A trillion-dollar leak the federal ledger ignores
Decades of consensus research — most prominently the National Academies of Sciences, Engineering, and Medicine reports on the costs of child poverty — converge on a number near $1.03 trillion per year. That is roughly 5.4% of US gross domestic product. It is the largest quantifiable economic loss the federal government does not appear on any budget line, because it accrues as productivity that never occurs, taxes never collected, and prison and hospital costs paid elsewhere on the ledger.
The loss compounds through three documented channels.
Sources: NASEM consensus studies; U.S. Treasury; Joint Committee on Taxation; independent economic policy research centers. The cost is aggregated across direct fiscal outlays, foregone tax revenue, and economic productivity that never materializes.
$135–180 billion buys the whole problem
The 2021 temporary expansion of the federal Child Tax Credit lifted millions of children out of poverty at an annualized cost in the $100–120 billion range. That episode demonstrated, at policy scale, that direct cash transfers and tax supports reduce child poverty dramatically and quickly.
Extending and stabilizing that mechanism produces two well-modeled funding tiers:
| Annual outlay | What it buys | Source family |
|---|---|---|
| ~$109B / yr | Universal supports & work package — cuts child poverty in half. | NASEM consensus model |
| $135–180B / yr | Targeted child allowances closing the gap between current household incomes and the federal poverty line — eradicates child poverty. | NASEM + Treasury models |
The cost-benefit ratio is maximized at complete eradication, not at the halfway tier. The marginal cost of lifting the remaining percentiles of children out of poverty is heavily outweighed by the societal savings those children produce as adults. Stopping at the lower tier leaves billions of dollars in economic gains uncaptured for the sake of a budget line that is already an order of magnitude smaller than the cost it offsets.
The Accord's delivery vehicles are sized to this envelope. The Social Stack anchors the Universal Child Allowance (beginning at $800 per child per month, over $1,000 in high-cost regions, tapering with child number and age), Baby Bonds ($19,000 by age 18), the Childcare Plan (ages 0–5, 50/25/25 operating split), and Skills Wallet ($20,000 lifetime credential fund). Each rides on FedCard and arrives automatically — no application, no caseworker, no eligibility cliff.
The math that should end the debate
Investments in children require upfront funding. The bulk of the societal benefits — adult earnings, reduced incarceration, lower healthcare utilization — accrue years or decades later. Honest cost-benefit accounting applies a discount rate to delayed benefits. At the federal long-term cost of capital (a defensible 3% annual discount) and a 15-year horizon to typical adult workforce entry or criminal-justice-system exposure age:
A 4.49-to-1 present-value ratio survives almost any reasonable parameter perturbation. Doubling the discount rate to 6% still leaves a ratio above 2.9. Extending the lag to 20 years still leaves a ratio above 3.9. There is no defensible set of inputs at which the investment under-performs the status quo.
That is the test a public expenditure should pass. It does, by a wide margin.
The funding is already in the architecture
A $135 billion annual program is well below the revenue mobilized by the Accord's existing tax architecture, which closes the longest-running avoidance channels in the code. None of the offsets below are speculative new instruments. Each closes a current carve-out, ends a deferral, or applies an existing tax to its full base.
| Accord instrument | What it does | Order-of-magnitude |
|---|---|---|
| Flat 28% payroll tax, uncapped | Replaces FICA. A single rate on all compensation — wages, bonuses, equity, options, perks, partnership distributions — with no $168,600 cap. The PDF's “remove the FICA cap” line ($320B/yr) is structurally part of this instrument; the cap simply does not exist in Accord canon. | $300B+ |
| Eliminate stepped-up basis at death | Unrealized capital gains are taxed at inheritance rather than erased. The Joint Committee on Taxation estimates this loophole alone costs about $72.5 billion annually, with the benefits accruing overwhelmingly to the largest estates. | ~$72B |
| Progressive rate ladder above $1M | The unified income tax extends graduated brackets into the extraordinary-income range. The PDF's “5–10% surtax on incomes over $1M” concept is delivered as the top of the progressive ladder rather than a bolted-on surtax. | $60–80B |
| Carried-interest closure | Private equity and hedge fund performance fees are taxed as ordinary compensation rather than capital gains. A small line item alone, but emblematic — and recurring revenue. | $9–10B |
The four mechanisms together mobilize roughly $440–460 billion a year — more than triple the cost of eradicating child poverty. The remainder flows into the Accord's 50-year debt-retirement schedule under the Debt Sunset Governor.
For the full revenue architecture (the lifecycle capture across compensation, income, consumption, externalities, wealth, and settlement), see /taxladder.
Tolerating the loss is the radical choice
The framing fight matters. The conventional posture treats child-poverty investment as a discretionary add-on — something the country may fund when the budget permits. That framing inverts the math. The economy is already paying a trillion dollars a year. The question is whether it keeps paying after the productive capacity is lost or pays a fraction up front and recovers the productive capacity.
Investing in child capacity is not new. It is also not partisan in any historical sense. The Land-Grant Acts, the GI Bill, the school lunch program, and the universal expansion of secondary education all proceeded on the same balance-sheet logic: nations that fail to develop the next generation's capacity decline. Nations that invest in it compound.
The Accord's position is straightforward. The country can keep paying $1.03 trillion a year in foregone productivity, criminal-justice cost, and avoidable healthcare load. Or it can spend $135 billion a year on the children themselves, recover the productive capacity, and retire the debt that the wasted potential helped to compound. The second option is the fiscally conservative one.
The Accord's broader causal framing applies here too: a century of asymmetric incentives accumulated into the present arrangement. Each institution that holds child poverty in place was built for some other purpose and was never tended after the world changed around it. The repair is overdue maintenance.
Where this connects in the Accord
- /social-stack — Engine 3. The delivery architecture: UCA, Baby Bonds, Childcare, Skills Wallet, Social Security 2.0, all riding on FedCard.
- /calculator/family — household-level impact calculator. Income, household size, region; the calculator runs the program-by-program math against the status quo.
- /taxladder — the full revenue architecture. Where the offsets above sit inside the lifecycle-capture model.
- /scenarios — eight household profiles modeled at scale. Three of the eight are families with children at varying income tiers.