Expenditure per student increases with the educational level in nearly all OECD countries, although by how much varies substantially among countries. On average, expenditure per student amounts to about USD 11 900 at primary level, USD 13 300 at secondary level and USD 20 500 at tertiary level.
Increasing expenditure per student typically correlates with improved outcomes, but only up to a point after which additional investment shows little impact on performance. For instance, Japanese 15-year-olds’ mathematics scores are higher than those of their US peers, even though Japan spends about 30% less per student aged 6 to 15 than the United States.
National averages sometimes hide substantial variation in spending within countries. In Canada, the highest-spending region records double the expenditure per student at primary and secondary level of the lowest-spending region, while in the United States it is three times the amount. Belgium and Lithuania have few regions and small differences by region in expenditure per student at primary and secondary education.
Education at a Glance 2024
Education finance plays an important role in ensuring equitable access to high-quality education and analysing the expenditure dynamics is crucial to understand the challenges of achieving equity in education while maintaining quality. Meeting the policy goals of expanding access to educational opportunities and providing high-quality education can mean higher costs which must be balanced against other demands on public expenditure and the overall tax burden. As a result, understanding the impact of different spending approaches is crucial in determining where investments can achieve the best outcomes. Although it is difficult to assess the optimal level of resources needed to prepare students for life and work in modern societies, international comparisons of spending on educational institutions per student can provide useful reference points.
This chapter first looks at expenditure per student at different levels of education. It then looks at spending per student on core services (e.g. teachers' salaries – see Chapter D3), ancillary services (e.g. meals, residence halls and healthcare) and, in the case of tertiary education, research and development. This is followed by an analysis of recent trends in spending per student and data on subnational variations in spending. Finally, this chapter also explores the relationship between expenditure and learning outcomes by combining data from the 2022 Programme for International Student Assessment (PISA) and cumulative expenditure on students from the ages of 6 to 15.
The average annual spending per student from primary to tertiary education in OECD countries is around USD 14 200, ranging from around USD 3 500 in Mexico to over USD 30 100 in Luxembourg.
Appropriately dividing government spending between core services (such as teaching costs), ancillary services (such as canteen services), and research and development (R&D) can be challenging. On average across OECD countries, expenditure on core education services represents 86% of total expenditure per student from primary to tertiary educational institutions.
In Denmark and Sweden, more than half of the expenditure at tertiary level is dedicated to R&D, well above the OECD average. Across OECD countries, 65% of total expenditure on educational institutions at tertiary level goes to core services (about USD 14 100 per student), while over 30% of total expenditure goes to R&D (about USD 7 000 per student).
More is spent per child on early childhood education for 0-2-year-olds than for pre-primary programmes, which cater to children over 2. This gap, partly driven by differing child-adult ratios, is particularly pronounced in Denmark, Finland, and Romania, where spending per child is about two to three times higher for programmes aimed at children aged 0 to 2 than for pre-primary programmes.
In 2021, the annual spending per student from primary to tertiary education in OECD countries was around USD 14 000 on average. But this average masks a wide range, from around USD 3 500 per student in Mexico to over USD 30 000 in Luxembourg (Table C1.1)The drivers behind these spending levels vary across countries and by level of education: in Luxembourg, for example, low ratios of students to teaching staff and high teachers’ salaries at primary and secondary levels (see Chapter D3) are reflected in high levels of expenditure per student. In contrast, Mexico has one of the highest ratios of students to teaching staff, which tends to drive costs down (see Indicator D7 in (OECD, 2023[1])). Differences can also be attributed to differences in gross domestic product (GDP) per capita, which are reflected in different teacher salaries and other costs, with GDP per capita in Mexico and Luxembourg at the opposite ends of the scale for OECD countries (see Annex 2).
The distribution of expenditure at different levels of education reflects the relative costs of education provision. Expenditure per student on educational institutions rises with the level of education in almost all countries, but by how much varies markedly. On average across OECD countries, expenditure per full-time equivalent student is around USD 11 900 at primary level, USD 13 300 at secondary level and USD 20 500 at tertiary level (Figure C1.1.). In pre-primary education, expenditure per child, rather than per full-time equivalent student, is around USD 11 700 (Box C1.1.).
Primary and secondary education take place in settings with generally similar organisations (with the exception of vocational programmes) and similar patterns of expenditure per student. In contrast, at tertiary level, higher salaries, the specialised resources needed such as advanced laboratory equipment and the investment in R&D contribute to the comparatively higher expenditure per student.
Among OECD countries, Italy and Korea stand out as the only countries where expenditure per student is higher at primary level than tertiary level (Table C1.1). In Italy, expenditure per student has increased faster at primary than tertiary level in recent years, to the point where expenditure per student at primary exceeded that at tertiary level. This is partly due to the integration of students with special educational needs, supported by the implementation of a law in 2020 providing a noticeable increase of teachers for special need students at primary level.
In contrast, in Hungary, Luxembourg, Romania, Türkiye, and the United Kingdom, expenditure per tertiary student is at least 2.4 times higher than it is for each primary student. In Luxembourg, where expenditure per student is the highest, the difference between the two levels is about USD 34 700, but it falls to about USD 9 900 when excluding R&D. In Denmark, Finland, Norway and Sweden, expenditure per student at tertiary level is lower than at primary level when R&D is excluded. While this is typical of Nordic countries, this is also true in a few other OECD countries where a substantial share of tertiary expenditure is dedicated to R&D (Table C1.1).
There is extensive research evidence showing that high-quality early childhood education and care brings both short- and long-term benefits, in terms of learning, well-being and subsequent educational, social and employment outcomes (Currie and Almond, 2011[2]; Taggart et al., 2015[3]; Van Huizen and Plantenga, 2018[4]; Heckman and Karapakula, 2019[5]). Yet, children from socio-economically disadvantaged families are less likely to participate in early childhood education and care (ECEC) and receive high quality care (see Chapter B1). Investing in ECEC is particularly important to address sources of inequality and promote intergenerational mobility. Complementing data on participation in early childhood education in Chapter B1, Figure C1.2 shows how much is spent per child in such programmes across OECD countries.
Figure C1.2. shows that, among countries with data available, expenditure ranges from less than half of the OECD average of about USD 11 700 per child in pre-primary education to more than twice that figure. In Denmark, Finland and Romania, there is a substantial difference between early childhood educational development and pre-primary programmes, while in Australia, Chile, and Lithuania spending per child is similar. Countries with above-average pre-primary spending per child also tend to be the ones with a wider gap between the two levels.
The financing of ECEC education programmes reflects – and may influence – quality of staffing, minimum space requirements, pedagogy and other process quality measures (Edwards, 2021[6]). Taking staffing as an example, higher minimum qualifications for teachers often raise staff compensation. Child-adult ratios can often be associated with the quality of care, but they also are determined by needs at different age levels (see Table D2.1.); this is reflected in the higher spending per child in early childhood education development compared to pre-primary for many countries.
The data in Figure C1.2 refer to expenditure per child, rather than per full-time equivalent student (which is the standard measure used in this chapter for primary to tertiary education). This means that a child who spends eight hours a day in ECEC and one that spends four hours a day will both count as one child (if they were in primary education, the first would count as one full-time equivalent, the second as 0.5). It is therefore important to acknowledge the wide differences in ECEC hours across countries. For 0-2 year-olds, average hours can range from 20 to 40 hours per week (OECD, 2023[7]). For 4-6 year-olds, data from 13 OECD countries show that minimum compulsory ECEC hours range from 500 to 1 300 hours per year (OECD, 2022[8]).
From an equity perspective, three factors can influence the intensity of participation and expenditure on ECEC, involving both households and governments. First, policies that support parents in the labour market (e.g. parental leave, job protection laws) may help them return to work after childbirth and encourage participation in ECEC, as happened in the United States after the passage of maternal leave packages (Hofferth and Curtin, 2003[9]). Second, high childcare costs can discourage participation in ECEC, especially among low-earning parents. This was found to be the case in Ireland and Germany (ESRI, 2018[10]; Hermesa et al., 2022[11]); research in multiple European countries has shown that subsidies would help families to rely on ECEC and return to work (Narazani et al., 2022[12]). Lastly, a country’s culture may affect the role of informal provision (e.g. family members looking after children, even when parents could afford formal childcare) and affect spending on ECEC.
Spending on education can be split between core services, ancillary services and R&D, each playing a different role in ensuring desirable learning outcomes and addressing potential inequalities. Resources allocated to ancillary services such as meals, transportation and accommodation play an important role in allowing students to pursue their education in the right conditions. Ensuring equitable access to these services is essential for addressing disparities in educational outcomes among different student populations. Meanwhile, spending on R&D plays a vital role in fostering innovation and driving economic growth.
Core services encompasses essential components directly related to instruction and account for most of the expenditure. On average across OECD countries, expenditure on core education services represents 86% of total expenditure per student from primary to tertiary educational institutions and it exceeds 90% in Chile, Israel, Latvia and Romania. At primary and secondary levels, among OECD countries with data available, on average about 94% of the expenditure on educational institutions is devoted to core educational services (equivalent to about USD 11 900 per student at primary level and USD 13 400 at secondary level). However, in Finland, France, Hungary, the Slovak Republic and the United Kingdom, ancillary services account for 10% or slightly more of the expenditure on educational institutions (Table C1.2).
At tertiary levels R&D expenditure can be substantial, reducing the share dedicated to core services (Figure C1.3). On average across OECD countries, 65% of total expenditure on tertiary educational institutions goes to core services (about USD 14 100 per student), while over 30% is on R&D (about USD 7 000 per student). More than half of the expenditure at tertiary education is dedicated to R&D in Denmark (equivalent to about USD 14 500 per student) and Sweden (USD 14 100). This might be explained by the essential R&D role universities play in these countries, while in others more R&D may be conducted outside educational institutions.
Expenditure per student in tertiary education is below average in Bulgaria, Chile and Romania. These countries also stand out by their very low spending on R&D in tertiary education. Spending on R&D represents only 4% of expenditure on tertiary institutions in Chile (USD 421 per student) and Bulgaria (USD 531 per student), and less than 1% in Romania, where it amounts to just USD 29 per student (Table C1.2). The historical development and structure of higher education systems in these countries may not have prioritised R&D activities within tertiary education institutions. For example, Bulgaria and Romania both inherited the Soviet model of the organisation of the science system, where the main institutional bodies were industrial research institutes, leaving public higher education institutions with very weak R&D (Radosevic and Auriol, 1999[13]; European Commission, 2021[14]).
Changes in spending on educational institutions primarily mirror fluctuations in the size of the school-age population (especially at primary and lower secondary levels, where enrolment is near universal) and the allocation of funds towards teachers' salaries, which constitute a key component of education expenditure.
On average across OECD countries, the total expenditure on primary to tertiary educational institutions per full-time equivalent student increased by 1.8% between 2015 and 2021. This was the result of a slight increase in expenditure (2.1%) and nearly stable numbers of students (a 0.2% increase in the number of full-time-equivalent students). However, the average hides cross-country variation in both student numbers and expenditure. While most countries experienced moderate growth in both students and overall expenditure, some have seen a decline in student enrolment accompanied by moderate to high growth in expenditure. For example, in Bulgaria, Hungary, Poland and Romania this resulted in high growth in expenditure per student, ranging from 5% to 9% (Figure C1.4).
Provisional data on education expenditure in 2022 are available for a small number of countries. After accounting for inflation, expenditure on primary to tertiary educational institution per student fell between 2021 and 2022 in Germany, Lithuania, New Zealand, Slovenia and Spain. The difference was the largest in Lithuania and Spain where both countries observed a decrease of 3% (Table C1.4, available online).
In some countries, subnational governments play an important, sometimes even a leading, role in financing education (see Chapter C4). Geographical disparities in economic activity mean that regions and municipalities may have varying levels of capacity to raise resources for education. This may be balanced by funding from central governments (e.g. a funding formula might allocate more resources to poorer areas). But there are many other potential factors at play; for example rural areas with smaller classes will have higher spending per student, all other things being equal. The combination of such factors can lead to substantial within-country variation in expenditure per student. Examining regional expenditure per student can help identify potential sources of inequality within countries. It also enables cross-country comparisons – a region that ranks highly within its own country may still lag behind internationally.
Figure C1.6Figure C1.5 shows that regional disparities in expenditure vary greatly across countries, especially among countries with highly decentralised education funding systems and many regions. For example, very little funding originates from central government for primary education in Germany (3%), Canada (5%), Spain (9%) and the United States (12%) (see Figure C4.2.). But regional variation is greater in Canada and the United States than in Germany or Spain. In Canada, the region with the highest expenditure per student is double that of the lowest region while in the United States, it is three times higher. In Germany the federal state with highest per-student expenditure spends 1.5 as much as the one with the lowest spending. Belgium and Lithuania, each with only two regions with data, have minimal differences in expenditure per student.
The data also show that two capital cities, Madrid and Berlin, are at opposite ends of the spectrum within their own national expenditure contexts. The absence of a clear expenditure pattern regarding capital cities underscores the decentralised nature of education governance and funding structures within countries (see Chapter C4). Although capital cities often serve as administrative and political centres, their expenditure per student is shaped by a wide range of factors.
Figure C1.6 shows the association between average student performance in PISA and cumulative expenditure per student between the ages of 6 and 15. Increasing expenditure per student typically correlates with improved outcomes, up to a certain threshold. Specifically, there is a positive association between cumulative expenditure and average performance until the threshold of about USD 100 000 per student from ages 6 to 15 is reached. Beyond this point, additional investment shows little to no discernible relationship with student performance.
For example, the cumulative expenditure per student between ages 6 to 15 is over USD 150 000 in the United States. Yet, the mathematics scores of 15-year-old students in Japan are higher, even though Japan spends about 30% less. This shows that increasing financial resources for schools alone may not address all educational challenges. Rather, the key lies in the strategic allocation of funds supporting quality education (Table C1.5, available online).
Ancillary services are services provided by educational institutions that are peripheral to their main educational mission. The main component of ancillary services is student welfare. In primary, secondary and post-secondary non-tertiary education, student welfare services include meals, school health services and transport to and from school. At the tertiary level, they include residence halls (dormitories), dining halls and health care.
Core educational services include all expenditure that is directly related to instruction in educational institutions, including teachers’ salaries, construction and maintenance of school buildings, teaching materials, books, and school administration.
Research and development includes research performed at universities and other tertiary educational institutions, regardless of whether the research is financed from general institutional funds or through separate grants or contracts from public or private sponsors.
The annual average growth rate is calculated using the compound annual growth rate which shows the geometric progression ratio that provides a constant rate of return over the time period under analysis.
Expenditure per child on educational institutions for early childhood education development and pre-primary levels is calculated by dividing total expenditure on educational institutions at that level by the corresponding sum of full-time and part-time enrolment, resulting in total expenditure on educational institutions per head count as opposed to per full-time equivalent student. Expenditure per student on educational institutions at a particular level of education (primary to tertiary) is calculated by dividing total expenditure on educational institutions at that level by the corresponding full-time equivalent enrolment. Only educational institutions and programmes for which both enrolment and expenditure data are available are taken into account. Expenditure in national currencies is converted into equivalent USD by dividing the national currency figure by the purchasing power parity (PPP) index for GDP. The PPP conversion factor is used because the market exchange rate is affected by many factors (interest rates, trade policies, expectations of economic growth, etc.) that have little to do with current relative domestic purchasing power in different OECD countries (see Annex 2 for further details).
Data on subnational regions on how much is spent per student are adjusted using national PPPs. Future work on the cost of living at subnational level would be required to fully adjust the expenditure per student used in this section.
Full-time equivalent student: The ranking of OECD countries by annual expenditure on educational services per student is affected by differences in how countries define full-time, part-time and full-time equivalent enrolment. Some OECD countries count every participant at the tertiary level as a full-time student, while others determine students’ intensity of participation by the credits that they obtain for the successful completion of specific course units during a specified reference period. OECD countries that can accurately account for part-time enrolment have higher apparent expenditure per full-time equivalent student on educational institutions than OECD countries that cannot differentiate between the different types of student attendance.
Vocational education and training expenditure: Expenditure on workplace training provided by private companies is only included when it is part of combined school- and work-based programmes, provided that the school-based component represents at least 10% of the study over the whole programme duration. Other types of employer-provided workplace training (e.g. entirely work-based training or employee training that takes place 95% at work) are excluded. Expenditure on VET programmes include the expenditure on training (e.g. salaries and other compensation of instructors and other personnel, as well as the cost of instructional materials and equipment). However, it excludes apprentices’ wages and other compensation to students or apprentices.
Please see the OECD Handbook for Internationally Comparative Education Statistics (OECD, 2018[15]) for more information and Education at a Glance 2024 Sources, Methodologies and Technical Notes, for country-specific notes.
Data refer to the financial year 2021 (unless otherwise specified) and are based on the UNESCO, OECD and Eurostat (UOE) data collection on education statistics administered by the OECD in 2023 (for details see Education at a Glance 2024 Sources, Methodologies and Technical Notes. Data from Argentina, China, India, Indonesia, Saudi Arabia and South Africa are from the UNESCO Institute of Statistics (UIS).
Provisional data on educational expenditure in 2022 are based on an ad-hoc data collection administered by the OECD and Eurostat in 2023.
Data from Table X2.2. are used to compute expenditure in constant 2015 prices and in equivalent USD converted using PPPs.
References
[2] Currie, J. and D. Almond (2011), “Human capital development before age five”, in Handbook of Labor Economics, Elsevier, https://doi.org/10.1016/s0169-7218(11)02413-0.
[6] Edwards, S. (2021), “Process quality, curriculum and pedagogy in early childhood education and care”, OECD Education Working Papers, No. 247, OECD Publishing, Paris, https://doi.org/10.1787/eba0711e-en.
[10] ESRI (2018), Maternal Employment and the Cost of Childcare in Ireland, Economic and Social Research Institute, Dublin, https://doi.org/10.26504/rs73.
[14] European Commission (2021), PSF Review of the Romanian R&I system: Background Report, Publications Office, https://doi.org/10.2777/38334.
[5] Heckman, J. and G. Karapakula (2019), “Intergenerational and intragenerational externalities of the Perry Preschool Project”, NBER Working Papers, No. 25889, National Bureau of Economic Research, https://doi.org/10.3386/w25889.
[11] Hermesa, H. et al. (2022), “Early child care and labor supply of lower-SES mothers: A randomized controlled trial”, DICE Discussion Paper, No. 394, Heinrich-Heine-University Düsseldorf, https://hdl.handle.net/10419/267152.
[9] Hofferth, S. and S. Curtin (2003), “The impact of parental leave statutes on maternal return to work after childbirth in the United States”, OECD Social, Employment and Migration Working Papers, No. 7, OECD Publishing, Paris, https://doi.org/10.1787/826588878522.
[12] Narazani, E. et al. (2022), “The impact of alternative childcare policies on mothers’ employment for selected EU countries”, JRC Working Papers on Taxation and Structural Reforms, No. 08/2022, Joint Research Centre, European Commission, https://joint-research-centre.ec.europa.eu/reports-and-technical-documentation/impact-alternative-childcare-policies-mothers-employment-selected-eu-countries_en.
[1] OECD (2023), Education at a Glance 2023: OECD Indicators, OECD Publishing, Paris, https://doi.org/10.1787/e13bef63-en.
[7] OECD (2023), “PF3.2: Enrolment in childcare and pre-school”, OECD Family Database, OECD, https://www.oecd.org/els/soc/PF3_2_Enrolment_childcare_preschool.pdf.
[8] OECD (2022), Ad-hoc survey on early childhood education, Unpublished.
[15] OECD (2018), OECD Handbook for Internationally Comparative Education Statistics 2018, OECD Publishing, Paris, https://doi.org/10.1787/9789264304444-en.
[13] Radosevic, S. and L. Auriol (1999), “Patterns of restructuring in research, development and innovation activities in central and eastern European countries: an analysis based on S&T indicators”, Research Policy, Vol. 28/4, pp. 351-376, https://doi.org/10.1016/s0048-7333(98)00124-3.
[3] Taggart, B. et al. (2015), “How pre-school influences children and young people’s attainment and developmental outcomes over time”, Research Brief, Department for Education, London, https://assets.publishing.service.gov.uk/media/5a803cb240f0b62305b89fbf/RB455_Effective_pre-school_primary_and_secondary_education_project.pdf.pdf.
[4] Van Huizen, T. and J. Plantenga (2018), “Do children benefit from universal early childhood education and care? A meta-analysis of evidence from natural experiments”, Economics of Education Review, Vol. 66, pp. 206-222, https://doi.org/10.1016/j.econedurev.2018.08.001.
Cut-off date for the data: 14 June 2024. Any updates on data and more breakdowns can be found on the OECD Data Explorer (http://data-explorer.oecd.org/s/4s).
In equivalent USD converted using PPPs for GDP, direct expenditure within educational institutions, by level of education
Note: See under Chapter C1 Tables for StatLink and Box C1.2. for the notes related to this Table.
Note: See under Chapter C1 Tables for StatLink and Box C1.2. for the notes related to this Table.
In equivalent USD converted using PPPs for GDP, direct expenditure within educational institutions, by level of education
Note: See under Chapter C1 Tables for StatLink and Box C1.2 for the notes related to this Table.
Note: See under Chapter C1 Tables for StatLink and Box C1.2 for the notes related to this Table.
Constant 2015 prices in equivalent USD converted using PPPs, by level of education
Note: See under Chapter C1 Tables for StatLink and Box C1.2 for the notes related to this Table.
Table C1.1 Total expenditure on educational institutions per student (2021)
Note: Expenditure per child/student is based on headcounts for early childhood education and on full-time equivalent students for all other levels of education.
1. Post-secondary non-tertiary figures are treated as negligible.
2. Upper secondary vocational programmes include lower secondary vocational programmes.
3. Data do not cover day care centres and integrated centres for early childhood education.
Table C1.2 Total expenditure on educational institutions per student on core educational services, ancillary services and R&D (2021)
Note: Based on full-time equivalent students. Some levels of education are included with others. Refer to "x" and "d" codes in Table C1.1 for details.
1. Primary education includes pre-primary programmes.
2. Upper secondary vocational programmes include lower secondary vocational programmes.
Table C1.3 Average annual change in total expenditure on educational institutions per student (2015 to 2021)
Note: Data on R&D are included in tertiary education, unless otherwise specified.
1. Primary education includes pre-primary and lower secondary programmes.
See Definitions and Methodology sections and Education at a Glance 2024 Sources Methodologies and Technical Notes (https://doi.org/10.1787/e7d20315-en) for more information.
Data and more breakdowns are available on the OECD Data Explorer (http://data-explorer.oecd.org/s/4s).
Please refer to the Reader's Guide for information concerning symbols for missing data and abbreviations.