Annual global spending on education is approximately US$4.7 trillion, of which 65% is spent in high income countries

Chapter 19 PDF

CREDIT: Muse Mohammed/IOM. These Syrian refugee children used International Organization for Migration (IOM) transportation to access this school, which is far away from their camp.

  • Globally, governments account for 79% of total spending, households for 20% and donors for 0.3% (12% in low income countries).
  • Of all money spent on education, just 0.5% is spent in low income countries.
  • In 2017, median public education spending was 4.4% of GDP and 14.1% of total public spending.
  • The fiscal consequences of immigration, including the cost of immigrant education, were modest at ±1% of GDP.
  • Few countries explicitly recognize migrant status in formula-based allocation of funds to schools. More address their needs indirectly by considering student language or socio-economic status.
  • In 2016, aid to education reached its highest level since records began at US$13.4 billion.
  • Multilateral development banks, such as the World Bank, have been reducing the share of education in their loans to middle income countries. A proposal for an International Financing Facility for Education aims to address this issue, but loans would need to be equity-oriented.
  • Policy-makers should not hold high expectations on the role aid can play in controlling migration.
  • Humanitarian and development aid provided about US$800 million for refugee education in 2016, but without joint planning.
  • The Education Cannot Wait fund set up in 2016 is indicative of recent efforts to bridge humanitarian and development aid.
  • Remittances increased education spending by up to 35% in 18 countries in sub-Saharan Africa, Central and Southern Asia and South-eastern Asia. Lowering the cost of remitting from 7.1% to the SDG 10.c target of 3% would provide households with an extra US$1 billion a year to spend on education.

The three main sources of education financing are governments, donors and households. Analysis for this report estimates annual spending on education at US$4.7 trillion worldwide. Of that, US$3 trillion (65% of the total) is spent in high income countries and US$22 billion (0.5% of the total) in low income countries (Figure 18a), even though the two groups have a roughly equal number of school-age children. Governments account for 79% of total spending and households for 21%. Donors account for 12% of total education expenditure in low income countries and 2% in lower middle income countries (Figure 18b).

Figure 18: Governments account for four out of five dollars spent on education

Public expenditure

Globally, median public education expenditure was 4.4% of GDP in 2017, above the minimum benchmark of 4% specified in the Education 2030 Framework for Action; by region, it ranged from 3.4% in Eastern and South-eastern Asia to 5.1% in Latin America and the Caribbean. The median share of total public expenditure dedicated to education was 14.1%, below the minimum benchmark of 15%, with regional rates ranging from 11.6% in Europe and Northern America to 18% in Latin America and the Caribbean. In total, 43 out of 148 countries do not meet either benchmark.

The median share of public education expenditure dedicated to primary education is 35%, in a range from 47% in low income to 26% in high income countries. A global median of 35% of total education spending was allocated for secondary education; the range was from 27% in low income to 37% in high income countries. Countries in Europe and Northern America spend the same amount per primary and post-secondary student. Sub-Saharan Africa spends 10 times more per post-secondary than per primary student.

Public debate often focuses on immigration’s negative effects on host community welfare. On the one hand, immigrants are more likely than natives to be of working age but typically generate less tax revenue because they earn less. On the other hand, immigrants are more likely to depend on social benefits and use public services, such as education. However, public education expenditure on immigrant children is best understood as an investment: They typically contribute more in tax and social security contributions than they receive over their lifetime. The fiscal consequences of immigration are relatively modest – typically within 1% of GDP, whether positive or negative.

Aid expenditure

In 2016, aid to education reached its highest level since records began in 2002. Compared with 2015, aid to education grew by US$1.5 billion, or 13% in real terms, to reach US$13.4 billion. Aid to basic education accounted for two-thirds of the increase; aid to secondary and post-secondary education rose at a lower rate, so the share of basic education in total education aid reached its highest level at 45%.

Aid disbursements to basic education are still not allocated to the countries most in need. The share of basic education aid to low income countries fell from 36% in 2002 to 22% in 2016. The share to least-developed countries was 34% in 2016, down from a 2004 peak of 47%.

The share of basic education aid to low income countries fell from 36% in 2002 to 22% in 2016

Lower middle income countries also face a financing gap. Multilateral development bank lending for education in these countries remains low. For instance, the median share of education between 2002 and 2017 was 10.5% for concessional loans from the International Development Association (IDA) but only 6.4% for non-concessional loans from the International Bank for Reconstruction and Development (IBRD). Moreover, the share of education in IBRD loans fell from 8.2% in 2012 to 4.7% in 2017, or one-quarter the amount for the energy and extractives sector (Figure 19). Deliberations continue on a proposal by the International Commission on Financing Global Education Opportunity to establish an International Finance Facility for Education to increase development bank lending for education in lower middle income countries.

Figure 19: A low and declining share of World Bank non-concessional loans goes to education


The idea that external assistance can reduce migration receives some support in policy circles. Raising disposable income in origin countries would reduce a key emigration incentive. The European Agenda on Migration and some countries’ aid policies support the idea.

A study of migration flows from 210 origin countries to 22 donor (and destination) countries showed that countries sending large numbers of migrants received the largest amount of foreign aid. However, the causality of the relationship is difficult to establish. If aid reaches poorer families, it may help them finance their migration costs. Aid may also increase information on donor countries and lower migrants’ transaction costs, encouraging emigration. Policy-makers may therefore need to temper expectations on the role aid can play in controlling migration.

The impact of education aid, in particular, on migration is very difficult to distinguish due to its small size. Even if aid for education contributes to reducing emigration, it is unlikely to do so quickly. However, education can have an important mediating role in migration for both origin and destination countries. A study of trends in migration from Northern Africa to OECD countries suggested that aid-induced income growth in origin countries was a push factor for migrants with low education. But satisfaction with local public services, including schools, may deter migration. And overall, demographic characteristics of the destination country – population density, urban population growth, age dependency ratios – and rewards for education in destination countries were much more important in determining migration rates than aid levels.

The causality of the relationship between international aid and migration flows is difficult to establish


Humanitarian aid increased in 2017 for the fourth year in a row, with global humanitarian funding to education reaching US$450 million. However, education’s share in total humanitarian aid was 2.1%, far below the target of at least 4%. Education has also historically had one of the lowest shares of funded requirements, compared with other sectors. However, the 2016 World Humanitarian Summit included a new commitment to education in emergencies through the establishment of the multilateral Education Cannot Wait (ECW) fund.

Multiple layers of coordination mechanisms, both vertical (from global to local) and horizontal (across sectors and actors), make the humanitarian aid architecture complex and contribute to the difficulty of increasing education’s part in emergency response. UN-coordinated interventions are organized in two ways. Humanitarian response plans (HRPs) provide a country strategy, usually over one year. Flash appeals propose responses to sudden emergencies to address acute needs over three to six months. HRPs have had limited education-related content, and education has often not featured in appeals.

There is no obligatory reporting system for international humanitarian assistance expenditure, making full tracking of education spending difficult. The OECD Development Assistance Committee Creditor Reporting System database does not break down humanitarian aid by sector, although it suggests that about US$400 million of development aid was spent on refugee education in 2016. The voluntary United Nations Office for the Coordination of Humanitarian Affairs Financial Tracking Service tracks education, but 42% of funding in 2017 was in the multisector category, shared among sectors or unspecified, resulting in potentially inaccurate education spending estimates.

Education underfunding results from insufficiencies in several areas: political will, vertical and horizontal coordination, humanitarian actor capacity and information or accountability mechanisms. A 2015 costing study calculated the funding gap at US$8.5 billion, which would need to be covered by the international community. This amounts to US$113 per child — a tenfold increase in current per-student spending. ECW aims to raise US$1.5 billion per year and increase education’s share of humanitarian aid by 1.2 percentage points by 2021. Its partners will need to share fundraising efforts, either prioritizing humanitarian objectives within development aid or increasing the priority of education within humanitarian aid.

Plans should not be determined by the activities organizations are used to implementing; rather, there is a need for actions that serve inclusive and equitable education of good quality. The Global Education Cluster has established guidelines for needs assessment. As a major new actor, ECW could enforce these guidelines by requiring partners to use them in programme design and developing their capacity.

Humanitarian aid has struggled with coordination among and within main actors. Education should be key in multisector humanitarian intervention plans. In the Rohingya crisis in Bangladesh, education was included from the first stage of relief response, and education service delivery began very quickly. However, lack of coordination meant learning centres were sometimes empty, as non-education interventions competed for children’s time.

Joint planning is also needed between humanitarian and development actors. While multisector planning helped ensure education’s place in the Rohingya humanitarian response, it did not ensure inclusion in the Bangladeshi education system. Multiyear appeals in protracted crises have been underfunded, and national education plans have not historically focused on resilience or crisis response. Short-term humanitarian education planning frameworks need to be combined with more structural concerns, as in Uganda, where the government’s Refugee and Host Community Education Response Plan, announced in May 2018, allocates US$395 million over three years to reach about 675,000 refugee and host community students per year.

Donors such as Australia, Canada, Norway and the United Kingdom are shifting to multiyear funding frameworks to ensure predictable funding to humanitarian partners, and ECW has a ‘multi-year resilience’ instrument, which also aims to bridge the humanitarian–development divide. Accountability in humanitarian aid needs to be strengthened. ECW has emphasized transparency and results by setting high standards in its strategic results framework, which covers not only its ultimate results but, crucially, its way of working and its operational efficiency.

Household expenditure

Lack of data on household contributions has long restricted analysis of total education expenditure. The UIS first released data for households in 2017, but coverage remains limited. The share of households in total education expenditure can be very high in some low and middle income countries, e.g. El Salvador (50%) and Indonesia (49%) (Figure 20).

Figure 20: Household education spending is considerable in some low and middle income countries

In general, household expenditure data do not appear to change dramatically from year to year, with some exceptions. In Chile, massive student protests about insufficient government education financing caused changes in public policy. As a result, the household share of total education spending fell between 2005 and 2015 from almost 50% to below 20%.

Previous years