Report No: . World Bank Developing a methodology for measuring the impact of hosting, protecting and assisting refugees (Phase II) The Global Cost of Inclusive Refugee Education: 2023 Update . 06 December 2023 . Poverty and Equity . . © 2023 The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved This work is a product of the staff of The World Bank. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of the Executive Directors of The World Bank 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. Attribution—Please cite the work as follows: “World Bank. 2023. The Global Cost of Inclusive Refugee Education: 2023 Update. © World Bank.� All queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. This work is a product of the staff of The World Bank. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of the Executive Directors of The World Bank 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. All queries on rights and licenses, including subsidiary rights, should be addressed to ksilva2@worldbank.org. 2 Table of Contents ACKNOWLEDGEMENTS ....................................................................................................................................................4 EXECUTIVE SUMMARY.....................................................................................................................................................5 INTRODUCTION..............................................................................................................................................................7 METHODOLOGY...........................................................................................................................................................10 SUMMARY OF FINDINGS ................................................................................................................................................13 INVESTING IN INCLUSIVE REFUGEE EDUCATION ..................................................................................................................19 3 Acknowledgements This paper represents a joint effort by the World Bank FCV Group and Education Global Practice and UNHCR. The core team was composed by Karishma Silva, Johannes Hoogeveen, and Dina Abu-Ghaida, under the guidance of Xavier Devictor (Advisor – Fragility, Conflict and Violence, World Bank). The team benefitted from the leadership of Sajjad Malik (Director for the Division of Resilience and Solutions, UNHCR) and the support of UNHCR including Rebecca Telford, Theresa Beltramo, and Betsy Lippman. The team is grateful for input from host country governments represented at the technical workshops on Measuring the Impact of Hosting, Protecting and Assisting Refugees held between 2019 and 2023. The team is grateful to reviewers Meskerem Mulatu, Harry Patrinos, Kevin Carey, Noah Yarrow, Thomas Poulsen, Caroline Sergeant, Robert Hopper, Gabriel Demombynes, and other members of the World Bank Fragility, Conflict and Violence Group (FCV) and Education Global Practice teams that participated in discussions on the paper. The World Bank team gratefully acknowledges support from the State and Peacebuilding Fund, funded through the Multi- Donor Trust Fund for Forced Displacement (FDTF) administered by the World Bank. Questions about this report should be directed to Karishma Silva at the World Bank (ksilva2@worldbank.org). Executive Summary The global refugee population has increased by over one-third between 2019 and 2022, mainly driven by crises in Ukraine and Afghanistan. Low- and middle-income countries host over three-quarters of all refugees. Among these refugees are 15 million school-aged children, many of whom will spend a considerable portion of their schooling years in forced displacement. Education is critical for refugees, host countries, and countries of origin, empowering refugees with the knowledge and skills needed to rebuild their lives, contribute to their host country’s economy, and support the sustainable return to and peaceful reconstruction of their home countries. It also helps reduce vulnerabilities associated with child labor, early marriage, and violence, and provides stability for children overcoming the trauma of forced displacement. Inclusion in national education systems of host countries is the most sustainable solution for educating refugees. Inclusion means that refugee children attend schools that are part of the host country’s national public system, funded through government channels, and attended by both refugee and host community children together. While many host countries have adopted inclusive education policies, they continue to face implementation challenges primarily because of inadequate multi-year predictable financing. As a result, nearly half of all school-age refugee children are out of school, with lower enrollment rates compared to host populations, especially at higher grade levels. The 2018 Global Compact on Refugees (GCR) places predictable and equitable responsibility-sharing at the center of the international protection agenda. A key step towards operationalizing responsibility-sharing for refugee education is to have an estimate of the costs to host countries. In 2019, the World Bank and UNHCR, through a participatory process with host governments and donors, developed a simple and transparent methodology to estimate these costs. In preparation for the 2023 Global Refugee Forum, this note provides an update to the global education costing exercise done in 2019 building on the same methodology.1 In 2019, the paper reported the fiscal cost of inclusive refugee education as an average across thirteen years of the K-12 education cycle to account for multi-year financing needs but assumed that the refugee population would remain constant. However, since 2019, refugee populations have increased by a third driven mainly by crises in Ukraine and Afghanistan. Many new hosting countries are middle-income countries where the cost of refugee education is substantially higher than in low-income countries.2 To reflect these new trends, this paper reports the total annual fiscal cost of inclusive refugee education for a single year without making any assumptions about changes in refugee populations. The estimated annual fiscal cost of educating all school-age refugee children hosted in low- and middle- income countries through national education systems is approximately US$9.3 billion. The fiscal cost of inclusive refugee education for Ukrainian children in middle-income host countries is around US$1.4 billion or 15 percent of the total costs. While the average unit cost of refugee education varies by host country income group and level of education, the total annual fiscal cost of educating all refugees in low- income countries, is only US$309 million. Delivering this level of financing could reach over a quarter of all refugee children. The middle-income host countries account for the rest of the total financing envelope or approximately US$9 billion. 1 These estimates provide a normative assessment of the fiscal costs of inclusive refugee education and do not represent current expenditure on refugee children in host countries. 2 Refugees hosted in high income countries are excluded from this exercise. 5 To meet these financing needs, renewed discussions around responsibility-sharing are needed. Most donor allocations to refugee education flow through humanitarian systems and respond to each refugee crisis as its own event. What is required is a more structural framework that is rule-based, ensuring transparency and equity across refugee crises as well as contributing to strengthening host country education systems for the benefit of all, as envisaged under the GCR. Ultimately, sharing the costs of refugee education will require negotiation between host governments and partners guided by objective criteria. For example, host country contributions could be determined based on the estimated costs of refugee education as a share of public education expenditure. If the international community paid for any refugee education costs exceeding 1 percent of public education expenditure in lower middle-income countries and 2 percent in upper-middle income countries, with the host country contributing the rest, total financing required from the international community would fall from US$9 billion to US$5.7 billion. Host governments already contribute to the recurrent costs of refugee education in most countries where refugees are included in national education systems. Clearly identifying these contributions can further support negotiations with donor partners. Increased commitments of financing from donors and host governments need to be accompanied by innovations in financing instruments. Where refugees are included in national schools, the financing instruments would allow for funding to flow through government systems and to address the bulk of inclusive education costs which are recurrent and multi-year. Further, any financing for refugee education should be provided as an addition to current donor allocations for each host country so that refugee education investments are not made at the expense of host country investments in its own population. Finally, the share of external financing should be guided by the host country’s needs and ability to pay. Current financing instruments come woefully short of meeting these financing needs and considerable innovation is needed. Providing host governments with adequate multi-year, predictable financing will allow them to operationalize their commitments to inclusive refugee education. To ensure the sustainability and efficiency of investments in refugee education, it is essential for financing to align with the shift toward inclusive education policies. This includes increasing developmental financing for refugee education and improving financing instruments to meet medium-terms recurrent needs. This report highlights the importance of the donor community and host countries engaging in discussions about responsibility-sharing and financing mechanisms, especially in middle-income countries, where the responsibility for covering costs is less clear. Financing refugee education is a critical challenge that requires a shift from short-term humanitarian financing to long-term developmental financing, with a focus on integrating refugee children into national education systems. Such a transition requires careful planning, policy dialogue, and collaboration between donors, humanitarian and developmental partners, and host governments. 6 Introduction The global refugee population has been increasing at an alarming rate in recent years. When the first edition of the Global Cost of Inclusive Refugee Education report was released, the total refugee population stood at 26 million in 2019, relatively stable from the previous year. In 2022, that figure rose to 35.3 million,3 representing a one-third increase in just three years. This was largely driven by crises in Ukraine and Afghanistan. The number of Venezuelans displaced abroad also increased, from 3.6 million in 2019 to 5.2 million in 2022. Among these refugees are 15 million school-aged children.4 With 67 percent of refugees living in protracted situations lasting at least five consecutive years,5 many refugee children will spend a large portion of their schooling years in forced displacement. Over three-quarters of refugees are hosted by low- and middle-income countries where resources are limited and learning poverty is high. These countries host a disproportionately large share of refugees relative to the resources available to them. While low-income countries (LICs) account for only 0.5 percent of global gross domestic product (GDP), they host 16 percent of refugees. In comparison, high-income countries (HICs) account for over 60 percent of global GDP and host less than a quarter of all refugees. LICs have inherently weaker education systems than HICs and thus additional inflows of refugee children can further deteriorate already weak systems. This proportion is much higher than in recent years due to the number Ukrainian refugees hosted in Europe.6 Further, learning poverty, which measures the share of children unable to read and understand a simple text by age 10, was as high as 57 percent7 in LICs and MICs signaling weak education systems.8 Without adequate support, LICs and MICs are ill-equipped to manage the education needs of refugee children and the impact of refugee inflows on host population education outcomes. Enrollment rates of refugees are far lower than those of host populations, with gaps widening with each grade level. Of the 15 million school-age refugee children, more than half are estimated to be out-of- school. Average gross enrollment rates (GER) for refugees stood at 65 percent at the primary level, 41 percent at the secondary level, and 6 percent at the tertiary level.9 This compares to global GERs of 102 percent, 77 percent, and 40 percent at primary, secondary and tertiary levels, respectively.10 While on average there is gender parity with refugee enrollment rates similar for boys and girls at the primary and secondary level, in some host countries with large refugee populations like Kenya and Ethiopia, boys access schooling at much higher rates than girls.11 Yet investing in education is critical for refugees, host countries and countries of origin. Quality education empowers refugees with the knowledge and skills needed to rebuild their lives with dignity. It opens doors to job opportunities resulting in individual returns, increased self-sufficiency and reduced dependency on aid, and the ability to contribute to the economies of the countries that host them. Education 3 UNHCR. 2023. Global Trends: Forced Displacement in 2022. 4 UNHCR. 2023. UNHCR Education report 2023 – Unlocking Potential: The Right to Education and Opportunity. 5 UNHCR. 2023. Global Trends: Forced Displacement in 2022. 6 UNHCR. 2023. Global Trends: Forced Displacement in 2022. 7 This figure is likely to have risen to 70 percent following school closures during the COVID-19 pandemic. Learning poverty in HICs was 8 percent. 8 World Bank. 2022. The State of Global Learning Poverty: 2022 Update. 9 UNHCR. 2023. UNHCR Education report 2023 – Unlocking Potential: The Right to Education and Opportunity. 10 World Bank. 2023. World Development Indicators, https://data.worldbank.org. Gross enrollment rates can exceed 100 percent as a result of over-age students repeating grades. Gross enrollment rates in low and middle- income countries, where most refugee children live, were 102 percent, 74 percent and 37 percent at primary, secondary, and tertiary levels, respectively. 11 UNHCR. 2023. UNHCR Education report 2023 – Unlocking Potential: The Right to Education and Opportunity. 7 can help reduce vulnerabilities associated with child labor, early marriage,12 and sexual violence,13 and can provide a sense of stability and normalcy for children overcoming the trauma of forced displacement, violence, and personal loss. Education is also associated with better health outcomes, including for the next generation where educated women have healthier children. Finally, education is important for peace building, reconstruction, and sustainable return to countries of origin. The cost of not investing in refugee education and youth is potentially very high. As most refugees are displaced for long periods of time, the cost of not integrating refugees into national education systems includes a dead weight loss on potential future earnings of the refugee and income tax collection. Refugees integrated into national education systems are more likely to be prepared to access the labor market and have higher earnings. Aside from building their human capital and ability to make positive economic contributions, inclusive refugee education is also a pathway to a peace dividend with positive social spillovers. Inclusion in national education systems of host countries is the most sustainable solution for educating refugees. Inclusion refers to refugee children attending schools that are (a) part of the host country’s national public system (e.g., curriculum, teachers’ qualifications, and oversight mechanisms); (b) funded through government channels; and (c) attended by both refugee and host community children together.14 Forced displacement situations are increasingly protracted with many refugees, and often multiple generations, unable to return to countries of origin in safety and dignity. Parallel humanitarian solutions provide a stop-gap measure in the immediate onset of a crisis but are not adequately financed to offer long-term solutions. Further, the scale of refugee education needs is too large to be addressed through parallel systems. Despite decades of concerted efforts by humanitarian agencies, more than half of all refugee children remain out of school. Inclusion in national systems is more likely to result in formal accreditation and recognition of qualifications, making it easier for refugees to continue their education and find employment in the host country or their countries of origin. Inclusion is also important to promote social cohesion with host communities, improve equity in the education outcomes of refugees and host populations, and strengthen service delivery for both refugees and host communities in underserved areas where refugees are often located. Inclusion in national education systems can strengthen sustainability and efficiency of investments. Investments in national education systems can help build capacity and strengthen local systems’ ability to respond to new crises. It also ensures that financing allocated to refugee education can benefit both refugees and host community children; and conversely, investments in improving the general quality of education accrue to both refugees and host community children. Further, there can be efficiency gains from reduced transactional costs to implementing partners and from using results-based financing to focus on refugee enrollment and attainment. There is a discernible trend towards the adoption of inclusive education policies, with varying success in implementation. A notable example is the 2017 Djibouti Declaration which demonstrates the commitment of the Intergovernmental Authority on Development (IGAD) country members – Djibouti, 12 Wodon, Q., Male, C., Nayihouba, A., Onagoruwa, A., Savadogo, A., Yedan, A., ... & Petroni, S. 2017. Economic impacts of child marriage: global synthesis report. 13 Evans, D. K., Hares, S., Holland, P. A., & Mendez Acosta, A. 2023. Adolescent Girls’ Safety In and Out of School: Evidence on Physical and Sexual Violence from across Sub-Saharan Africa. The Journal of Development Studies, 59(5), 739-757. 14 UNHCR. 2023. Multistakeholder Pledge: Securing sustainable futures: Towards a shared responsibility to uphold the right to education and include refugee children in national education systems. Available at: https://globalcompactrefugees.org/pledges-contributions/multistakeholder-pledges-2023/multistakeholder- pledge-securing-sustainable. 8 Ethiopia, Eritrea, Kenya, Somalia, Sudan, South Sudan, and Uganda – to establishing inclusive education systems. Several other host countries including Colombia, Ecuador, Brazil, Turkiye, Jordan, Lebanon, and Iran have implemented policies that ensure the right to education for all children regardless of their legal status.15 However, the degree of implementation of these policies can vary widely depending on the host country’s resources, capacity, and the size of the refugee population. Some host countries have completely integrated refugees in national education systems with refugee children accessing public schools alongside host community children and funding directed through government channels (e.g., Turkiye, Colombia, Ecuador, Brazil). In others, most refugee children access schools in camps that have transitioned to the public system (e.g., Chad) or second shifts in public schools that almost exclusively enroll refugee children (e.g., Jordan). In some countries, while refugees may be integrated in public schools, financing is directed through external channels (e.g., Lebanon). While recognizing the realities of this continuum of inclusion, the benefits of inclusion as described above best accrue when refugee children are integrated in public schools attended by refugee and host community children, and financing is directed through government channels. 15 Abu-Ghaida, D. & Silva, K. 2021. Educating the Forcibly Displaced: Key Challenges and Opportunities. 9 Methodology This report builds on the methodology used to estimate the cost of inclusive refugee education developed by the World Bank and UNHCR in consultation with donors and host countries.16 Under the 2018 Global Compact on Refugees, the donor community committed to equitable burden- and responsibility-sharing to ease pressures on hosting countries. An important step towards operationalizing responsibility-sharing is to measure the impact of hosting and assisting refugees on host countries, using a simple, transparent, and commonly agreed methodology. The methodology used in this report was developed in consultation with Member States through technical workshops organized by the UNHCR in 2019 and 2020.17 It relies on official, publicly available data sources where possible to ensure transparency. The purpose of this costing exercise is to provide indicative estimates of the fiscal cost of including all refugee children in national education systems; however, host countries should develop costed plans to obtain more accurate figures based on updated costs of refugee education needs and enrollment targets. These figures provide normative estimates of the fiscal costs of inclusive education for refugees and do not represent current expenditure on refugee education in host countries. Scope of the costing exercise. This exercise only estimates the recurrent costs of inclusive refugee education, and not the associated investment or economic costs.18 It estimates the cost of educating refugees aged 5-17 years in the public education systems of their current host countries with a specific focus on LICs and MICs with 7000 or more refugees.19 This includes 72 LICs and MICs which host approximately three- quarters of all refugee children.20 Since 2019, there has been an increase in the number of middle-income hosting countries. Among these are six UMICs – Moldova, Montenegro, Belarus, Georgia, Paraguay, and Namibia; and four LMICs – Tajikistan, Uzbekistan, Bolivia, and Tunisia. Refugee numbers include refugees, asylum-seekers and Venezuelans displaced abroad, who are registered with UNHCR. The UNHCR refugee population dataset includes age-disaggregation data.21 Refugee children aged 5-11 years are assumed to be uniformly distributed across one year of pre-primary education and six years of primary education. Those aged 12-17 years are assumed to be uniformly distributed across six years of secondary education. The methodology is based on the premise that refugee education is embedded in the host country education system, facing the same cost drivers, and efficiency and quality constraints. The base unit 16 World Bank and UNHCR. 2021. The Global Cost of Inclusive Refugee Education. 17 UNHCR. 2022. Measuring the impact of hosting, protecting and assisting refugees – Progress Report. https://www.unhcr.org/sites/default/files/legacy-pdf/62bc1cc24.pdf. 18 Recurrent costs are the costs borne by host governments and humanitarian agencies on an ongoing basis to ensure that refugees can access education. These include costs associated with teacher salaries, procurement of teaching and learning materials, teacher training, maintenance of school infrastructure, and amortization of investment outlays. Investment costs, for instance construction of schools and facilities, that benefit both refugees and host populations, are not included in this exercise. However, host countries, particularly those with substantial refugee inflows, will incur capital investment costs, and it is important that these are estimated in national education plans and adequately financed. Similarly, economic costs incurred because of the inclusion of refugees in the economy, for instance changes in individual or societal returns from refugee education, social tensions etc. are not included in this exercise. 19 This exercise does not include Palestinian refugees under the mandate of UNRWA. 20 A little over a quarter of all refugee children are hosted in high-income countries. There are 35 other low- and middle-income host countries that collectively host less than 12,000 school-aged refugee children and are excluded from this analysis. 21 Where demographic data is available for less than 20 percent of refugees, age distribution of refugees is assumed to follow the average of other host countries. 10 cost is the per-capita public education expenditure at each level of education,22 calculated as initial public expenditure by level of education divided by total public enrollment at that level of education. This data was obtained from the UNESCO UIS database to ensure transparency. Public expenditure includes both current and capital amortization costs, as well as spending at all levels of education. The costing methodology also takes into account the importance of early childhood education (ECE) for long-term education outcomes and the historically inadequate public spending in the area. As such, an ECE coefficient of 30 percent has been added to the per-capita primary education costs to estimate the per-capita pre-primary base costs.23 A refugee education coefficient is added to the base cost to support refugee-specific education interventions. Considering the distinct educational needs of refugees, including psychosocial support, language assistance, bridging programs, remedial/accelerated education programs, and so on, a mark-up or refugee coefficient is added to the base unit costs to estimate the unit costs of refugee education. As there is no systematic database of the costs of these additional services across host countries, a uniform coefficient is used across all countries. However, host countries can adjust this coefficient based on the unique educational needs of refugees in their countries, as well as the changing costs in the immediate and protracted phases of forced displacement situations. The uniform refugee coefficient is adopted from the EFA-GMR 2015 report figures on the additional cost of reaching marginalized children with education. 24 The coefficients applied are 20 percent for pre-primary and primary education and 35 percent for secondary education. Consequently, the unit cost for refugee education becomes (1 + refugee coefficient) multiplied by the base unit cost at a specific level of education. Analysis of the unit cost differences between host and refugee students in Uganda provides confidence in these refugee coefficient estimates.25 As highlighted in the 2021 global costing report, while adding the refugee coefficient creates a higher unit cost for refugees compared to host country children, this is not perceived as an additional investment in refugees vis-à-vis host country children. Instead, it serves as a mechanism to bridge the initial gap between refugees and local students, thereby “leveling the playing field� as they integrate into national education systems.26 Given limited evidence to date due to few finalized country-level refugee education costed plans, more country level data is needed to validate the refugee coefficient applied in practice. Future exercises in updating this education costing figure would benefit from validating the EFA-GMR coefficient with further education costed plans. The annual cost of inclusive refugee education for each country is then calculated as the unit cost of refugee education by level of education multiplied by the estimated number of school-aged refugee children at that level of education. These costs are then summed across all host countries to derive the total annual cost. The 2021 costing paper reported the average fiscal cost of refugee education across the thirteen-year K-12 education cycle. Starting with the existing cohort of refugee children, it assumed that refugee students transitioned from one grade to the next. This implied that the refugee student population fell each year, with a share of secondary school-aged children transitioning out of the education system each year; and as a result the average cost across the K-12 cycle was lower than the total cost in the first year. However, given that refugee population numbers have been increasing year-on-year and it is not 22 For host countries where public unit cost data was not reported, unit cost was estimated based on government expenditure by level of education as a percentage of GDP per capita. This data was obtained from the UNESCO UIS database or, if unavailable, from the World Bank Development Indicators database. Where data was not available for the last ten years, the average for countries in the same region and income-group was used. 23 World Bank and UNHCR. 2021. The Global Cost of Inclusive Refugee Education. 24 Education for All Global Monitoring report. 2015. Pricing the right to education: the cost of reaching new targets by 2030. Policy Paper 18. Paris: UNESCO. 25 Ministry of Education and Sports. 2018. Education Response Plan for Refugees and Host Communities in Uganda. 26 World Bank and UNHCR. 2021. The Global Cost of Inclusive Refugee Education. 11 possible to accurately predict changes in refugee population, this report focuses on the annual fiscal cost of inclusive refugee education for a single year. These figures are only estimates and rely on incomplete datasets. There are significant gaps in the availability of data on forcibly displaced populations. Not all refugees are registered with UNHCR, as such host country estimates of refugee populations might differ from the UNHCR dataset. Further, demographic data was only available for 76 percent of refugees.27 This data is important for determining the distribution of school-age refugee children across different levels of education. Host country public education expenditure data is collected by UNESCO UIS but is often not reported or outdated. UNESCO and the World Bank impute missing data for the annual Education Finance Watch report, but imputed figures are not verified with governments. The cost of refugee-specific education programs will differ by country and might not match the uniform refugee coefficients. They may differ by geography and evolve over time (e.g., costs of language bridging programs, accelerated or remedial education, mental health, and psychosocial support etc.) Strengthened data systems across all these areas can help improve the accuracy of these estimates. This report is accompanied by a dashboard where host countries can modify or update per student costs, refugee student numbers, refugee coefficients, and targets for refugee enrollment. This dashboard is not a substitute for detailed costed education plans but can be a useful planning tool. 27 UNHCR. 2023. Global Trends: Forced Displacement in 2022. 12 Summary of Findings The total annual fiscal cost of educating all 9 million refugee children through national education systems is estimated to be US$9.3 billion.28 This compares to the estimated annual cost of educating all refugees in 2019 of US$7.1 billion.29 The increase in total annual costs is driven by the increase in school- age refugee population from around 7 million in 2019 to over 9 million in 2022, particularly in MICs. For instance, the cost of inclusive education for Ukrainian refugees in MICs is estimated at US$1.4 billion. Ukrainian refugee children account for around five percent of the total school-aged refugee population in LICs and MICs but nearly 15 percent of the total annual costs. The average estimated unit cost of refugee education fell from US$1,051 to US$1,020, likely as a result of reduced public education expenditure following the COVID-19 pandemic.30 The top 15 host countries by school-age refugee population account for US$7.5 billion or over 80 percent of the total annual cost. Of these host countries, four are UMICs, six are LMICs and five are LICs. Turkey, Lebanon, and Jordan, in response to the Syrian crisis, account for the largest share or US$3 billion of the total annual cost. Colombia and Peru, which host many Venezuelans displaced abroad, account for another US$1.8 billion of the total annual cost. In Pakistan, the total annual cost is estimated at US$209 million and in Bangladesh it is US$67 million.31 Uganda, Sudan, Ethiopia, DRC, Kenya, and Chad collectively host over 2 million school-age refugee children and account for US$334 million of the total annual cost. Finally, the estimated total annual cost for educating refugees in Russia and Iran is US$1.3 billion and US$0.75 billion, respectively (Figure 1). Figure 1. Top 15 host countries – school-age refugee population and total annual costs 1,500,000 2,500 School-age Refugees Total Annual Costs - UMIC 1,200,000 Total Annual Costs - LMIC 2,000 900,000 1,500 600,000 1,000 300,000 500 - - 28 An additional US$457 million is the estimated annual cost of delivering education to refugee children under the UNRWA mandate. UNRWA schools enroll approximately 543,000 children annually at an average unit cost of US$841.50. 29 The estimated annual cost of educating all refugees of US$9.3 billion is significantly larger than the estimated figure in the 2021 Global Cost of Inclusive Refugee Education Report of US$4.85 billion. This difference comes from an increase in the school-age refugee population from 7 million in 2019 to 9 million in 2022, as well as from a change in reporting from annual costs vs. average costs. 30 UNESCO and World Bank. 2023. Education Finance Watch 2023. 31 Bangladesh is not considering integrating refugees into national education systems at this point. This cost estimate reflects the potential fiscal cost of refugee education if the country introduced inclusive education strategies. 13 The average unit cost for refugee education varies by host country income-group and level of education. The average annual unit cost for refugee education is estimated at US$1,020. In LICs, LMICs and UMICs, this figure is US$130, US$661, and US$2,051, respectively. The average unit cost for refugee students in UMICs is nearly 16 times higher than that for LICs and 3 times higher than that for LMICs because of higher levels of public education expenditure in these countries resulting in higher base costs for refugee inclusion in national systems. There are also large variations by level of education, particularly in LICs and LMICs where secondary education unit costs are around double the unit costs for primary education. Table 1 provides the average unit cost of refugee inclusion in national systems by income-group and level of education. Table 1. Average unit costs for refugee inclusion in national systems Average Unit Cost of Inclusive Refugee Education (US$) Pre-primary Primary Secondary Low income 142 109 218 Lower-middle income 628 483 817 Upper-middle income 2,563 1,971 2,205 Low and middle income 1,199 922 1,160 Prioritization of investments in LICs and LMICs will ensure that nearly two-thirds of all school-age refugee children are reached. While LICs and LMICs account for 64 percent of school-age refugee children, their share of the total annual costs of inclusive refugee education is only 28 percent. In fact, the total annual cost of educating nearly 2.4 million refugee children in LICs is only US$309 million (Table 2). Table 2. Distribution of school-age refugee children and annual costs by host country income-category Share of Number of School-Age Total Annual Cost Share of Total Refugee Refugee Children (US$ Millions) Annual Costs Children Low income 2,373,900 309 26% 3% Lower-middle income 3,502,155 2,317 38% 25% Upper-middle income 3,268,411 6,705 36% 72% Low and middle income 9,144,465 9,330 For many host countries, the annual cost of educating all refugee children through national education systems would account for less than one percent of their public education expenditure. The ratio of total annual costs of refugee inclusion to host country public education expenditure is averagely 2.8 percent.32 It averages 3.8 percent and 2.4 percent in LICs and MICs, respectively. However, these averages mask large variations between host countries as refugee populations are concentrated in a few countries. For 36 out of the 67 LICs and MICs for which data is available, the annual cost of educating refugees was less than one percent of public education expenditure. In some countries like Somalia, Sudan and Jordan, the estimated annual cost of educating refugees is more than 15 percent of public expenditure on education, signaling either the large number of school-age refugee populations that the countries host or limited public 32 Data on public education expenditure is computed based on public education expenditure as a percentage of GDP multiplied by GDP for the year 2020. Data was obtained from the Global Education Finance Dashboard and the World Bank Development Indicators datasets. Data on public education expenditure was unavailable for five out of the 72 host countries in this study. These include Iraq, Lebanon, Libya, Montenegro, and Nigeria. 14 expenditure on education, or both. Figures 3a – 3c show the distribution of annual cost of refugee education as a percentage of public expenditure on education by country and income-group. Figure 3a. Projected annual cost of refugee education as a percentage of public expenditure on education – Low-income countries 25.00 20.23 20.00 15.44 15.00 10.00 8.93 7.84 6.02 5.00 2.95 2.71 2.40 1.65 1.07 0.82 0.81 0.43 0.40 0.30 0.26 0.19 0.19 0.11 - Figure 3b. Projected annual cost of refugee education as a percentage of public expenditure on education – Lower middle-income countries 25.0 23.4 20.0 15.0 10.0 8.7 5.8 5.7 5.0 3.7 2.7 2.0 1.9 1.8 1.4 0.8 0.7 0.5 0.4 0.3 0.2 0.2 0.2 0.11 0.11 0.07 0.06 0.05 0.03 0.02 - 15 Figure 3c. Projected annual cost of refugee education as a percentage of public expenditure on education – Upper middle-income countries 10.0 8.8 8.6 9.0 8.0 8.0 6.8 6.7 7.0 6.0 5.0 4.0 2.8 3.0 2.3 2.1 1.9 2.0 1.6 1.0 0.8 0.8 1.0 0.5 0.5 0.5 0.3 0.3 0.2 0.14 0.13 0.12 0.01 - The total financing gap will be determined by the amount donors and host country governments contribute to the total annual costs of inclusive refugee education. In high-income countries (HICs), the financing gap would be zero as host countries would contribute the total costs of inclusive refugee education. In LICs, it is expected that donors fully cover these costs. However, how these costs should be shared between donors and middle-income host countries is currently unclear. Table 3 below presents three scenarios of cost-sharing between donors and host countries and how this impacts the total financing required from the donor community. In this analysis, benchmarks for responsibility-sharing are determined using the metric of annual cost of refugee inclusion as a percentage of public education expenditure. This provides an indication of the scale of need as well as host countries’ ability to pay. This analysis is only indicative and other metrics could also be used to determine benchmarks. In all three scenarios, donors cover all inclusive refugee education costs in LICs. In scenario 1, donors cover any expenses that exceed 1 percent of public education expenditure in LMICs and 2 percent of public education expenditure in UMICs. This reduces the annual amount required from the donor community from US$9.3 billion to US$5.7 billion. In scenario 2, donors cover any expenses that are over 2 percent of public education expenditure in LMICs and 4 percent of public education expenditure in UMICs. In this case, total financing required from the donor community would be US$4.2 billion and host country contributions would amount to around half of the total costs of inclusive refugee education. In scenario 3, donors cover any expenses that are over 4 percent of public education expenditure in LMICs and 6 percent of public education expenditure in UMICs. In this case, the annual cost to be borne by the donor community falls to US$2.9 billion, but this would require that host countries contribute nearly 70 percent of the total annual costs. In 2022, donor financing directed to refugee education in LICs and MICs was around US$2.6 billion.33 This compares with scenario 3 described above. However, as host government and philanthropic contributions amounted to only US$2.4 billion in 2022, there is a large financing gap of around US$4.3 billion. Closing this gap will require additional commitments from donors and middle-income host countries. 33 World Bank. Forthcoming. Financing for Refugee Education – Key facts and findings: Insights into the costs and financing of refugee education in low- and middle-income countries. 16 Table 3. Donor financing requirement with increasing contributions from host countries Annual Cost (US$ millions) Base Scenario 1 Scenario 2 Scenario 3 Host country contribution as % of LIC: 0; LMIC: 1%; LIC: 0; LMIC: 2%; LIC: 0; LMIC: 4%; LIC: 0, MIC: 0 public expenditure of education UMIC: 2% UMIC: 4% UMIC: 6% Low-income countries 309 309 309 309 Lower middle-income countries 2,317 1,873 1,607 1,329 Upper middle-income countries 6,705 3,484 2,311 1,313 Total 9,330 5,666 4,228 2,951 Host country contribution as % of annual costs of inclusive refugee 0% 39% 55% 68% education The annual cost of inclusive refugee education above refers to the cost of reaching all school-age refugee children; however, this target may not be realistic given that nearly half of all school-age refugee children are out of school and that a large majority of them live in LICs and MICs where public education systems have limited capacity. Relaxing this assumption, Table 4 below provides the annual cost estimates for different targets. The SDG indicators 4.1 and 4.2 target universal access and completion at pre-primary, primary and secondary education levels by 2030. Achieving these targets for refugee children would cost US$9.3 billion annually. Given that the average primary education completion rate in LICs was only 67 percent in 2022,34 the goal of universal secondary completion became clearly unachievable. As such, countries established national targets or benchmarks based on their education priorities, institutional capacity, and availability of resources.35 Providing refugee children access to education in line with the national SDG 4 benchmarks would cost US$8.8 billion annually. Acknowledging that the pre-primary education sub-sector has been historically underfunded and that secondary education is expensive, we also estimate the cost of achieving SDG 4 national benchmarks at pre-primary and secondary levels but universal access at primary education for refugee children. This would cost US$8.9 billion annually. Finally, we estimate the cost of providing refugee children with education at parity with host country access rates in 2022, noting that these figures are likely to increase for host countries and therefore are inadequate targets for refugee children. The annual cost of educating refugee children in line with host country GERs is US$8.5 billion. Table 4. Annual cost of inclusive refugee education across different benchmarks Annual Cost (US$ Millions) Target LICs LMICs UMICs Total SGD 4 Targets - Universal pre-primary, primary and 309 2,317 6,705 9,330 secondary education SDG 4 National Benchmarks 260 2,124 6,454 8,838 Universal Access at Primary, SGD 4 National Benchmarks 267 2,153 6,482 8,902 at Pre-Primary and Secondary Access at Host Country GERs in 2022* 162 1,881 6,420 8,463 * GER capped at 100 percent 34 World Bank Development Indicators. https://data.worldbank.org/indicator/SE.PRM.CMPT.ZS 35 UNESCO. 2023. Can countries afford their national SDG4 benchmarks? Policy Paper 49. https://unesdoc.unesco.org/ark:/48223/pf0000385004 17 While these figures provide an indicative cost of integrating refugee children in national education systems, it is important to note that these costs are not static and do not encompass the costs of improving the quality of education. Public education expenditure fell following the COVID-19 pandemic,36 and increases in public expenditure may be required to achieve pre-pandemic levels of spending as well as to address learning loss through remedial support, implement re-enrollment strategies, provide second-chance education solutions to vulnerable children, and invest in new or refurbished infrastructure in line with public health protocols.37 Further, as host countries make greater investments in improving foundational learning, reducing pupil teacher ratios, reducing class sizes, equipping classrooms for digital learning and other quality improvements, public education expenditure is likely to increase. The Education Commission called for an increase in public education expenditure in LICs and MICs from 4 percent to 5.8 percent of GDP by 2030, requiring an annual rate of growth in public spending of 7 percent,38 for all children to be on track to achieve basic secondary skills. Such increases in public education spending will increase per capita public expenditure thereby increasing the base cost of refugee inclusion in national education systems. 36 UNESCO and World Bank. 2023. Education Finance Watch 2023. 37 UNESCO. 2020. Act Now: Reduce the impact of COVID-19 on the cost of achieving SDG 4. Policy Paper 42. 38 Education Commission. 2016. The Learning Generation: Investing in education for a changing world. https://report.educationcommission.org/wp-content/uploads/2016/09/Learning_Generation_Full_Report.pdf 18 Investing in Inclusive Refugee Education While educating refugees is a developmental challenge, it has historically been financed through humanitarian response. This mismatch has resulted in short-term, inadequate, unpredictable, and inflexible financing for refugee education, posing a serious challenge to host countries’ efforts to integrate refugee children in national education systems. While almost all host countries grant refugees the right to education alongside host community children, in LICs and MICs where public education expenditure is already stretched, only a few are able to operationalize this commitment. Parallel education systems, delivered through humanitarian agencies, continue to operate long after the initial stages of refugee crises. Without the assurance of adequate, multi-year, predictable financing, host countries will remain unable, or able but reluctant, to absorb these parallel system schools. Yet, only a few host countries access developmental financing for refugee education.39 There is an urgent need for financing to follow the shifts in policies towards inclusive refugee education, both in terms of increased volume of developmental financing available for refugee education and in the number of host countries accessing this financing. Current donor allocations to refugee education respond to each refugee crisis as its own event rather than within a structural framework. This approach has led to financing being inequitably allocated across refugee crises and host countries, driven by political considerations, bilateral relationships between donors and host countries, or competing and new crises with greater media attention.40 A shift towards a more data- driven, systematic approach to assessing host country needs and resource limitations can help address these inequities in donor allocations. A critical next step is for host countries and the donor community to engage on the degree of responsibility-sharing and associated financing instruments. The cost of educating refugees hosted in HICs do not feature in this exercise as it is expected that the host country covers the total fiscal cost of refugee inclusion. On the other hand, in LICs where public education expenditure is limited, it is reasonable to expect that the donor community fully finances the total fiscal costs of refugee inclusion in national education systems. However, many school-aged refugee children are hosted in LMICs and UMICs, where it is unclear what share of the fiscal costs should be covered by the donor community and what share by the host government. UMICs account for 72 percent of the total annual costs of inclusive refugee education. Many UMICs already include refugee children in national systems, with the largest share of the costs financed by the host governments with minimal external assistance (except in Turkiye).41 Yet many of these countries’ resources are limited, resulting in relatively low enrollment rates for refugee children. Commitments of increased financing should be accompanied by innovations in financing instruments to respond to medium-term recurrent needs. Inclusion in national systems will incur investment costs in host countries but in most cases these costs will be marginal. The bulk of the costs are recurrent. These costs cannot be financed through investment lending instruments. Instead, recurrent costs should be financed through budget support. However, there are inadequate financing instruments to provide predictable, multi-year financing for recurrent costs. Humanitarian agencies are funded through annual appeals – while humanitarian agencies finance recurrent expenditures, their financing appeals are rarely fully met, and financing is unpredictable, does not flow through government systems and has high 39 For instance, as of mid-2023, only eleven countries were accessing the World Bank Window for Host Communities and Refugees (WHR) and only two countries were accessing the Global Concessional Financing Facility (GCFF) for education projects in response to refugee crises. 40 World Bank and RAND Corporation. 2023. Financing Inclusive Refugee Education – Towards more equitable and transparent responsibility-sharing. 41 Abu-Ghaida et al. 2023. What could be more important than education? https://blogs.worldbank.org/dev4peace/what-could-be-more-important-education 19 transaction costs. As such, this funding channel is not suitable to finance inclusive systems. While the World Bank has several financing instruments to provide budget support to host governments (for instance, Development Policy Financing and Program for Results), many other partners, especially those that are at the nexus of humanitarian and development programming, do not. While LICs have access to interest-free loans or grants from IDA, most UMICs fall in the IBRD or blend lending category and have relatively limited access to concessional financing instruments for refugee education.42 Even where inclusive refugee education is financed through grants, there is an opportunity cost to finance as there is often a limited amount of donor financing available to each host country. Providing financing for inclusive refugee education as additions to existing donor allocations would incentivize host countries to access these funds. While external financing is limited, targeting investments in LICs and integrating refugees in traditional developmental financing where refugee populations are small can lead to immediate gains. LICs collectively host almost 2.4 million school-aged refugee children and account for only US$309 million or three percent of the annual cost of inclusive refugee education. Further, 23 out of the 72 LIC and MIC host countries have less than 10,000 school-aged refugee children. In these countries, existing developmental financing for education can be extended to include refugee children at relatively small marginal costs. Financing refugee education through developmental financing can lead to efficiency gains. Unlike humanitarian financing, developmental financing can embed a strong results-based orientation. Results- based financing has been successfully used to incentivize the integration of excluded populations into schools,43 and has been piloted in refugee-hosting countries like Lebanon and Ethiopia where Program for Results financing instruments have been used to implement incremental steps toward refugee inclusion in national systems. Further, directing financing through host governments can reduce overhead and transactional costs associated with hiring implementing partners in hard-to-reach areas. Recognizing that host countries are at different stages of implementing inclusive education policies, parallel systems may still be necessary but should be redesigned to be inclusion-enabling. Humanitarian response plays an important role in mobilizing resources quickly in the immediate onset of a refugee crisis or where there are political challenges to refugee inclusion, as well as during transition phases to ensure continued service delivery. In these cases, it is important that any support delivered is inclusion- enabling to the extent possible, i.e., follows the curriculum of the host government, provides language bridging or accelerated education programs to ease transition into national systems, follows the host country scale for teacher salaries etc. Further, humanitarian agencies should work with developmental partners and host governments from the very beginning to develop a plan to transition parallel support to national systems. Host countries will require financial, technical, and policy dialogue support to integrate refugee children in national education systems. Increased investments in refugee education should be complemented by costed inclusive refugee education plans, with clear multi-year financing targets and strengthened data systems to track refugee inclusion. It is important that host countries are supported to provide education for refugees even after they are integrated in national systems and become hard to distinguish from local children in education management information systems. This includes the recurrent and operational costs of education service delivery. Efforts to strengthen planning and implementation capacity where it is weak can have positive spillovers for the system as a whole. Understanding the status 42 World Bank and RAND Corporation. 2023. Financing Inclusive Refugee Education – Towards more equitable and transparent responsibility-sharing. 43 Adelman, M., Holland, P., & Heidelk, T. (2017). Increasing access by Waiving Tuition: Evidence from Haiti. Comparative Education Review, 61(4), 804-831. Retrieved from: https://www.journals.uchicago.edu/doi/ abs/10.1086/693904 20 quo of existing non-State education provision for refugees in terms of governance, financial flows, and links to other policy areas (such as freedom of movement) will be critical in the robust design of transition into national systems. Further, transitioning from parallel systems to national systems does not imply that humanitarian financing for refugee education will be redirected to host governments. Humanitarian response is often led by different line ministries than those responsible for developmental aid (among donors) or education (among host countries). As such, transitioning to an inclusive system will require careful planning and policy dialogue between donors, humanitarian and developmental partners and host governments. 21