Early childhood education produces the highest social return on investment

The claim

High-quality early childhood education programs — particularly those targeting children from low-income and otherwise disadvantaged families — produce lifetime benefits so large relative to program cost that they represent the single best return on public investment available. The claim, formalized by Nobel laureate economist James Heckman as the “Heckman curve,” holds that the earlier an investment in human capital is made, the higher its return: investments in infants and toddlers yield more than investments in school-age children, which in turn yield more than investments in adolescents or adults. By this account, the United States chronically underinvests in early childhood while overinvesting in remediation and incarceration — a policy pattern that is both economically irrational and structurally regressive, since market provision of quality early care is accessible primarily to families who can pay for it.

Proponents of the claim point to three converging streams of evidence: experimental evidence from model programs, quasi-experimental evaluations of large-scale public programs, and cross-national comparisons between the US and peer nations with universal early childhood systems. They argue that the benefits — higher adult earnings, lower crime rates, better health outcomes, reduced dependence on public assistance — accrue primarily to participants but generate large fiscal externalities that repay public investment many times over.

The mechanism

The underlying causal mechanism runs through neuroscience and skill formation economics simultaneously. Human brain development is not uniform across the lifespan. The first five years of life are characterized by exceptionally rapid synaptogenesis — the formation of neural connections — and by elevated neuroplasticity in domains including language processing, executive function, and emotional self-regulation. These domains are what economists call “foundational skills”: once formed, they lower the cost of acquiring all subsequent skills, and once absent or stunted, they are expensive to remediate.

Heckman’s skill formation model formalizes this as a dynamic complementarity: early skills beget later skills. A child with secure attachment, adequate vocabulary, and basic executive function at age 5 enters kindergarten at a learning advantage that compounds through the school years. A child who arrives at kindergarten with deficits in all three enters a remediation process that rarely fully closes the gap and consumes disproportionate instructional resources. The mechanism implies that programs which intervene early — providing cognitive stimulation, language-rich environments, responsive caregiving, and nutritional and health support — shift the developmental trajectory at a moment when the marginal return to investment is highest.

The market failure that motivates public intervention is straightforward: the benefits of early childhood investment accrue over decades and partly to third parties (lower crime rates, higher tax revenues, reduced public health costs). Parents, particularly low-income parents facing severe liquidity constraints, cannot privately finance the level of investment that would be socially optimal. Private market provision of quality early care is expensive — median full-time childcare costs in the US exceed $15,000 per year in most states, exceeding average public university tuition — and quality is difficult for consumers to verify, creating adverse selection problems. The result is chronic market underprovision of quality care for children from low-income families precisely during the developmental window when the return to investment is highest.

The evidence

Perry Preschool: the foundational RCT

The High/Scope Perry Preschool Program, begun in 1962 in Ypsilanti, Michigan, remains the most studied early childhood intervention in history. Researchers randomly assigned 123 African-American children from low-income families to treatment (high-quality half-day preschool plus weekly home visits) or control conditions. The program ran for two years, at ages 3 and 4. Follow-up data have been collected at ages 14, 19, 27, and 40.

The age-40 results, analyzed by Heckman and colleagues (2010) using intent-to-treat identification, showed that treatment participants had significantly higher employment rates, higher median earnings, lower rates of criminal arrest (with arrests for violent crime 32% lower in the treatment group), higher rates of high school graduation, and better self-reported health. The IQ advantage of the treatment group, which had been large at age 5, faded to statistical insignificance by age 8 — a finding that initially seemed to undercut the program but which Heckman reinterpreted as evidence that the program’s lasting effects operated through non-cognitive channels: persistence, self-regulation, and social competence. The economic return, accounting for program costs, crime reduction savings, higher tax contributions, and reduced public assistance, was estimated at $7–12 per dollar invested in 2010 dollars.

Abecedarian Project: intensive intervention from infancy

The Abecedarian Project, a randomized trial begun at the University of North Carolina in 1972, was more intensive than Perry: treatment children received full-day, year-round, high-quality childcare from infancy (average entry age: 4.4 months) through age 5, plus a subsequent school-age phase. The intervention was unusually comprehensive, including nutritional support, health screening, and a structured language-enrichment curriculum.

By age 30, Abecedarian treatment participants showed substantially better outcomes across multiple domains. Campbell and colleagues (2002) reported that treatment participants were 23% more likely to be in skilled employment, had completed more years of education, had a 26% lower rate of depressive symptoms, and showed lower rates of tobacco and drug use. A 2012 follow-up by Garcia and colleagues estimated a benefit-cost ratio of $7.30 per dollar invested after accounting for criminal justice savings, health cost differences, and earnings gains. Critically, the effects on health outcomes — which persist and grow into adulthood through pathways including lower rates of hypertension and metabolic syndrome — suggest that early childhood environments affect biological developmental trajectories in ways that have long-run consequences for healthcare costs.

Chicago Child-Parent Centers: scaling up

The Chicago Child-Parent Centers (CPC), operating since 1967 in Chicago public schools, provide a quasi-experimental test of whether model-program effects survive at larger scale. Reynolds and colleagues (2002) used a matched comparison design to evaluate the effects of CPC participation (beginning at age 3) on a cohort followed to age 24. CPC participants showed a 29% lower rate of violent arrest, higher rates of high school completion, and lower rates of special education placement than matched comparisons. The benefit-cost analysis estimated $10.83 returned per dollar invested — higher than Perry, partly because CPC operates at lower per-child cost than the intensive model programs.

The CPC evidence is particularly important because the program operates through the public school system at realistic funding levels and staff-to-child ratios, suggesting the returns estimated from model programs are achievable in public-sector implementation with adequate quality standards.

Head Start: the contested evidence

The Head Start Impact Study (HSIS), a nationally representative RCT conducted by the Department of Health and Human Services with approximately 5,000 children, produced a substantially more modest finding: cognitive and social-emotional benefits at the end of the Head Start year largely faded by first or third grade. By third grade, differences between treatment and control groups were small and mostly statistically insignificant.

This finding has been widely cited as evidence that Head Start does not work. The interpretation is contested. Program quality in Head Start varies substantially across its roughly 1,700 grantees, and the HSIS estimated the effect of access to Head Start against the background that many control children accessed other childcare and preschool programs — narrowing the treatment-control contrast. Studies of specific high-quality Head Start centers, and long-run administrative data analyses by Ludwig and Miller (2007) and Deming (2009), found substantial positive effects on educational attainment, health, and criminal outcomes. Deming’s (2009) analysis of the National Longitudinal Survey of Youth found that Head Start participation, compared to parental care, produced effects on a composite outcome index approximately one-third the size of Perry Preschool’s — modest but real and cost-effective.

The Head Start evidence does not refute the core claim but qualifies it: the return to early childhood investment is strongly contingent on program quality. Underfunded programs with high staff turnover and poor curriculum implementation do not produce the returns observed in well-resourced model programs. This is an argument for investment in quality, not a refutation of the mechanism.

Neural development and sensitive periods

The neuroscientific evidence for sensitive periods in early development has strengthened considerably since the Perry Preschool era. Studies using MRI and EEG technologies have documented that language network development shows maximal plasticity in the first three years of life; executive function networks in the prefrontal cortex develop most rapidly between ages 2 and 7; and the stress-response system (HPA axis calibration) is substantially determined by caregiving quality in the first two years of life. Children raised in conditions of chronic stress — poverty, housing instability, parental mental health crises — show altered cortisol profiles and executive function deficits that track the neurobiological disruptions these stressors produce. High-quality early childhood programs that provide stable, responsive caregiving reduce cortisol dysregulation and improve executive function in children from high-stress backgrounds, with effects observable at the neurological level.

Cross-national evidence

France provides the clearest natural comparator. The French écoles maternelles system provides universally available, publicly funded preschool from age 3 (and increasingly from age 2 in disadvantaged areas), attended by over 97% of three-year-olds. France’s PISA results show an SES gradient — family background still predicts outcomes — but the gradient is substantially smaller than in the US. OECD analysis attributes part of this difference to the compression of early childhood cognitive disadvantage achieved by universal preschool provision.

Denmark and Sweden offer an even more comprehensive model: publicly subsidized, high-quality childcare is available from approximately 12 months of age, with parent fees capped as a percentage of income. Enrollment rates for children aged 1–5 exceed 90%. Denmark’s 1990s childcare expansion, which reached previously unserved families in rural areas, has been analyzed as a quasi-experiment: children in municipalities that gained access to subsidized childcare showed significantly better long-run educational outcomes than comparable children in municipalities that did not, with effects concentrated among children from lower-income families.

Germany’s 2013 legal right to a childcare place from age 1 provides a more recent natural experiment. Early analyses suggest it improved maternal employment and child development outcomes, though the effects were heterogeneous by childcare quality. Finland’s near-universal subsidized childcare, combined with a year of paid parental leave, means that Finnish children from low-income families enter kindergarten with less developmental deficit than their American counterparts — a structural precondition of Finland’s celebrated educational equity outcomes.

The US, by contrast, has no universal childcare or pre-K entitlement. Publicly funded preschool programs cover roughly 30% of four-year-olds in state pre-K programs; enrollment among three-year-olds from low-income families is far lower. Childcare deserts — defined as areas with more than three children per licensed childcare slot — encompass 43% of American children under 5. This geographic and income-based rationing of quality early care is the direct structural cause of the developmental gap that arrives at kindergarten door.

Who benefits

The concentrated interest groups opposing universal early childhood investment include: the private, for-profit childcare industry, which opposes public subsidy structures that would impose quality regulations on their operations; the US Chamber of Commerce and member businesses, which have consistently opposed employer childcare mandates and paid family leave provisions that would reduce the cost burden on parents; and the conservative policy network — including the Heritage Foundation, Cato Institute, and American Enterprise Institute — which has emphasized the fading Head Start effects as evidence against public pre-K expansion, while opposing the increased per-child spending that would close the quality gap between model programs and average public programs.

The political economy of inaction also benefits from geographic dispersion of the childcare desert problem: childcare access crises are worst in rural and low-income urban areas, populations with lower political representation in state legislatures, which control most early childhood policy. Employer coalitions in sectors that rely on low-wage female labor — retail, food service, domestic work, care work — have structural incentives against childcare subsidies that would reduce the labor supply of mothers with young children and tighten labor markets in those sectors.

The counter

The strongest challenge to the cost-benefit claims is the external validity problem: the two most-cited RCTs, Perry Preschool and Abecedarian, enrolled very small samples (58 and 57 treatment children, respectively) in intensive programs run by motivated researchers at relatively high per-child costs. The generalizability of $7–12 returns from these programs to large-scale public implementation is not guaranteed, and the Head Start evidence provides a cautionary example of effects that attenuate substantially at scale.

A second legitimate challenge concerns the opportunity cost of early childhood investment. The Heckman curve claims that early investment has higher returns than later investment — but this comparison depends on what the late investment is. Rigorous interventions in adolescence (high-dosage tutoring, career and technical education, certain workforce development programs) also show substantial returns. The binary framing of early versus late can obscure that multiple points in the life course are responsive to intervention.

A third challenge is distributional: universal pre-K programs, while often justified on equity grounds, deliver per-child subsidies to middle- and upper-income families who were already purchasing quality care privately. The marginal return to additional investment is concentrated in the most disadvantaged children, suggesting that means-tested expansion of high-quality targeted programs may generate larger returns per dollar than universal programs — at the cost of political coalition breadth and potential stigmatization.

These are genuine complications, not refutations. The evidence base for early childhood investment is broader and stronger than for almost any other social program. The question is not whether to invest but how — in what quality standards, at what coverage level, and through what institutional design — to replicate the returns observed in the best programs.

References

Heckman, J. J., Moon, S. H., Pinto, R., Savelyev, P. A., & Yavitz, A. (2010). The rate of return to the HighScope Perry Preschool Program. Journal of Public Economics, 94(1–2), 114–128. https://doi.org/10.1016/j.jpubeco.2009.11.001

Schweinhart, L. J., Montie, J., Xiang, Z., Barnett, W. S., Belfield, C. R., & Nores, M. (2005). Lifetime effects: The HighScope Perry Preschool study through age 40. HighScope Press.

Campbell, F. A., Ramey, C. T., Pungello, E., Sparling, J., & Miller-Johnson, S. (2002). Early childhood education: Young adult outcomes from the Abecedarian Project. Applied Developmental Science, 6(1), 42–57. https://doi.org/10.1207/S1532480XADS0601_05

Reynolds, A. J., Temple, J. A., Robertson, D. L., & Mann, E. A. (2002). Age 21 cost-benefit analysis of the Title I Chicago Child-Parent Centers. Educational Evaluation and Policy Analysis, 24(4), 267–303. https://doi.org/10.3102/01623737024004267

Deming, D. (2009). Early childhood intervention and life-cycle skill development: Evidence from Head Start. American Economic Journal: Applied Economics, 1(3), 111–134. https://doi.org/10.1257/app.1.3.111

Ludwig, J., & Miller, D. L. (2007). Does Head Start improve children’s life chances? Evidence from a regression discontinuity design. Quarterly Journal of Economics, 122(1), 159–208. https://doi.org/10.1162/qjec.122.1.159

Heckman, J. J. (2006). Skill formation and the economics of investing in disadvantaged children. Science, 312(5782), 1900–1902. https://doi.org/10.1126/science.1128898

Garcia, J. L., Heckman, J. J., Leaf, D. E., & Prados, M. J. (2020). Quantifying the life-cycle benefits of an influential early-childhood program. Journal of Political Economy, 128(7), 2502–2541. https://doi.org/10.1086/705718

Havnes, T., & Mogstad, M. (2011). No child left behind: Subsidized child care and children’s long-run outcomes. American Economic Journal: Economic Policy, 3(2), 97–129. https://doi.org/10.1257/pol.3.2.97

OECD. (2023). Education at a glance 2023: OECD indicators. OECD Publishing. https://doi.org/10.1787/e13bef63-en