Measuring and forecasting progress in education: what about early childhood?

The paper addresses a critical gap in recent global analyses of educational progress: the exclusion of SDG Target 4.2, which calls for universal access to quality early childhood development, care, and pre-primary education by 2030. While prior work indicates advances toward universal primary education and narrowing gender gaps, substantial disparities persist—particularly in secondary education and in regions such as sub-Saharan Africa, North Africa, and the Middle East. The authors argue that learning begins before birth and that inequalities are evident from the earliest years, making early childhood development foundational to lifelong learning and achievement of SDG 4. The study’s purpose is to highlight the importance of early learning, document disparities, and provide new empirical analyses demonstrating associations between pre-primary participation and adolescent academic outcomes, as well as the macroeconomic costs of failing to universalize pre-primary education.

The literature underscores rapid brain development and learning in early life, driven by responsive, nurturing interactions in homes and early care settings. The Nurturing Care Framework articulates the multi-dimensional supports—health, nutrition, safety, responsive caregiving, and early learning—that underpin early development. Evidence from LMICs shows pervasive early disadvantage (poverty, stunting, limited stimulation, low pre-primary attendance) with strong socioeconomic and rural-urban gradients, while gender gaps in early indicators are small. Pre-primary enrolment has increased globally since 2000, but remains low in sub-Saharan Africa and varies with national legal frameworks (free and/or compulsory provisions correlate with higher enrolment and better primary completion). Research documents complementarities between early and later skills, with universal, high-quality ECE reducing learning gaps. Longitudinal studies show that pre-primary participation enhances later academic performance and schooling attainment. Policy levers (e.g., paid parental leave, minimum wages, tuition-free pre-primary, targeted income support) can mitigate early disadvantages and improve developmental outcomes.

The study comprises two analytical components. 1) Pre-primary attendance and adolescent test scores: Using the 2018 PISA dataset (n=430,264 adolescents across 73 middle- and high-income countries), the authors examined associations between retrospectively reported years of pre-primary education (ISCED-0) and standardized mathematics and science scores at age 15. Plausible values were averaged and standardized (mean 0, SD 1). Multivariate regressions estimated the associations of one year and two or more years of pre-primary (binary indicators) with test scores, controlling for age, sex, household socioeconomic status (wealth index), parental education, age of school entry, plus country and subnational fixed effects (geographical region or school type). Separate models were estimated by income groups (lower-middle: N=6, upper-middle: N=24, high: N=43) and by region (East Asia & Pacific: N=11; Europe & Central Asia: N=42; Latin America & Caribbean: N=10; Middle East & North Africa: N=8; North America: N=2). The authors caution that these are non-causal associations, though robustness checks with fixed effects and alignment with prior causal estimates increase confidence. 2) Cost of inaction (COI) simulation: For 134 countries, the authors estimated the present discounted value of forgone future income (net of programme costs) if pre-primary enrolments remain at 2018 levels rather than reaching universal coverage (SDG 4.2.2). Inputs included: 2018 UNESCO gross enrolment rates (GER) for current coverage; target coverage set to 100% or higher if GER exceeded 100% (to account for catch-up); assumed causal impact of pre-primary on lifetime earnings set at 8% (conservative relative to small-scale experimental estimates ~14%); discount rate 3%; and per-child programme costs adjusted across countries using PPP conversion factors to reflect wage/price level differences. The simulation sums discounted earnings gains over working life (from entry at age 18 for 45 years) and subtracts programme costs for the additional children needed to reach universality. Sensitivity analyses varied the assumed earnings impact (−1 to −2 percentage points), costs (+10%/20%), and discount rate (4–5%).

• Association with adolescent achievement: Relative to <1 year or no pre-primary, 1 year of pre-primary was associated with +0.10 SD (95% CI 0.09, 0.11) in mathematics at age 15; 2+ years with +0.22 SD (95% CI 0.21, 0.23). For science: +0.09 SD (95% CI 0.08, 0.10) for 1 year; +0.20 SD (95% CI 0.19, 0.21) for 2+ years. Associations were slightly larger in lower-middle- and high-income countries than in upper-middle-income countries; strongest in East Asia & Pacific; weakest in Europe & Central Asia and North America.
• Cost of inaction (COI): Median COI of not reaching universal pre-primary coverage (as % of GDP): high-income 0.58%; upper-middle-income 2.54%; lower-middle-income 6.24%; low-income 9.06%. Losses are inversely related to country income level and often exceed annual governmental education expenditures (averages: 4.1% of GDP in low-income; 4.4% lower-middle; 4.3% upper-middle; 5% high-income).
• Coverage trends and disparities: Global pre-primary gross enrolment rose from 35% (2000) to 62% (2019), but was only 32% in sub-Saharan Africa (2019). Within-country disparities by wealth and rural/urban status are large, with poorer rural households most disadvantaged.
• Policy and legal context: Among 194 countries, 68 legally mandate free and/or compulsory pre-primary; 46 are both free and compulsory. Tuition-free pre-primary is associated with 16% higher gross enrolment; providing at least one year free and compulsory is associated with 10% higher primary school completion.
• Financing gaps: Most LMIC governments allocate <5% of education budgets to pre-primary; donors allocate ~2% of basic education aid to pre-primary. Households bear large out-of-pocket costs (e.g., 63% in Nepal; 100% in Uganda).
• COVID-19 impacts: Six-month pre-primary closures are estimated to reduce future earnings by ~2.5% of GDP for current preschool-age cohorts; 12-month global closures may push >22 million additional children behind developmentally.

Findings suggest that pre-primary participation is associated with meaningful gains in adolescent cognitive outcomes and that delaying universalization of pre-primary imposes substantial macroeconomic costs, especially in lower-income countries where losses may exceed current total education spending. Given strong complementarities between early and later skills, early childhood care and education can enhance subsequent learning, narrow socioeconomic gaps, and improve human capital formation. However, current investment levels are inadequate: government allocations to pre-primary are low, donor support is minimal, and households shoulder heavy financial burdens, exacerbating inequities. To meet SDG 4 holistically and reduce inequalities—particularly in the context of COVID-19 disruptions—countries should prioritize quality early childhood care and education and adopt supportive family policies (paid parental leave, minimum wage policies, tuition-free pre-primary, and income support), coupled with improved programme quality and data systems to monitor early development and learning.

The study brings early childhood to the forefront of SDG 4 monitoring by (1) demonstrating positive associations between pre-primary participation and adolescent mathematics and science performance across 73 countries and (2) quantifying the economic costs of failing to reach universal pre-primary coverage across 134 countries. The results underscore that investing in early childhood development and learning yields substantial individual and societal returns and that the costs of inaction are high, particularly in lower-income contexts. To accelerate progress, governments and donors should increase and prioritize funding for quality pre-primary programmes, implement supportive family policies, and strengthen data systems to track early development and access. Future research should generate more causal evidence at scale on programme quality and equity impacts, refine country-specific COI parameters (including broader societal benefits and general equilibrium considerations), and evaluate policy packages that integrate early childhood with primary and secondary education systems.

• The PISA-based analyses are observational and non-causal, despite controls and fixed effects; unobserved confounding may remain.
• COI simulations include only adult earnings effects and omit other benefits (e.g., reduced crime, maternal labour force participation), likely making estimates conservative; general equilibrium effects with scale-up could lower returns.
• Parameter assumptions (e.g., 8% earnings impact, 3% discount rate, PPP-adjusted costs) are stylized and may not capture country-specific contexts; GER-based coverage can over/underestimate needs due to age structure and catch-up enrolment.
• Data availability constraints (e.g., country coverage) and heterogeneity in programme quality limit generalizability; results should be refined for country-specific policymaking.
• COVID-19-related projections carry uncertainty given evolving pandemic dynamics and policy responses.