The Price of Shortchanging: Early Childhood Education (ECE)
One-paragraph takeaway
The United States is not saving a coin by under-investing in the first years of life; it is borrowing against its own future at a very high interest rate. Our best estimate is that the gap costs the economy about $350 billion a year (roughly 1.2% of GDP), with a plausible range of $250 to $550 billion, while closing most of it would take about $100 to $200 billion a year in smart, quality-focused spending. The return is not automatic: it depends on paying and keeping excellent teachers, keeping groups small, and building programs that are coherent from birth through third grade. The choice is not between spending and saving. It is between paying early, on purpose, or paying later, in remediation, lost earnings, and lost potential.
The economic cost of the U.S. gaps in paid parental leave, affordable child care, universal preschool, and full-day, right-sized kindergarten
1. Executive summary
The question.
What does the United States actually pay for having no national paid parental
leave, an expensive and fragmented child care market, partial preschool
coverage, and uneven full-day kindergarten, often in crowded classrooms?
The short answer.
The cost is large, recurring, and mostly hidden because it shows up in other
budgets: lower parental earnings, weaker employer productivity, remediation and
special education, reduced adult earnings and tax revenue, higher health and
justice costs, and slower growth. Our synthesis of the published evidence puts
the steady-state annual cost of the early-childhood investment gap at roughly $250 billion to $550 billion per year, with
a central estimate of about $350 billion (about 1.2% of GDP). Lifetime
losses attached to each year's birth cohort are of a similar order of magnitude
and are counted once, not twice (see Section 6).
The comparison that
matters. Closing most of the gap would cost on the order of $100 billion to $200 billion per year
in additional public and shared spending. Most rigorous benefit-cost studies of
high-quality programs find returns well above one dollar per dollar spent, with
long-run annual returns in the high single digits to low teens in the
best-studied programs. The strongest conclusion is therefore not "every
program pays for itself." It is that quality,
targeting, and implementation determine whether the return is 0x or 7x, and
that the current U.S. approach, which is cheap in the budget year and costly
over a lifetime, is the worst of both worlds.
Five findings.
1.
The U.S. is an
outlier. It is the only OECD country without national paid parental leave,
and it spends roughly half the OECD average share of GDP on early childhood
education and care.
2.
The largest
measurable near-term cost is parental labor supply. One widely cited
industry-research estimate puts the annual cost of the infant-toddler child
care crisis alone at about $122 billion in lost earnings, productivity, and tax
revenue.
3.
The largest
long-run cost is human capital. Gaps present at kindergarten entry persist,
compound, and show up as lower reading proficiency, higher dropout and special
education rates, and lower lifetime earnings.
4.
The public pays
twice. Taxpayers fund remediation, safety-net, and justice costs later
because they under-fund prevention earlier.
5.
The evidence is
strong but not uniform. Fadeout of test-score gains, and at least one large
randomized study with negative later outcomes, mean the case is for high-quality, well-implemented
programs, not for any spending whatsoever.
2. Scope, definitions, and method
2.1 What is being priced
The counterfactual is a system comparable to peer-country
norms and to the best evidence-based U.S. models. It has four levers:
|
Lever |
Age
span |
Current
U.S. gap |
|
A. Paid parental and family leave |
Birth to about 12 months |
No national program; FMLA unpaid and covers only part of
the workforce; paid leave in a minority of private jobs and a growing set of
states |
|
B. Affordable infant-toddler care |
0 to 3 |
Market-based, expensive, thin supply, low wages for staff |
|
C. Universal high-quality preschool |
3 to 4 |
Roughly one-third of 4-year-olds and a small fraction of
3-year-olds in state-funded pre-K; Head Start reaches a minority of eligible
children |
|
D. Full-day, right-sized kindergarten |
5 to 6 |
Most public kindergartners are in full-day programs, but
state requirements vary and class sizes and teacher supports differ widely |
2.2 How we keep the analysis MECE
We organize costs along two axes so that no dollar is
counted twice and nothing is left out.
Axis 1: Mechanism
(who or what is harmed first).
•
M1. Parents' labor market outcomes (earnings, career,
retention)
•
M2. Children's human capital (cognitive, language,
executive function, social-emotional)
•
M3. Fiscal and social-system costs triggered by weaker
outcomes (remediation, safety net, justice, health)
•
M4. Child health and welfare effects (birth outcomes,
maltreatment, early detection of delays)
•
M5. Macroeconomic and intergenerational effects
(growth, inequality, mobility)
Axis 2: Bearer (who
pays).
•
Households, employers, governments (federal, state,
local), and the broader economy.
Rules to prevent
double counting. Each effect is assigned to the single mechanism where it
first appears. Macro effects (M5) are reported as a memo item, not added to the
total, because they are the aggregate of M1 through M4. Fiscal costs are
reported as a transfer view
(taxpayer cost) and are not added to lost earnings except for the tax-revenue
component, which is a government loss rather than a household loss.
2.3 Two time frames
•
Annual flow:
what the economy loses in a typical year at current coverage.
•
Cohort lifetime:
the present value of losses over the life of one birth cohort of about 3.6
million children.
The annual flow is the most useful comparison to budget
decisions; the cohort view is the one that matches how education investments
actually pay back.
2.4 Method and honesty about uncertainty
This is a synthesis model that combines published estimates
and transparent assumptions, not a new econometric study. Ranges reflect
different discount rates, program quality, take-up, and attribution rules.
Every number should be treated as an order-of-magnitude planning estimate, and
the verification list in Section 11 names what to confirm before publishing.
3. The current state of play
3.1 International position
•
The United States is the only OECD member with no
national statutory paid maternity, paternity, or parental leave. Most peers
offer many weeks to many months of paid leave, often with job protection and
incentives for both parents.
•
Public spending on early childhood education and care
in the U.S. is roughly half the OECD average as a share of GDP, and far below
the Nordic countries and France.
•
Enrollment of 3-year-olds in organized early education
is well below the OECD average and far below the near-universal coverage in
countries such as France, Spain, Sweden, and Norway.
•
U.S. maternal labor-force participation, which once led
the world, fell behind several peers beginning in the 1990s. Research by Blau
and Kahn attributes a meaningful share of that relative decline to the absence
of family-friendly policies such as paid leave and part-time protections.
3.2 Household affordability
•
Average annual center-based infant care costs in many
states exceed in-state public college tuition. National averages in recent
years have been in the range of roughly $11,000 to $13,000 per child per year,
and far higher in high-cost states.
•
Families with low and middle incomes, who benefit most
from care, are least able to pay for it; subsidy eligibility is limited, and
only a minority of eligible children receive federal child care subsidies.
3.3 Supply and workforce
•
Child care and early education staff are among the
lowest-paid occupations in the country, with median wages near those of
fast-food and retail workers despite credential expectations.
•
Turnover is high, programs close or cut slots when they
cannot hire, and "child care deserts" cover large parts of rural
America and many urban neighborhoods.
•
This is the quality constraint: low wages and high
turnover directly undermine the teacher-child relationships that drive
developmental benefits.
3.4 Kindergarten
•
Kindergarten attendance is not compulsory in roughly
half of states, and requirements for districts to offer full-day programs vary.
Most children now attend full-day kindergarten, but the dosage, class size, and
curriculum quality differ widely.
•
Crowding matters. Evidence from the Tennessee STAR
experiment and follow-up work by Chetty and colleagues shows that early class
size and classroom quality in the earliest grades predict later college
attendance and adult earnings.
3.5 Early intervention
•
Only a small share of infants and toddlers are served
under the federal early intervention program (IDEA Part C), well below
estimates of the share with developmental delays. Missed early detection is a
direct feeder into later special education costs.
4. Why the early years carry outsized economic weight
4.1 Developmental science
The brain's architecture, including language, attention, and
emotional-regulation circuits, develops fastest between birth and age five.
Early experience, including responsive caregiving, language exposure, stable
routines, and stress levels, shapes that architecture. Chronic adversity and
under-stimulation raise the later cost of building the same skills.
4.2 The economics
James Heckman's work frames this as a return-on-investment curve that is steepest for the youngest
children: skills beget skills, so a dollar spent early compounds, while a
dollar spent on remediation later works against a larger accumulated deficit.
Heckman and colleagues estimate annual rates of return of roughly 7 to 10
percent for the Perry Preschool program and as high as the low teens for
comprehensive birth-to-five programs such as the Abecedarian Project. Even
conservative analysts, such as those at the Washington State Institute for
Public Policy, find positive benefit-cost ratios for many well-run pre-K
programs.
4.3 The reading pipeline
Outcomes at the end of third grade are a hinge. A widely
cited longitudinal analysis by Donald Hernandez found that children who were
not reading proficiently by third grade were about four times more likely to
leave high school without a diploma than proficient readers, with the risk
higher still for children who also lived in poverty. National reading results
remain stubbornly low: only about one in three fourth-graders reads at or above
the NAEP proficient level, and the share has not recovered to its earlier peak.
5. The evidence base, including the counter-evidence
A credible report must show where the evidence is strongest
and where it is contested.
Strong long-run
positive evidence
•
Perry Preschool
(Michigan) and Abecedarian (North
Carolina): small randomized trials with decades of follow-up showing higher
earnings, better health, and lower crime. Abecedarian follow-ups also found
improved adult cardiovascular risk markers.
•
Chicago
Child-Parent Centers: a large quasi-experimental study showing higher
educational attainment, lower special education placement, lower maltreatment,
and benefit-cost ratios commonly reported in the range of high single digits.
•
Boston and Tulsa
public pre-K: lottery-based and quasi-experimental studies showing gains in
school outcomes, with Boston evidence of higher high-school graduation and
college enrollment.
•
Head Start
long-run studies by Deming and by Bailey and colleagues: positive adult
outcomes, including educational attainment and economic self-sufficiency, even
where test-score gains faded.
•
Paid leave:
studies of California's program, and of programs abroad, find more
leave-taking, higher post-birth attachment to employers, and in several studies
better infant health and breastfeeding outcomes. Findings on long-run maternal
earnings are mixed and depend on leave length. Very long leaves can reduce
women's career progression, which is why design matters.
Contested or negative
evidence
•
Test-score
fadeout is common: early advantages in test scores often shrink by third
grade. The long-run gains seen in some studies suggest that non-cognitive
skills, behavior, and attainment persist even when scores do not, but this
remains an active debate.
•
Tennessee
Voluntary Pre-K randomized study found that children randomized into the
program did worse on several academic and behavioral measures by sixth grade.
It is a warning that low-quality or poorly aligned programs can fail or even
backfire.
•
Head Start
Impact Study: modest early gains that were mostly gone by the end of
kindergarten and third grade on measured cognitive outcomes.
•
"30 million
word gap" (Hart and Risley) is widely cited but its methods and
replication are disputed; we do not rely on it for any figure here.
Reading of the
evidence. The most defensible position is conditional: programs with trained, well-paid, stable teachers,
rich language and play-based instruction, strong curricula, small groups, and
family engagement show durable returns; programs that are large, underfunded,
and academically narrow show weak or negative returns. The cost of the status
quo is therefore real, but the benefit of any fix depends on funding it at a
level that preserves quality.
6. The cost stack
All figures are annualized, rounded, in approximately 2026
dollars, and are synthesis estimates with explicit logic.
M1. Parents' labor market losses (households, employers, government)
What happens.
Without paid leave and with expensive, scarce child care, parents,
disproportionately mothers, reduce hours, leave jobs, decline promotions, or
exit the labor force. Employers pay in turnover, absenteeism, and lost
productivity. Governments lose income and payroll taxes.
Evidence anchors.
•
A widely cited industry-research estimate puts the
infant-toddler child care problem alone at about $122 billion per year in lost earnings, productivity, and tax
revenue (ReadyNation / Council for a Strong America, 2023 update; a 2021
version estimated about $57 billion on narrower assumptions).
•
Child-penalty research (Kleven and colleagues) finds
long-run earnings gaps for mothers relative to fathers that are large in the
U.S., on the order of roughly 30 percent.
•
Only a minority of private-sector workers have access
to employer-paid family leave, and the lowest-wage workers have the least
access.
Our estimate.
Extending the $122 billion infant-toddler estimate to preschool-age children
and adding the leave gap yields $150
billion to $250 billion per year, of which roughly 20 to 30 percent falls
on governments as lost tax revenue, and the rest falls on households and
employers.
M2. Children's human-capital losses (households and the economy)
What happens.
Children who lack quality early care enter kindergarten with gaps in language,
executive function, and social-emotional skills. These gaps compound through
the reading pipeline, producing lower attainment and lower lifetime earnings.
Model logic.
•
Cohort size: about 3.6 million births per year.
•
Share of children meaningfully under-served relative to
a high-quality benchmark: we assume roughly 40 to 50 percent (all children from
low-income families plus a large share of middle-income families with no access
to quality).
•
Present-value earnings and productivity loss per
under-served child attributable to the early gap: about $40,000 to $120,000,
consistent with the lower half of published program benefit estimates after
applying a probability that a given program is fully effective.
•
Result: roughly $60
billion to $215 billion per cohort-year, which we round to $70 billion to $200 billion per year.
Cross-check.
McKinsey and Company's 2009 analysis of the U.S. achievement gap estimated that
the shortfall relative to better-performing countries cost the economy roughly
$1.3 to $2.3 trillion in 2008 (9 to 16 percent of GDP), with income-based gaps
alone costing $400 to $670 billion. Early childhood is one portion of that K-12
gap, so a slice of this size is consistent with the larger figure. Hanushek and
Woessmann's growth models imply cumulative gains in the tens of trillions of
present-value dollars if achievement were raised to the level of top-performing
countries.
M3. Fiscal and social-system costs (taxpayers)
What happens.
Weaker early outcomes lead to more grade retention, special education
placement, and remedial services; higher rates of high-school non-completion;
and more reliance on safety-net programs and the justice system.
Evidence anchors.
•
Studies of Perry, Chicago CPC, and several state pre-K
programs show reductions in special education placement, grade retention, and
criminal justice involvement.
•
Belfield, Levin, and colleagues estimate that
"opportunity youth" (young people neither in school nor working)
impose taxpayer and social costs on the order of tens of billions of dollars
per year for each cohort-age group, with lifetime costs well above $1 trillion
for the existing population.
•
Earlier work by Levin and colleagues puts the lifetime
public cost of each high-school dropout in the hundreds of thousands of dollars
when lost taxes, health spending, and justice costs are combined.
Our estimate.
After removing the lost-tax portion already included in M1 and M2, the
incremental fiscal cost attributable to the early-years gap is about $20 billion to $60 billion per year.
M4. Child health and welfare (households, Medicaid and public health
systems)
What happens.
Short or no paid leave is associated with lower breastfeeding rates, delayed
well-baby care, higher maternal stress and postpartum depression, and in some
studies worse infant health. Poor-quality or unsafe care raises injury and
neglect risks; missed early detection delays intervention for developmental
delays.
Evidence anchors.
CDC-based estimates place the total annual economic burden of child
maltreatment in the hundreds of billions of dollars. Early childhood programs
such as Chicago CPC reduced maltreatment, but only a fraction of that total is
plausibly preventable through early education and care.
Our estimate. $10 billion to $40 billion per year
attributable to the gap.
M5. Macroeconomic and intergenerational effects (memo item, not added)
Lower workforce participation, lower skills, and lower
innovation capacity reduce potential GDP, widen inequality, and reduce
mobility; weaker outcomes also carry into the next generation. These effects
are the aggregate expression of M1 through M4, so we report them as context and
do not add them.
Summary table
|
Mechanism |
Primary
bearer |
Low |
Central |
High |
|
M1. Parents' labor market |
Households, employers, governments |
$150B |
$190B |
$250B |
|
M2. Children's human capital |
Households, economy |
$70B |
$120B |
$200B |
|
M3. Fiscal and system costs |
Taxpayers |
$20B |
$35B |
$60B |
|
M4. Child health and welfare |
Households, public systems |
$10B |
$20B |
$40B |
|
Total (annual,
rounded) |
|
about $250B |
about $365B |
about $550B |
|
Share of U.S. GDP |
|
about 0.8% |
about 1.2% |
about 1.8% |
The low case sums to less than the simple column total
because we assume partial overlap in the low scenario; the central and high
cases assume overlap is already removed by the allocation rules in Section 2.2.
Cohort-lifetime view
Using a 3 percent real discount rate, the present value of
lifetime losses for one birth cohort is of the order of $300 billion to $800 billion, mostly from M2 and M3. This view is
the same cost expressed over a lifetime, not an additional cost, so it should
not be added to the annual flow.
7. Who bears the cost
|
Bearer |
Share
of central estimate |
Main
channels |
|
Households |
about 45% |
Lost earnings, out-of-pocket care costs, lower child
lifetime income |
|
Employers |
about 15% |
Turnover, absenteeism, lost productivity |
|
Governments |
about 25% |
Lost tax revenue, remediation, safety net, justice, health |
|
Wider economy |
about 15% |
Lower growth, weaker consumer demand, lower innovation |
The key political-economy insight is that the people who would pay for the fix
(taxpayers, in the budget year) are not the people who capture most of the
benefit (children, parents, and future taxpayers). That mismatch explains
much of why the U.S. under-invests.
8. The other side of the ledger: what closing the gap costs
|
Lever |
Rough
additional annual cost |
Notes |
|
A. National paid family leave (about 12 weeks, wage
replacement of about two-thirds, capped) |
$20B to $50B |
Typically financed through small payroll contributions,
often well under 1% of wages |
|
B. Affordable, high-quality infant-toddler care
(sliding-scale subsidy and workforce wage floor) |
$40B to $90B |
The largest single cost; driven by workforce pay and
ratios |
|
C. Universal high-quality pre-K for 3- and 4-year-olds |
$30B to $60B |
Net of existing state and Head Start spending |
|
D. Kindergarten quality (full-day access, class size, aide
support) |
$5B to $20B |
The smallest lever; many costs already in school budgets |
|
Total |
about $100B to
$200B |
About 0.3% to 0.7% of GDP |
Planning numbers vary widely by quality standard, wage
assumptions, and take-up; these are consistent in scale with published
proposals and cost analyses but are not a substitute for a detailed costing.
9. Net return, scenarios, and sensitivity
9.1 Scenarios
|
Scenario |
Program
quality |
Share
of gap closed |
Annual
cost |
Annual
benefit |
Benefit-cost
ratio |
|
Weak implementation |
Low wages, big groups, narrow academics |
40% |
$100B |
$50B to $100B |
0.5 to 1.0 |
|
Base case |
Solid quality, partial wage floor |
65% |
$150B |
$230B to $330B |
about 1.5 to 2.2 |
|
High-quality system |
Strong wages, small groups, play-rich curricula |
85% |
$200B |
$310B to $470B |
about 1.5 to 2.4 in the first decade, rising as cohorts
mature |
Note that benefits ramp over time. Leave and parental
labor-supply effects arrive within a year or two; child-outcome benefits arrive
over 10 to 40 years, so the benefit-cost ratio in the first decade understates
the lifetime ratio. Published lifetime benefit-cost ratios for the best-studied
programs range from roughly 3:1 to 10:1 or more.
9.2 What moves the answer most
1.
Quality and
workforce pay. The biggest driver; low-quality scale-up can erase most of
the return.
2.
Targeting vs
universality. Benefits per dollar are highest for low-income children;
universal programs gain political durability, broader labor-market effects, and
peer benefits, but at lower return per dollar.
3.
Discount rate.
At 7 percent, lifetime benefits shrink substantially relative to 3 percent.
4.
Fadeout and
attribution. If early gains fade more than assumed, M2 falls by a third or
more.
5.
Parental labor
supply response. The M1 estimate is sensitive to the share of caregivers
who would move into work and to wage assumptions.
10. Strategic implications
For policymakers
•
Treat early childhood as infrastructure with a measurable return, not as welfare spending.
Use multi-year, life-cycle budgeting so that early spending is scored against
later savings.
•
Sequence and
package. Paid leave (low cost, fast labor-market return) plus affordable
infant-toddler care (the largest cost, with the largest parent-labor effect)
plus universal pre-K and kindergarten quality create the full pipeline; funding
only one segment leaves leaks.
•
Fund the
workforce. Wage floors and credential pathways for early educators are the
quality lever that keeps returns positive.
•
Measure what
matters. Track kindergarten readiness, third-grade reading, attainment, and
adult outcomes, not only test scores in year one.
•
Use
pay-for-success and state-federal matching to reduce the mismatch between
who pays and who benefits.
For employers
•
Employer-supported leave and child care show positive
returns in retention and productivity, and are a competitive advantage in tight
labor markets.
For educators and
advocates
•
Anchor the case in outcomes the public already values:
literacy by third grade, high-school completion, workforce readiness, and
fiscal prudence.
•
Pair the numbers with the counter-evidence.
Acknowledging fadeout and the Tennessee findings builds credibility and
sharpens the argument for quality.
11. Limitations and verification list
This white paper synthesizes published research from
memory-informed knowledge and was not built from a live literature pull. Before
publishing, confirm each item below against its primary source and update to
the latest edition.
1.
ReadyNation / Council for a Strong America: the $122
billion annual infant-toddler child care cost (2023) and the 2021 comparison
figure.
2.
McKinsey and Company (2009), The Economic Impact of the Achievement Gap in America's Schools:
$1.3 to $2.3 trillion and the income-gap range.
3.
Heckman and colleagues: return-on-investment estimates
for Perry Preschool and Abecedarian; update to the most recent versions.
4.
OECD Education at
a Glance and Family Database:
U.S. ECEC spending as a share of GDP, enrollment of 3-year-olds, and paid leave
comparisons.
5.
NIEER State of
Preschool yearbook: current enrollment shares for 3- and 4-year-olds and
state funding.
6.
Child Care Aware of America: current national average
child care price.
7.
Bureau of Labor Statistics National Compensation
Survey: share of private-sector workers with paid family leave.
8.
NAEP: current fourth-grade reading proficiency.
9.
Hernandez (2011), Annie E. Casey Foundation:
third-grade reading and dropout risk.
10.
Chetty and colleagues (2011), How Does Your Kindergarten Classroom Affect Your Earnings?, and
Tennessee STAR follow-up.
11.
Durkin and colleagues, Tennessee Voluntary Pre-K
randomized study; Deming (2009); Bailey and colleagues (2021).
12.
Belfield and Levin on opportunity youth and dropout
costs; CDC maltreatment cost estimates; Kleven and colleagues on child
penalties; Blau and Kahn on female labor force participation.
13.
Cost figures in Section 8 should be checked against
recent Congressional Budget Office, Center for American Progress, and
state-level costings.
The modeling assumptions in Section 6 (share of children
under-served, present-value loss per child, overlap adjustments) are our own
and are the first things to test with sensitivity runs if this paper is used in
a formal setting.
12. One-paragraph takeaway
The United States is not saving a coin by under-investing in
the first years of life; it is borrowing against its own future at a very high
interest rate. Our best estimate is that the gap costs the economy about $350 billion a year (roughly 1.2% of GDP),
with a plausible range of $250 to $550 billion, while closing most of it would
take about $100 to $200 billion a year in smart, quality-focused spending. The
return is not automatic: it depends on paying and keeping excellent teachers,
keeping groups small, and building programs that are coherent from birth
through third grade. The choice is not between spending and saving. It is
between paying early, on purpose, or paying later, in remediation, lost
earnings, and lost potential.
