3. Selected types of social investment
3.2. Impact of investment in school education on learning outcomes
Education underpins long-term gains for individuals and enhances society’s human capital.It does so by increasing people’s labour market opportunities and mitigating poverty and social exclusion risks. Individual benefits accrue to the economy and society as a whole, fostering the productivity, competitiveness and innovation necessary for successful implementation of the green and digital transitions. Simulations assuming a gradual improvement over 15 years show that increasing basic skills by 25 PISA (151) points could lead to a 0.5 pp higher average annual growth rate in EU GDP in the long term. (152) Recent evidence from Portugal suggests that well-designed reforms expanding the Vocational Education and Training (VET) offer in public schools increase upper secondary graduation rates. The reform introduced a VET track alongside an academic track (high school), which particularly benefitted those that otherwise tend to leave school earlier. The effects are larger for low-achieving students, children of less educated parents, and welfare recipients, including leading to higher wages and other positive outcomes for participants over several years.(153) Evidence shows that low-income countries with higher shares of graduates with tertiary education experienced stronger catching-up towards the average GDP per capita between 2008 and 2021.(154)
During the COVID-19 pandemic, the composition of total public expenditure shifted away from education.Spending on education was 9.5% of total public expenditure in 2022, 0.6 pp below pre-COVID-19 pandemic levels in 2019 (see Chapter 1, Section 3.6). However, the latest available data from 2022 show that expenditure increased slightly (by 0.1 pp) compared to 2021, potentially indicating some signs of recovery. In 2022, over 70% of public spending on education went to school education, which includes both (pre-)primary and secondary levels, each accounting for roughly similar shares (34% and 36.8%, respectively) (Chart 3.4). Tertiary education received 16.6% of public expenditure on education. At Member State level, the allocation of public expenditure to different education levels is driven by various factors, such as the duration of compulsory education, relative wages in the education sector, or class size. (155) Between 2012 and 2022, the share spent on school education increased slightly (+0.4 pp), driven by pre-primary and primary levels (+2.5 pp), but decreased for tertiary education (-0.5 pp) which might (partly) reflect the changes in demographic composition.
Chart 3.4
School education receives the largest share of public expenditure on education
Weighted average EU share of government expenditure on education, by education level, 2012-2022
Note: School education includes (pre-)primary and secondary levels. Further categories of public expenditure on education not included in the chart are post-secondary non-tertiary education, education not definable by level, subsidiary services to education, research and development (R&D) education, and education not elsewhere classified.
Source: Eurostat [gov_10a_exp].
PISA outcomes point to worsening basic skills in the EU over time, with potential negative impacts on EU competitiveness and wage developments in the longer term. The average PISA score across reading, mathematics and science subjects decreased by 10 points since the last wave in 2018, and by 18 points over the last 10 years in the EU (Chart 3.5). (157) While this declining pattern was also observed in other industrial countries (with the average PISA score dropping by 9 points since 2018 and 19 points since 2012 in OECD countries on average), the EU performed quite poorly compared to other advanced economies such as the US, the UK, Canada or Japan. In 2022, for example, it had the highest percentage of underachievers across all three subjects, and one of the lowest shares of top performers among those countries. Although PISA outcomes had already started to decline prior to the COVID-19 pandemic, learning losses during lockdowns might have intensified this downward trend among other possible factors. Box 3.3 presents a case study on Italy, showing the persistent impacts of COVID-19 pandemic school disruptions on student achievement, with learning losses still significant in 2021 and 2022.
Chart 3.5
Basic skills and efficiency of cumulative expenditure on education per student decreased post-COVID-19 pandemic
Average EU PISA score across reading, mathematics and science, and PISA EU efficiency score, 2012-2022
Note: PISA efficiency score measures the number of PISA points produced by USD 1 000 (expressed in purchasing power parity (PPP)) of cumulative expenditure on education per student, following the methodology in European Commission (2023d). It is calculated as 1 000 x (average PISA score over reading, mathematics and science)/cumulative expenditure per student from age 6-15 (in USD, PPP), and can be interpreted as an approximate measure of the pay-off from additional expenditure for a school education system. Yearly averages for Member States for which data are available.
Source: OECD PISA data for 2012, 2015, 2018, 2022.
Box 3.3: Impact of COVID-19 pandemic schooling disruptions on student achievement in Italy
Italy was one of the countries first and most severely hit by the COVID-19 pandemic. It experienced 13 weeks of complete school closure in spring 2020 and 22 weeks of partial closures during the 2020-2021 school year. The study found greater learning losses in reading than in mathematics and among boys rather than girls, with impacts found to be mixed for different socioeconomic backgrounds.
In Italy, learning losses due to online or hybrid learning were greater in reading than in mathematics. (1) This is somewhat inconsistent with previous research suggesting that learning in mathematics is more dependent on in-person instruction than reading, and that, in the EU, learning deficits in mathematics were significantly larger than in languages and other subjects. (2) As the study on Italy considered students of two different secondary level grades, this inconsistency could not be explained by the better ability of most parents to help children with reading rather than mathematics.
There were significant gender differences in the impact of physical school closures in Italy. On average, girls did better than their male peers. For reading and mathematics, the learning loss in 2021 was much lower for girls than for boys, with the gender gap widening further in 2022 and 2023 (Table A3.1 in the Annex). These results are consistent with both pre- and post-COVID-19 pandemic international patterns, with the difference ascribed to higher levels of intrinsic motivation and self-discipline among girls than boys.
Learning losses for students from different economic, social and cultural backgrounds presented a mixed pattern. The OECD measured parents’ socioeconomic and cultural status (ESCS measure). In Italy, students in the highest ESCS quartile suffered the largest learning loss in 2020-2021 (Table A3.1 in the Annex). (3) This finding for Italian students is inconsistent with EU-level findings (4) and may be due to differential baselines, with standardised scores for the lowest ESCS quartile considerably below the population average and thus potentially too low to register substantial losses due to remote learning measures.
Learning losses among Italian students in 2021 showed no significant difference in the impact of remote and hybrid learning between native students and students of foreign origin. In the first two years after the outbreak of the COVID-19 pandemic, natives and first- and second-generation migrants showed the same learning loss associated with online and hybrid learning in reading and mathematics (Table A3.1 in the Annex). However, in 2023, the learning loss among native pupils was larger than among their first- and second-generation migrant peers in mathematics.
Italian students’ performance did not bounce back to pre-COVID-19 pandemic learning levels upon their return to regular schooling modalities. Similar to other research, (5) the study does not detect signs of learning loss recovery. The considerable learning loss observed in 2021 remained largely stable in 2022 and 2023 compared to 2021. (6) If targeted policies do not mitigate learning losses and increase education systems’ resilience to disruption, the accumulation of losses could amplify the long-term negative effects on human capital and, eventually, the economy. (7)
- 1. (Amer-Mestre and Flisi, 2024). The study selected students in Italy at the end of upper secondary school (grade 13) as the treatment group, and students at the end of lower secondary school (grade 8) as the control. It implemented a difference-in-differences strategy to identify the causal effects of online and hybrid learning on performance, estimating the effect of school closures not only on one cohort who was in grade 13 in 2020-2021, but also for subsequent cohorts in grade 13 up to the 2022-2023 school year.
- 2. (European Commission, 2024e).
- 3. Due to data quality concerns for later years, only the impact on the 2020-2021 school year was analysed.
- 4. Immigrant or ethnic minority backgrounds of students and parents contributed to learning disadvantages during school disruptions (European Commission, 2022d).
- 5. (European Commission, 2024e)
- 6. The observed return of educational achievement to pre-COVID-19 pandemic levels in some countries (e.g. France) may have resulted from changes in examination content rather than learning recovery (European Commission, 2022d).
- 7. (Di Pietro, 2023).
Higher public expenditure on education is not sufficient to achieve better outcomes in basic skills pointing to other structural factors and the importance of reforms.Countries with higher levels of spending per student from ages 6-15 do not always have higher PISA scores. In 2022, Cyprus and Malta achieved similar average PISA scores compared to Bulgaria and Croatia, respectively, but at much higher cumulative expenditure per student (Chart 3.6). At the same time, some countries with similar levels of spending per student (Austria, Malta and Cyprus) achieved different levels of PISA scores (with the PISA score in Austria being 27 and 83 points higher than Malta and Cyprus, respectively). Prior to the COVID-19 pandemic, the efficiency of spending on education outcomes, as measured by the PISA efficiency score, (157) showed some smaller variations from year to year but remained quite stable on average (Chart 3.5). In 2022, the average PISA efficiency score recorded a sharper decline, with an additional investment of USD 1 000 per student translating into an increase of merely 5.1 points of average PISA scores, down from 6.8 points in 2018. Efficient spending ensures that resources are channelled to areas where they have sufficient impact and are spent effectively, enhancing value for money. Research suggests that reducing inefficiencies in spending on education could lead to substantial gains in the EU, with the potential to increase annual growth of GDP per capita by 0.8 pp in the long term. (158)
Chart 3.6
Higher expenditure on education does not necessarily lead to better PISA outcomes
Average EU PISA score across reading, mathematics and science, and cumulative expenditure on education per student from ages 6-15 (USD, PPP), 2022
Note: Data missing for Luxembourg (full set of data) and Estonia (cumulative expenditure per student).
Source: OECD PISA 2022 data.
Equity and excellence in education can be promoted at the same time.Evidence from PISA studies suggests that Member States with lower levels of underachievement tend to reach higher levels of top educational performance (i.e. a higher share of students achieving a high level of competence in a specific grade).(159) However, alongside the quantity of spending, other factors such as the quality of education and allocation of resources can also play a key role in determining the effectiveness of investment in education. In 2019, the Council (160) identified several policy actions that could help to improve efficiency, effectiveness and equity of education.(161) These include enhancing quality and equal opportunities through policies ranging from ECEC to higher education, VET and adult learning, fostering competencies in line with future labour market needs, reconsidering financing models for education and training (including through synergies with EU funds), and fostering synergies with complementary structural policies.
High-quality education for all will help Europe achieve its economic and social objectives and is thus a key element of the European Pillar of Social Rights. The Commission is working with Member States on the development of the European Education Area. This will be underpinned by a comprehensive policy, governance, and investment framework, EU-level targets on education and training to measure progress, and EU-level actions and funding to support European cooperation and national reforms in Member States.
Notes
- 151.See OECD's Programme for International Student Assessment (PISA) here It measures 15-year-olds’ ability to use their reading, mathematics and science knowledge and skills to meet real-life challenges.
- 152.(European Commission, 2019b).
- 153.(Ferreira and Martins, 2023).
- 154.(Eurofound, 2024b).
- 155.(European Commission, 2023g).
- 156.To approximate the average annual pace of learning among 15-year-olds, OECD (2023) used 20 score points in PISA as a common benchmark for all countries. However, this benchmark should not be used to convert the differences across countries and changes over time in PISA scores as years-of-schooling equivalents, given differences in the pace of learning at a given age, organisation of schooling, resources invested in education, and quality of education across countries, as well as possible changes in the pace of learning over time.
- 157.The definition of PISA efficiency score follows the methodology used in European Commission (2023d). It measures how many PISA points are produced by USD 1 000 (in PPP) of cumulative expenditure per student from age 6-15 and is calculated as 1 000 x (average PISA score over reading, mathematics and science)/cumulative expenditure per student from age 6-15 (in USD, PPP), and can be interpreted as a rough measure of the pay-off from additional expenditure for a school education system.
- 158.(European Commission, 2020c).
- 159.(European Commission, 2024e).
- 160.In its joint Economic and Financial Affairs (ECOFIN) and Education, Youth, Culture and Sport (EYCS) formation, 8 November 2019.
- 161.(European Commission, 2021b).