segregation in the field of study and gender gaps in digital literacy rates. Interventions that increase girls’ digital skills and increase their interest in STEM- and ICT-related fields could address economic gender gaps later in life. Potential strategies to increase the share of girls in these fields are role models, partnering with the private sector, addressing gender biases in learning materials, including parents, and encouraging girls’ participation in extracurricular activities (Hammond et al. 2020). Other strategies are interactions with matched-background mentors (Kricorian et al. 2020) and norm-based interventions that target stereotypes (OECD 2022). Potential strategies to address the enforcement of gender stereotypes in the educational system are the introduction of bias-free schoolbooks, gender-responsive teacher training, and parental workshops. Examples of potential interventions to address gender inequality are parenting workshops (Bruck and Cater 2016), open communication between teachers and parents about gender equality and raising awareness among parents about potential gender stereotypes (Warner and Barerra 2005), and replacing current schoolbooks with bias-free ones (Concordă 2018). Educating teachers about the influence of gender stereotypes and informing them about potential school-based interventions that target gender equality could also have a positive effect (Unterhalter et al. 2014). The implementation of gender-responsive pedagogies is crucial and should be incorporated into teacher training and professional development (UNICEF 2023). Instrumentalizing media pieces that empower both boys and girls to access and invest in their education could also be beneficial (Farré 2012). 5) Understand gender equality as a multidimensional subject 48 This report reveals that educational gender inequalities are multidimensional. We show that enrollment rates are often higher in urban than rural areas and for high-income groups compared to low-income groups . Furthermore, the sizes of gender gaps also differ between these groups. Other vulnerable groups, such as Roma girls and boys or those with migration backgrounds, might face additional challenges, but the possibility of empirically measuring these intersectional vulnerabilities is limited by data and evidence gaps. In addition, addressing marginalization and discrimination faced by vulnerable groups, such as Roma boys and girls, is crucial. To achieve greater inclusion and gender equality across different population groups, it is important to factor in the multiple vulnerabilities some boys and girls face. There is an important rural-urban divide in several of the educational indicators studied. The evidence in this report makes the case for a persistent rural-urban divide in Romania. To avoid further disparities in the short and long runs between regions, it is recommended to invest in the educational sector in Romania in those regions that lag behind. Prioritizing rural areas is crucial to avoid further inequalities. Several studies note that Roma women and girls face additional barriers in the educational sector; a bottom-up approach would address these. Andrei, Tkadlecova, and Martinidis (2015) show that Roma women and girls face additional challenges in comparison to non-Roma women and girls. They relate the ineffectiveness of current policies and interventions to address these challenges to the top- down approach that is currently used. The researchers recommend a paradigm shift in this area, stating that a bottom-up approach would be more efficient in addressing the additional needs of Roma women and girls. These findings might also apply to other vulnerable groups, such as persons with migration backgrounds, persons living in poverty, or persons with disabilities. Teachers should focus on addressing the marginalization and discrimination that impede Roma boys and girls from fulfilling their potential in the schooling system (Rotaru 2019). 6) Design and implement policies that address reversed gender gaps Our analysis reveals that boys exhibit lower performance than girls across various educational indicators and among different demographic groups. Consequently, it is crucial for Romania to implement strategies that specifically address the issue of reversed gender gaps. These gaps exist because of complex and multifaceted constraints. For a summary of the current state of the literature, see UNESCO (2022). International evidence highlights the influence of gender norms, which can contribute to boys’ leaving school early. Additionally, labor market demands and the need for household income generation can disproportionately affect boys. Furthermore, boys may experience higher levels of school-based violence and bullying compared to girls. In response to these challenges, Romania should adopt international recommendations focused on the underachievement and low enrollment of boys and implement interventions at the macro-, meso-, and microlevels of the education system (UNESCO 2022). For instance, interventions targeting gendered social norms associated with traditional notions of masculinity (Farré 2012; Heilman 2018) and violence prevention programs are recommended (UNESCO 2022). A recent literature review conducted by UNESCO (2022) offers a comprehensive list of interventions at the micro-, meso-, and macrolevels that can inform policy and practice toward addressing reversed gender gaps in Romania. School-based violence-preventing interventions and eliminating violence at schools could transform the current school system into a safer environment, especially for boys. Several studies have noted that school-based violence and bullying are persistent problems in Romania and that discrimination based on gender, ethnicity, and socioeconomic status is common among students. Interventions tackling these phenomena are urgently needed in Romania to avoid cycles of abuse and violence, 49 which ultimately might also affect the current and long-run exposure of the population to gender- based violence. Examples of potential strategies are teacher training, incorporating curricular-based intervention via history or health classes, and teaching conflict resolution (Anderson 2011). Interventions that target social norms around traditional concepts of masculinity might improve the educational attainment of both boys and girls and could be implemented via the school system. Our results show that men see themselves confronted with traditional gender norms around masculinity that might disincentivize them from investing in higher education and skills. At the same time, girls are also subject to harmful gender norms. Based on these findings, interventions that target these social norms and facilitate the discussion of the role of men in society might improve the educational attainment of both men and women. Including boys and men in intervention programs that empower women is recommended, given that doing so would help to eliminate ill feelings against women and broaden the understanding of men of their role in achieving gender equality (Farré 2012). Inclusive intervention programs might result in males’ feeling more engaged, empowered, and involved in the mission to achieve equality between the sexes (Farré 2012). In addition, supporting positive masculinity, using fatherhood as the starting point to embrace gender equality, and discussion groups around concepts of masculinity have proven to be effective tools (Heilman 2018). 7) Fund gender-sensitive research projects and M&E initiatives Additional research and systematic monitoring and evaluation could help with tracking the progress toward gender neutrality in education and identify the most efficient and effective interventions. There are important knowledge gaps about the status of gender inequality in the educational sector and its underlying drivers. Moreover, there is no systematic application of monitoring and evaluation systems to track what works best. Introducing M&E initiatives and funding more research around gender-sensitive educational interventions and interventions tackling barriers that produce gender inequalities in educational outcomes could help generate knowledge about what works best. 8) Generate systematic information on Roma children compared to non-Roma children; generate data on child labor and time use of children There are several important data gaps around educational outcomes in Romania. One example is recent data on abortions by age group, income level, and other sociodemographic characteristics. Moreover, data on contraceptive usage by groups could improve the understanding of and address the high teen pregnancy rate. Additionally, it would be useful to update the learning-poverty indicator every year. As part of the research conducted for this report, we also noted important data limitations around the schooling outcomes of Roma women and men (boys and girls). Consequently, we recommend that the Ministry of Education start to track educational indicators for different groups, for example, Roma women and girls or women and girls with migration backgrounds. The generated data can help with the design of tailored policy interventions. To enhance understanding of educational disparities and inform targeted interventions, it is essential to generate systematic information that compares Roma children to their non-Roma counterparts. Currently, there is a lack of updated, representative, and systematic data on the educational outcomes of Roma children and children from other vulnerable groups, particularly when considering gender differences. To address this gap, we recommend incorporating ethnicity parameters in surveys and collecting administrative data on specific subgroups within the Roma population. Increasing the sample size of children with migration backgrounds is also recommended. Furthermore, it is crucial to make this information readily accessible to the public. 50 Similarly, in order to gain insights into the relationship between low enrollment rates, school dropout, poverty, and the necessity of income generation, it would be valuable to collect data on child labor and the time use patterns of children. With such data in hand, policy makers and stakeholders can develop targeted strategies to address the unique challenges faced by vulnerable populations and promote equitable educational opportunities for all children. 9) Increase public spending on education Romania should increase its government expenditure on education. The adverse trends in educational indicators for the country and the significant gaps between Romania’s scores and the EU averages could be related to Romania’s relatively low public spending share on education. Increasing the government spending share in the educational sector could create important benefits for capital accumulation, more broadly speaking, as well as address some of the potential drivers behind (reversed) gender gaps, such as the inadequate infrastructure and sanitation, limited resources available for teachers, supplies, and digital tool kits. Gender budgeting could improve government budgeting in the educational sector in Romania. Gender budgeting is a gender-sensitive approach to government budgeting that entails approaches such as gender-sensitive budget monitoring and evaluation systems, as well as the restructuring of fiscal policies and budgetary allocations (Stotsky 2016). Consequently, it takes place at different stages of the budgeting process (Stotsky 2016). Gender budgeting improves government budgeting based on the principles of progressive public finance. It is efficient in tackling gender gaps; externality correcting, especially with respect to the care economy; and equity enhancing (Stotsky 2016). Importantly, Stotsky (2016) identifies the need for spending ministries to take the lead in identifying and defining gender-oriented goals and then requesting budgets that enable the reaching of these goals. To close (reversed) gender gaps in Romania, the Ministry of Education should assume such a leading role and connect its own gender-oriented goals to the overarching gender budgeting plan of the Romanian government. Gender budgeting can increase transparency, accountability, tracking, and ultimately accelerate the achievement of gender-oriented goals embedded in public policies. 10) Develop gender-responsive education sector planning (GRESP) and an operationalized gender-sensitive assessment tool Gender-responsive education sector planning is an opportunity for Romania’s educational sector to address gender inequalities by design. Gender-responsive education sector planning (GRESP) is an efficient way to engrave gender equality into the heart of national education systems and consequently address gender equality by design (UNGEI 2023). This type of planning can help to identify both gender barriers within education systems and effective strategies and policy interventions to close these barriers. GRESP ultimately creates gender-sensitive policies, planning, and learning environments. It covers a broad set of interventions that target gender equality, such as mainstreaming gender issues in teacher training and curricula, eliminating gender-based violence and discrimination in schools, and gender-sensitive monitoring and evaluation systems. Implementing GRESP could be a valuable approach in the effort to achieve gender equality in Romania. Importantly, the preparation of this type of plan should be informed by a gender analysis of the educational sector that identifies critical gender disparities and their underlying factors (UNGEI 2023). When undertaking GRESP, it is recommended to follow best practices, such as those outlined in the “Guidance for Developing Gender-Responsive Education Sector Plans” by UNGEI (2018). An operationalized gender-sensitive assessment tool could help with tracking educational gender gaps more closely. Romania currently has two objectives within the educational sector. The first 51 objective is to reduce dropout and the second is to fight educational poverty among children and young people.41 Given Romania’s overall poor performance in gender equality when compared to the rest of the EU (EIGE 2022), it would be beneficial to include one gender-related objective in the overarching political framework. Given the specific country context in Romania, which is marked by partly positive but also reversed gender gaps, one possibility would be setting a goal of educating the Romanian people about gender equality, more broadly speaking, and building this education up from the bottom, starting in ECE facilities and providing this education through adulthood. Romania’s educational gender strategy should be both about educating males and females and empowering them to fulfill both their roles and responsibilities in an agenda to achieve gender parity. 7. Conclusion In this report, we analyze gender inequality in the educational sector in Romania and shed light on the underlying drivers of these inequalities. We do so for the population as a whole but also for certain subgroups, such as the low- and high-income groups and the rural and urban populations. A detailed assessment of the educational outcomes of Roma boys and girls versus non-Roma boys and girls, or those of children with migration backgrounds, is limited by the current data environment. We analyze gender inequality in educational outcomes for different educational levels and also analyze recent developments in educational performance among boys and girls. In investigating the underlying drivers, we analyze contextual supply-side factors, other contextual factors, and individual demand-side factors. We mainly rely on descriptive data analysis, though we also make use of household surveys. We find that there are multiple gender inequalities in the educational sector in Romania, and these are related to supply-side contextual factors, other contextual factors, and individual demand-side factors. While for some indicators, we find that girls lag behind boys, there are also cases in which boys lag behind girls. To achieve gender equality in the educational sector, the Romanian government should focus on closing both of these types of gaps. Moreover, given the complexity of the gender inequalities identified in this report, the government should implement a nuanced and tailored approach toward gender equality in education. The underlying reasons behind these inequalities are manyfold and are associated with both the contextual and the individual levels. They range from gendered social norms, lack of important infrastructure and public investment, and deficiencies in educational quality to discrimination and marginalization. Teenage pregnancy also interact negatively with educational participation. Moreover, there is evidence of intergenerational educational persistence, especially among girls. We identify ten policy areas and recommend multiple possible interventions to tackle gender inequalities in education in Romania. We identify ten areas for potential policy interventions and provide a menu of possible concrete measures to correct gender inequalities in these areas based on the evidence presented in this report as well as in previous studies. These measures range from increasing enrollment rates in ECE, overtargeting school dropout and falling enrollment rates, addressing gender segregation in the field of study, and reducing reversed gender gaps to funding more research and M&E initiatives and data. Moreover, it is recommended that gender-responsive education sector planning be developed, the public spending share on education be increased, and that gender inequality be understood as a multidimensional phenomenon. 41 These goals were shared in technical discussions conducted between a World Bank team and a team from the Ministry of Education in January 2023, in Bucharest, Romania. 52 References Aikman, S., and N. Rao. 2012. “Gender Equality and Girls’ Education: Investigating Frameworks, Disjunctures, and Meanings of Quality Education.” Theory and Research in Education 10(3): 211–28. https://doi.org/10.1177/1477878512459391. Akresh, R., D. Halim, and M. Kleemans. 2023. “Long-Term and Intergenerational Effects of Education: Evidence from School Construction in Indonesia.” The Economic Journal 133 (650): 582–612. Alan, S., S. Ertac, and I. Mumcu. 2018. “Gender Stereotypes in the Classroom and Effects on Achievement.” Review of Economics and Statistics 100 (5): 876–90. Anderson, Jill. 2011. “Violence: What Schools Can Do.” Harvard Graduate School of Education. https://www.gse.harvard.edu/news/11/06/violence-what-schools-can-do. Andrei, Roxana, Tana Tkadlecova, and George Martinidis. 2015. “Challenges faced by Roma Women in Europe on Education, Employment, Health, and Housing: Focus on Czech Republic, Romania, and Greece.” Balkan Social Science Review 4. https://www.researchgate.net/publication/313899271_Challenges_faced_by_Roma_Women_in_Eu rope_on_Education_Employment_Health_and_Housing_focus_on_Czech_Republic_Romania_and_G reece. Apostu, Ottilia. 2014. Romania Country Study: Analysing the Situation of Out Of School Children. UNICEF Romania Office. https://unesdoc.unesco.org/ark:/48223/pf0000221282. Azomahou, T. T., and E. A. Yitbarek. 2016. “Intergenerational Education Mobility in Africa: Has Progress Been Inclusive?” World Bank Policy Research Working Paper 7843, World Bank, Washington, DC. Baird, S., F. H. Ferreira, B. Özler, and M. Woolcock. 2014. “Conditional, Unconditional and Everything In Between: A Systematic Review of the Effects of Cash Transfer Programmes on Schooling Outcomes.” Journal of Development Effectiveness 6 (1): 1–43. https://doi.org/10.1080/19439342.2014.890362. Bennett, J. 2012. Roma Early Childhood Inclusion: Overview Report. Open Society Foundations. Brătucu, T. O., G. Brătucu, I. B. Chițu, and L. Dovleac. 2018. “Influence of Students Mentality Towards Gender on the Development of Romanian Women’s Career.” Journal of Smart Economic Growth 3 (2): 7–12. Breda, T., E. Jouini, C. Napp, and G. Thebault. 2020. “Gender Stereotypes Can Explain the Gender- Equality Paradox.” Proceedings of the National Academy of Sciences 117 (49): 31063–69. Bruck, Kjerstin A., and Åsa Cater. 2016. “Men for Gender Equality: Promising Projects for Preventing Men’s Violence against Women.” In Preventie van intergenerationeel geweld in Nederland en EU: Verkenning van wat werkt. Bijlagen, edited by Majone Steketee, Reneé Römkens, Eliane Smits van Waesberghe, Trees Pels, Katinka Lünnemann, Jodi Mak, Jamila Mejdoubi, and Hanna Harthoorn, 51– 63. Atria & Verwey-Jonker Instituut. Burger, K. 2015. “Effective Early Childhood Care and Education: Successful Approaches and Didactic Strategies for Fostering Child Development.” European Early Childhood Education Research Journal 23 (5): 743–60. 53 Burger, K. 2016. “Intergenerational Transmission of Education in Europe: Do More Comprehensive Education Systems Reduce Social Gradients in Student Achievement? Research in Social Stratification and Mobility 44: 54–67. https://doi.org/10.1016/j.rssm.2016.01.001. Carlana, M. 2019. “Implicit Stereotypes: Evidence from Teachers’ Gender Bias.” The Quarterly Journal of Economics 134 (3): 1163–1224. Checchi, D., C. V. Fiorio, and M. Leonardi. 2013. “Intergenerational Persistence of Educational Attainment in Italy.” Economics Letters 118 (1): 229–32. Ciolca, S. M. 2020. “School Dropout in Vaslui County: Causes, Risk Situations.” In Book of Proceedings of the 10th International Congress of Physical Education, Sport and Kinetotherapy, edited by Monica Stănescu, Manon Popescu, and Marius Stoicescu, 12–23. Discobolul Publishing House. Chisamya, G., J. DeJaeghere, N. Kendall, and M. A. Khan. 2012. “Gender and Education for All: Progress and Problems in Achieving Gender Equity.” International Journal of Educational Development 32 (6): 743–55. Cismaru, D. M., and N. Corbu. 2019. “The Multiple Impacts of Education Gaps in Romania.” In Development in Turbulent Times: The Many Faces of Inequality Within Europe, edited by Paul Dobrescu, 171–83. Springer. Ciucă, R. E. 2019. “The Bullying Phenomenon in Romanian Schools.” Revista Universitară de Sociologie 15 (1): 240–50. Concordă, E. 2018. “Gender Stereotypes in School Textbooks.” Revista Românească pentru Educaţie Multidimensională 10 (4): 65–81. Connell, R. W. 2003. “The Role of Men and Boys in Achieving Gender Equality.” United Nations, Division for the Advancement of Women. Demirel, M., and C. Okten. 2020. “Gender Gap in Intergenerational Educational Persistence: Can Compulsory Schooling Reduce It?” IZA Discussion Paper No. 13362, IZA (Institute of Labor Economics). Diaconescu, S., T. Ciuhodaru, C. Cazacu, L. Z. Sztankovszky, C. Kantor, and M. Iorga. 2015. “Teenage Mothers, an Increasing Social Phenomenon in Romania. Causes, Consequences and Solutions.” Revista de Cercetare si Interventie Sociala 51: 162–75. Dongarwar, Deepa, and Hamisu M. Salihu. 2019. “Influence of Sexual and Reproductive Health Literacy on Single and Recurrent Adolescent Pregnancy in Latin America.” Journal of Pediatric and Adolescent Gynecology 32 (5): 506–13. Drob, Eliza. 2020. “Gender-Based Violence in the Poorest County of European Union - Vaslui County in Romania.” World Vision Romania. https://www.wvi.org/stories/romania/gender-based-violence- poorest-county-european-union-vaslui-county-romania. Dryler, H. 1998. “Parental Role Models, Gender and Educational Choice.” British Journal of Sociology 49 (3): 375–98. Edström, J., A. Hassink, T. Shahrokh, and E. Stern (eds.). 2015. Engendering Men: A Collaborative Review of Evidence on Men and Boys in Social Change and Gender Equality. EMERGE: Engendering Men—Evidence on Routes to Gender Equality. 54 https://opendocs.ids.ac.uk/opendocs/bitstream/handle/20.500.12413/7059/EMERGE%20Evidence_ Report_final.pdf?sequence=1&isAllowed=y. EIGE. 2022. “Romania.” https://eige.europa.eu/gender-mainstreaming/countries/romania. EPF. 2023. “European Contraception Policy Atlas—Romania.” https://www.epfweb.org/node/747. EU-LFS. 2019. “Employment and Unemployment (LFS) — Overview.” https://ec.europa.eu/eurostat/web/lfs. European Commission. 2022. “Romania: Political, Social and Economic Background and Trends.” https://eurydice.eacea.ec.europa.eu/national-education-systems/romania/political-social-and- economic-background-and-trends. European Commission. 2023. “Eurydice”. https://eurydice.eacea.ec.europa.eu/national-education- systems/romania/overview Eurostat. 2017. “Teenage and Older mothers in the EU.” . Eurostat. 2023. “Legally induced abortions by mother's age.” https://ec.europa.eu/eurostat/databrowser/view/DEMO_FABORT__custom_7397075/default/table ?lang=en Farré, Lídia. 2012. “The Role of Men for Gender Equality.” Working Paper, World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/9237. Fessler, P., P. Mooslechner, P., and M. Schürz. 2012. “Intergenerational Transmission of Educational Attainment in Austria.” Empirica 39: 65–86. Fry, Richard, Brian Kennedy, and Cary Funk. 2021. “STEM Jobs See Uneven Progress in Increasing Gender, Racial and Ethnic Diversity.” PewResearch.org. https://www.pewresearch.org/science/2021/04/01/stem-jobs-see-uneven-progress-in-increasing- gender-racial-and-ethnic-diversity/. Gavrilescu, A. S., and F. Merloiu. 2021. “The Bullying Phenomenon in Romanian Schools.” New Trends in Psychology 3 (2): 34–44. Gheorgiu, Maria, Leslie Hawke, Joost de Laat, and Catalina Papari. 2021. “Using Evidence and Coalitions to Scale-Up a National Early Education Initiative: The Case of Law 248/2015 in Romania.” Frontiers in Public Health 8. Goldfarb, Eva S., and Lisa D. Lieberman. 2020. “Three Decades of Research: The Case for Comprehensive Sex Education.” Journal of Adolescent Health 68 (1): P13–27. https://doi.org/10.1016/j.jadohealth.2020.07.036. Goldin, C. 1994. “The U-Shaped Female Labor Force Function in Economic Development and Economic History.” In Investment in Women’s Human Capital and Economic Development, edited by T. P. Schultz, 61–90. Chicago: University of Chicago Press. Gray, C., and H. Leith. 2004. “Perpetuating Gender Stereotypes in the Classroom: A Teacher Perspective.” Educational Studies 30 (1): 3–17. Greulich, A., and A. Dasré. 2018. “Observing the Number of Children with EU-SILC: A Quantification of Biases.” Population 73(4): 685–720. 55 Gyaraki, Réka. 2022. “Education Challenges in Romania 2022.” Brokenchalk.org. https://brokenchalk.org/education-challenges-in-romania-2022/. Hall, William J., Benjamin L. H. Jones, Kristen D. Witkemper, Tora L. Collins, and Grayson K. Rodgers. 2019. “State Policy on School-based Sex Education: A Content Analysis Focused on Sexual Behaviors, Relationships, and Identities.” American Journal of Health Behavior 43 (3): 506–19. https://doi.org/10.5993/AJHB.43.3.6. Hamby, McKenzie. 2018. “Addressing the Barriers to a Complete Education in Romania.” borgenmagazine.com. https://www.borgenmagazine.com/education-in-romania/. Hausmann, R., L. Tyson, and S. Zahidi. 2008. The Global Gender Gap Report 2009. Geneva, Switzerland: World Economic Forum. Hammond, Alicia, Eliana Rubiano Matulevich, Kathleen Beegle, and Sai Krishna Kumaraswamy. 2020. The Equality Equation: Advancing the Participation of Women and Girls in STEM. The International Bank for Reconstruction and Development and The World Bank. https://doi.org/10.1596/34317. Hanushek, E. A., and L. Wößmann. 2006. “Does Educational Tracking Affect Performance and Inequality? Differences-in-Differences Evidence across Countries.” The Economic Journal 116 (510): C63–C76. https://doi.org/10.1111/j.1468-0297.2006.01076.x Heilman, Brian. 2018. “Breaking Gender Stereotypes: How Promundo Is Working to Shift Harmful Gender Norms Globally.” Preventconnect.org. https://www.preventconnect.org/wp- content/uploads/2018/07/2018_PC_ConferenceSummaries-Promundo.pdf?x47887. Horga, M., C. Gerdts, and M. Potts. 2013. “The Remarkable Story of Romanian Women’s Struggle to Manage their Fertility.” Journal of Family Planning and Reproductive Health Care 39 (1): 2–4. Huang, J. 2013. “Intergenerational Transmission of Educational Attainment: The Role of Household Assets.” Economics of Education Review 33: 112–23. ILO (International Labour Office) and UNICEF (United Nations Children’s Fund). 2021. Child Labour: Global Estimates 2020, Trends and the Road Forward. New York: ILO and UNICEF. INS (National Institute of Statistics). 2019. “Household Budget Survey 2019.” International Trade Administration. 2023. “Romania—Country Commercial Guide.” https://www.trade.gov/country-commercial-guides/romania-education-and-training-services. Ioana, M. I., M. I. Anda, P. Cornelia, and C. R. Mariana. 2015. “School Dropout–A Social Problem in Romania.” Procedia-Social and Behavioral Sciences 182: 623–28. Iorga, M., L. M. Pop, N. Gimiga, L. Păduraru, and S. Diaconescu. 2021. “Assessing the Opinion of Mothers about School-Based Sexual Education in Romania, the Country with the Highest Rate of Teenage Pregnancy in Europe.” Medicina (Kaunas) 57 (8): 841. Iorga, M., R. V. Socolov, and D. G. Socolov. 2016. “An 8 Years Analysis of Pregnancies and Births among Teenagers in a University Hospital in North-Eastern Romania.” Revista de Cercetare si Interventie Sociala 54. Islam, K. M. M., and M. N. Asadullah. 2018. “Gender Stereotypes and Education: A Comparative Content Analysis of Malaysian, Indonesian, Pakistani and Bangladeshi School Textbooks.” PloS One 13 (1): e0190807. 56 K12 Academics. “Kindergarten Education in Romania.” K12academics.com. https://www.k12academics.com/Education%20Worldwide/Education%20in%20Romania/kindergart en-education-romania. Kricorian, K., M. Seu, D. Lopez, E. Ureta, and O. Equils. 2020. “Factors Influencing Participation of Underrepresented Students in STEM Fields: Matched Mentors and Mindsets.” International Journal of STEM Education 7: 1–9. Lange, S., and M. von Werder. 2017. “Tracking and the Intergenerational Transmission of Education: Evidence from a Natural Experiment.” Economics of Education Review 61: 59–78. https://doi.org/10.1016/j.econedurev.2017.10.002. Levtov, R. G., G. Barker, M. Contreras-Urbina, B. Heilman, and R. Verma, R. 2014. “Pathways to Gender-Equitable Men: Findings from the International Men and Gender Equality Survey in Eight Countries.” Men and Masculinities 17 (5): 467–501. Loshkin, Michael, and Monica Fong. 2006. “Women’s Labour Force Participation and Child Care in Romania.” The Journal of Development Studies 42(1): 90–109. MEB Primary Education General Directorate and UNICEF Turkey. 2011. Late Enrollment in Primary Education: Causes and Recommendations for Prevention. https://www.unicef.org/turkiye/media/1046/file/TURreport_lateenrollment.pdf.pdf. Mihaileanu, Liviu. 2018. “The Condition of Children under Romania’s Child Protection System.” Romania Without Orphans Alliance. Nadel, H., M. Spellmann, T. Alvarez-Canino, L. Lausell-Bryant, and G. Landsberg. 1996. “The Cycle of Violence and Victimization: A Study of the School-Based Intervention of a Multidisciplinary Youth Violence-Prevention Program.” American Journal of Preventive Medicine 12 (5): 109–19. OECD. 2020. Public Spending on Childcare and Early Education (accessed on March 21, 2023). https://www.oecd.org/els/soc/PF3_1_Public_spending_on_childcare_and_early_education.pdf. OECD. 2022. Net Childcare Costs in EU Countries, 2021. https://www.oecd.org/els/soc/benefits-and- wages/Net%20childcare%20costs%20in%20EU%20countries_2021.pdf. OECD. 2023. “Gender Wage Gap” (indicator) (accessed March 9, 2023). doi: 10.1787/7cee77aa-en. Oliva, A., A. Parra, and I. Sánchez-Queija. 2019. “Intergenerational Transmission of Family Work Values and Adolescents’ Educational Aspirations and Expectations in Spain: A Longitudinal Study.” In Work and Family Interface in the International Career Context, edited by A. Parra and A. Oliva, 215–33. Springer International Publishing. https://doi.org/10.1007/978-3-030-17498-9_13. Oyvat, C., and Ö. Onaran. 2022. “The Effects of Social Infrastructure and Gender Equality on Output and Employment: The Case of South Korea.” World Development 158: 105987. Papadakis, S. 2018. “Gender Stereotypes in Greek Computer Science School Textbooks.” International Journal of Teaching and Case Studies 9 (1): 48–71. Patache, L., F. L. Ghencea, and O. Negurita. 2022. “Identifying Constraints on Roma Minority Education Povision in Romania.” Technological and Economic Development of Economy 28 (1): 836–48. https://doi.org/10.3846/tede.2022.16381. 57 Plummer, M., and A. Cossins. 2018. “The Cycle of Abuse: When Victims Become Offenders.” Trauma, Violence, & Abuse 19 (3): 286–304. Pratama, Rafly G. 2022. “Student Council: 3 out of 10 Romanian High School Students Witnessed Acts of Violence, Discrimination.” Romania-Insider.com. https://www.romania-insider.com/council- students-witness-acts-violence-discrimination. Rada, C. 2014. “Family Planning in Romania: A Cross-Sectional Study of Knowledge, Practice and Attitudes.” International Journal of Collaborative Research on Internal Medicine & Public Health 6 (9): 260–75. Radu, M. C., L. S. Manolescu, R. Chivu, C. Zaharia, C. Boeru, M. E. Pop-Tudose, A. Necsulescu, and M. Otelea. 2022. “Pregnancy in Teenage Romanian Mothers.” Cureus 14 (1): e21540. RDCC (Romanian Diversity Chamber of Commerce). 2021. The Absence of Women in STEM in Romania Is a Missed Opportunity. The absence of women in STEM in Romania is a missed opportunity (video) - RDCC Robayo-Abril, Monica, and Britta Rude. Forthcoming. Preparatory School Years and Maternal Employment in Romania. Romania-Insider. 2020. “Save the Children Romania: Almost a Quarter of Teenage Mothers in the EU Live in Romania.” Romania-Insider.com. https://www.romania-insider.com/teenage-mothers- romania-eu. Romania Journal. 2019. “Romania Ranks Third in Europe on Bullying in Schools.” https://www.romaniajournal.ro/society-people/romania-ranks-third-in-europe-on-bullying-in- schools/. Romania Journal. 2022a. “Students In Romania Are At a Minimum Functional Level of Digital Skills, Report Says.” https://www.romaniajournal.ro/society-people/education/students-in-romania-are- at-a-minimum-functional-level-of-digital-skills-report-says/. Romania Journal. 2022b. “Study: 76% Of Children Are Quarreled at School In Romania, 56% of Parents Punish Their Children.” https://www.romaniajournal.ro/society-people/study-76-of-children-are- quarreled-at-school-in-romania-56-of-parents-punish-their-children/. Rotaru, Ileana. 2019. Case Study. School Dropout of Roma in Romania: Between Antigypsyism, the Lack of Resources and the Lack of Interest of the Authorities. Nevo Parudimos Association. https://ergonetwork.org/wp-content/uploads/2019/11/School-dropout_Nevo-Romania.pdf. Rumberger, R. W. 2001. “Why Students Drop Out of School and What Can Be Done.” UCLA: The Civil Rights Project / Proyecto Derechos Civiles. https://escholarship.org/uc/item/58p2c3wp. Schneebaum, A., B. Rumplmaier, and W. Altzinger. 2016. “Gender and Migration Background in Intergenerational Educational Mobility.” Education Economics 24 (3): 239–60. Selander, Morgan, and Emily Walter. 2020. “Lack of Educational Opportunities for the Roma People in Eastern Europe.” Ballard Brief. www.ballardbrief.org. Sovič, A., and V. Hus. 2015. “Gender Stereotype Analysis of the Textbooks for Young Learners.” Procedia-Social and Behavioral Sciences 186: 495–501. 58 Stotsky, M. J. G. 2016. Gender Budgeting: Fiscal Context and Current Outcomes. International Monetary Fund. Vdovîi, Lina, and Michael Bird. 2019. “Over 30 Percent of Hospitals in Romania Are Refusing Legal Abortions.” Theblacksea.eu. https://theblacksea.eu/stories/quarter-hospitals-romania-are-refusing- legal-abortions/. UNESCO. 2020. Global Education Monitoring Report—Gender Report: A New Generation: 25 years of Efforts for Gender Equality in Education. Paris: UNESCO. https://unesdoc.unesco.org/ark:/48223/pf0000374514/PDF/374514eng.pdf.multi. UNESCO. 2022. Leave No Child Behind: Global Report on Boys’ Disengagement from Education. Leave no child behind: global report on boys’ disengagement from education - UNESCO Digital Library. UNESCO. 2023. “CSE Implementation at Regional and Country Levels.” https://csetoolkit.unesco.org/toolkit/getting-started/cse-implementation-regional-and-country- levels. UNGEI. 2018. Guidance for Developing Gender-Responsive Education Sector Plans. https://www.ungei.org/publication/guidance-developing-gender-responsive-education-sector-plans. UNGEI. 2023. “Gender-Responsive Education Systems by Design.” https://www.ungei.org/what-we- do/gender-responsive-education-systems. UNICEF. 2023. “Girls’ Education.” https://www.unicef.org/education/girls- education#:~:text=Girls%20who%20receive%20an%20education,strengthens%20economies%20and %20reduces%20inequality. United Nations Department of Economic and Social Affairs, Population Division (2022). World Contraceptive Use 2022. Link: https://www.un.org/development/desa/pd/data/world-contraceptive- use Unterhalter, E., A. North, M. Arnot, C. Lloyd, L. Moletsane, E. Murphy- Graham, J. Parkes, and M. Saito. 2014. Interventions to Enhance Girls Education and Gender Equality. Education Rigorous Literature Review. Department for International Development FID Research for Development. Verdonk, P., H. Seesing, H., and A. De Rijk. 2010. “Doing Masculinity, Not Doing Health? A Qualitative Study among Dutch Male Employees about Health Beliefs and Workplace Physical Activity.” BMC Public Health 10: 1–14. Warner, L., and J. Barerra, 2005. “The Role of Educators in Communicating Gender Equity to Parents.” In Gender and Schooling in the Early Years, edited by Janice Koch and Beverly J. Irby, 127–48. Greenwich, CT: Information Age Publishing. WHO Regional Office for Europe and BZgA. 2010. Standards for Sexuality Education in Europe. https://www.bzga-whocc.de/fileadmin/user_upload/BZgA_Standards_English.pdf Wolbrecht, C., and D. E. Campbell. 2007. “Leading by Example: Female Members of Parliament as Political Role Models. American Journal of Political Science 51 (4): 921–39. World Bank. 2019. Romania. Early Childhood Development. Saber Country Report 2019. Washington, DC: World Bank. 59 World Bank. 2022. “Invest in Childcare.” Washington, DC: World Bank. https://www.worldbank.org/en/topic/education/brief/investinchildcare World Bank. Forthcoming. Benchmarking Exercise. World Economic Forum. 2022. Global Gender Gap Report 2022. Romania. https://www.weforum.org/reports/global-gender-gap-report-2022/in-full/economy-profiles- 5b89d90ea5. World Population Review. 2023. “Education Rankings by Country 2023.” https://worldpopulationreview.com/country-rankings/education-rankings-by-country. World Values Survey. 2023. “World Values Survey.” https://www.worldvaluessurvey.org/wvs.jsp. 60 Appendix 1. Detailed Table on Diagnostic, Constraints, Interventions, and Recommendations Table A.1. Diagnostic and Recommendations High-Level Policy Evidence on Gaps Barriers Tailored Intervention Recommendations Gender gaps (a) Availability: Potential limited availability of formal public childcare centers as legal entities and increasing role of childcare provided by private institutions; negative trends in the number of available creches and (a) Undertake detailed and updated gender-sensitive nurseries and kindergartens as legal entities, both in demand and supply assessments of formal childcare and ECE rural and urban areas; (b) Affordability constraints (1) Study and address demand and services to design tailored interventions to increase access; ECD: Very low net enrollment rates in ECD for both boys among the poor, with gross and net childcare costs (as supply side constraints to access to (b) provide financial incentives to support enrollment among and girls (zero- to two-year-olds ) proprotion of women’s median full-time earnings) ECE for the zero- to two-year-olds the poor; (c) increase public spending, quality, and available below the EU average of 14 percent in Romania, but infrastructure in the area of ECE (interventions to tackle higher among the poor; (c) Acceptability: Parental acceptability) workforce participation and willingness to use institutional care--a large share of children ages less than three years are cared for only by their parents (76.8 percent), the largest in the EU Based on international evidence, under limited (2) Gender-sensitive policies (a) Alleviation of financial constraints through targeted ECE: Gender gaps in net enrollment rates of three- to resourcesfamilies might prefer to put boys into school targeting access, supply, and usage intervention strategies; (b) Study of girls facing additional five-year-olds for all groups (UNICEF 2011) of ECE constraints to ECE (a) Norm-based interventions tackling social norms that constrain boys’ and girls’ educational participation; (b) Interventions that target the awareness of the importance of educational attainment among parents and children (UNESCO 2014); (c) Teacher training on gender stereotypes and marginalization of other subgroups of the population Primary and Secondary: Negative trends for boys and Based on previous evidence (UNESCO 2014), there are (such as the poor or Roma) (UNESCO 2014); (d) School-based, girls. Gross enrollment rates are below the EU average three factors behind children being out of school and at violence-preventing interventions and elimination of violence for boys and girls, and have fallen over time in the case risk of dropping out in Romania: (a) sociocultural (3) Implement policies that target at schools (Anderson et al. 2011); (e) Increased public of primary and secondary education; school dropout is demand-side factors; (b) economic demand-side school dropouts and low spending and investment in educational quality and relatively high; factors; (c) supply-side barriers--we argue that each of enrollment rates attractiveness (UNESCO 2014); (f) Reentry programs and Performance drops for boys and girls according to PISA these factors affects boys and girls asymmetrically, childcare services for young mothers (Robayo-Abril and Rude results leading to (reversed) gender gaps forthcoming) and those that address teenage pregnancy; (g) Addressing of marginalization and discrimination faced by Roma boys and girls, which plays a crucial role in school dropout in Romania (UNESCO 2014) and requires bottom-up approaches (Andrei et al. 2015). Teacher training on the sensitivities around marginalization and discrimination of Roma students is also crucial (UNESCO 2014); (h) Previous studies show that (conditional) cash transfer programs lead to improved educational outcomes (Baird et al. 2014) (a) Facilitation of interactions with matched-background mentors (Krikorian et al. 2020); (b) Norm-based interventions that target stereotypes; (c) partnering with the private sector, addressing gender biases in learning materials and Educational segregation in the field of study and gender among parents, and encouraging girls’ participation in gaps in digital literacy; girls underperform boys in math Lack of role models and gender stereotypes extracurricular activities (Hammond et al. 2020); (d) while boys underperform girls in reading (4) Address gender segregation in Implement strategies that address the perpetuation of education: Incentivize and invest in gender stereotypes in the educational system, such as the the development of girls’ skills and introduction of bias-free schoolbooks, instrumentalization of interests in STEM and ICT and of the media, gender-responsive teacher training, and parental boys' skills in reading and female- workshops (Bruck and Cater 2016; Concordă 2018; dominated fields of study Unterhalter et al. 2014; UNICEF 2023; Farré 2012) Gender gaps are multidimensional: they differ across Severity of constraints differs across groups: The (5) Understand gender equality as a Targeted interventions that factor in the multidimensionality income groups and between rural and urban areas persistence of social norms that could limit boys' and multidimensional subject of (reversed) gender gaps are better suited girls' access to education differs across groups, as do the available infrastructure and economic conditions Reversed gender gaps Complex and multiple constraints : There is evidence of gender norms that could lead to children's leaving Follow international recommendations around the school early; international evidence also points to the Boys underperform in several educational outcomes, (6) Implement Policies that address underachievement and low enrollment of boys and roll out role of labor market demands, household needs to such as learning-poverty rates reversed gender gaps interventions in the macro-, meso-, and micro environment generate income, and boys' being asymmetrically (UNESCO 2022) affected by school-based violence and bullying (UNESCO 2022). Additional gaps Invest in research to identify what works best, for example Knowledge gaps: Lack of systematic, accessible, updated No systematic knowledge on what works best to (7) Fund gender-sensitive research via impact evaluations and rigorous monitoring and studies on what works best address the multidimensional (reversed) gender gaps projects and M&E initiatives evaluation strategies, in collaboration with the academic community and technical specialists Data gaps: Roma children perform worse on educational (8) Generate systematic Include ethnicity parameters in surveys; generate No updated, representative, systematic information on outcomes compared to non-Roma children, but gender- information on Roma children administrative data on subgroups and make information educational outcomes of Roma children sensitive information is missing compared to non-Roma children accessible to the public Conduct a gender-sensitive public expenditure review of the Relatively low public spending on education (as a (9) Increase public spending on Relatively low public spending in the educational sector educational sector and invest money where it is most percent of GDP) when compared to the EU average education at all education level needed, applying gender-budgeting tools Gender-responsive education sector planning (GRESP ) is an 10) Develop gender-responsive efficient way to engrave gender equality into the heart of Overall, gender inequalities in the educational sector in education sector planning (GRESP) national education systems and consequently address gender No gender-sensitive strategies to date Romania exist and an operationalized gender- equality by design (UNGEI 2023). An operationalized gender- sensitive assessment tool sensitive assessment tool could help to track educational gender gaps more closely 61 Appendix 2. Additional Regression Tables Table A.2. Results with B40 and gender as the interaction terms (1) VARIABLES Enrolled Age -0.0394*** (0.000828) Partner in HH 0.0276** (0.0125) No. of HH members with max. primary schooling -0.0195*** (0.00616) No. of HH members with secondary schooling -0.00592*** (0.00216) No. of HH members with tertiary education 0.0106* (0.00614) HH members 0.000193 (0.00226) No. of working HH members -0.0151*** (0.00388) No. of unemployed HH members -0.0419*** (0.0100) Bottom 40 -0.0107 (0.00760) Urban -0.00540 (0.00664) Female = 1 -0.235*** 1.b40#1.female 0.0142** (0.00672) Constant 0.926*** (0.0153) Observations 22,482 R-squared 0.199 Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. 62 Table A.3. Intergenerational persistence of education by gender (1) VARIABLES Enrolled Female -0.228*** (0.00529) Age -0.0394*** (0.000828) Partner in HH 0.0343*** (0.0126) No. of HH members with max. primary schooling -0.0218*** (0.00673) No. of HH members with secondary schooling -0.00595*** (0.00216) No. of HH members with tertiary education 0.0116* (0.00653) HH members 0.000266 (0.00226) No. of working HH members -0.0151*** (0.00388) No. of unemployed HH members -0.0418*** (0.0100) Bottom 40 -0.00944 (0.00709) Urban -0.00531 (0.00664) 1.female#c.hhnbhigher -0.0139** (0.00622) 1.female#c.hhnblower 0.0245*** (0.00669) Constant 0.925*** (0.0152) Observations 22,482 R-squared 0.199 Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. 63