The Ledger

Part One: How We Pay

The baseline, stated plainly

In fiscal year 2025 the federal government collected $5.2 trillion and spent $7.0 trillion, leaving a deficit of $1.8 trillion, about 5.9 percent of the economy. That is the ledger this platform inherits. Any agenda that pretends the starting point is a blank checkbook is not being honest, and any critic who pretends the current ledger is healthy is not being honest either. The status quo is not free. It is running $1.8 trillion a year in the red while leaving people uninsured, unhoused, and uncounted.

Where the money is

Before this platform asks anyone for a new dollar, it goes where the money already is. Five families of financing run through every brief on this site, and every big-ticket proposal names which family it draws on.

1. Collect what is already owed. The IRS projects the gross tax gap, tax legally owed but not paid, at $696 billion per year, with about $606 billion never recovered. Read that number the honest way: it measures the size of the problem, not the size of the available money, because the net gap is by definition the part enforcement does not reach, and the scored yield of rebuilt enforcement is what the ten-year table below actually books, on the order of $127 billion per decade per funding tranche, not hundreds of billions a year. A disproportionate share of unreported income is business and pass-through income, which is concentrated among higher earners, where audit coverage was allowed to wither; separate Treasury-affiliated research estimates the top of the distribution accounts for an outsized share, and that estimate is distinct from the IRS gap figure itself.
2. End the preferences that fail the test. Federal tax expenditures, the deductions, credits, and carveouts written into the code, sum to roughly $2.2 trillion a year (FY2025), and almost none of them are ever asked to prove they serve a public purpose. The Tax Code Public Value Test in Pillars 2 and 7 exists to keep the ones that earn their keep and end the ones that are just extraction with a filing requirement.
3. Stop paying twice for failure. The GAO counted $186 billion in improper federal payments in fiscal 2025 alone, up $24 billion from the year before, and its own caveat is that the true number is higher. Add the damage the public absorbs because pollution is priced at zero, and the money wasted before any new program is even proposed runs to hundreds of billions a year. Pricing pollution honestly, as Pillar 7 proposes, both cuts that damage and raises revenue while doing it.
4. Honest new revenue, named as such. Some of this platform requires real taxes, and we say so. Fixing the cap that lets the highest earners stop paying into Social Security partway through the year is a tax change. Taxing wealth no more gently than work is a tax change. We think both are fair, and we refuse to disguise either as something else.
5. Purchases that return more than they cost. The child allowance, universal school meals, the care agenda, and the tutoring corps are purchases. The evidence, pinned throughout the briefs and the Evidence page, is that they are purchases with returns, in child poverty cut nearly in half within a single year when the expanded Child Tax Credit was tried, and in lifetime earnings, health, and reduced crime. But a return on investment is not the same as free, and this platform does not use the words interchangeably.

What these numbers are, and are not

The five families above are honest pools, not a budget, and pretending otherwise would break this page's own rule. The $696 billion tax gap is a gross figure; about $606 billion stays uncollected after enforcement, and no enforcement regime collects it all, because audits cost money and returns diminish. Improper payments are not all fraud and not all recoverable; the GAO's category includes documentation failures and even underpayments. The $2.2 trillion in tax expenditures is not a pot to sweep; ending a preference changes behavior, interacts with other taxes, and some expenditures are policy commitments this platform would defend under its own Public Value Test. And a social return on investment is not a federal budget offset: a program that repays society several times over can still require an appropriation every year. What this page establishes is where the money is and what the platform refuses to pretend. The ten-year table below is the first stage of the full model, provisional and labeled as such until it has independent review.

The ten-year table (v1.2)

The rules: a number appears only if an official scorekeeper (CBO, JCT, the Social Security actuaries) or a clearly labeled independent estimate published it. Score windows differ by row and are stated. Rows without an honest score stay blank and say so, because a blank is more useful than a guess. The whole table is downloadable as JSON, and independent scoring or challenge is invited through the report form.

Revenue and savings

Spending

What the v1.2 table shows, read honestly: every row is now read against one canonical ten-year window, 2026-2035; rows whose official score used an earlier decade say so and are carried as order-of-magnitude for the canonical window, which for spending rows likely understates nominal costs, the conservative direction. Any total therefore remains approximate, never a precise sum. Read that way, and using net figures where dedicated revenue exists (paid leave nets about $228B after its own payroll tax) while setting aside what cannot honestly be pooled (the Social Security cap fix is trust-fund revenue that strengthens Social Security, not general funds; the pollution price returns as dividends), the picture is stark: roughly $2.4 trillion of scored and estimated general-fund spending against roughly $600 billion of scored general-fund revenue. And read the gap as a floor, not a ceiling: re-basing the older-window spending rows to 2026-2035 would widen it, and the healthcare shift sits entirely on top of it, at $1.5 to $3 trillion per year of federal ledger movement presented separately below. That gap is not a secret to hide; publishing it is the point. Closing it is the work of the rows still blank, the wealth-versus-work tax commitments not yet specified into scoreable designs, and the honest possibility that some proposals phase in slower than advocates want. A platform that showed you a table that magically balanced would be lying somewhere inside it.

The big one: healthcare, without flinching

The largest number on this platform is universal, comprehensive, affordable healthcare, with Medicare for All as the clearest path. The honest accounting comes from the Congressional Budget Office itself: a single-payer system would raise federal spending by $1.5 to $3.0 trillion a year by 2030, and, under most designs CBO analyzed, total national health spending would stay roughly flat or fall, by as much as $0.7 trillion, while covering everyone. Both halves of that sentence are true at once, and so is the assumption underneath the favorable half: the lower-spending results come from designs that pay providers at Medicare-adjacent rates and capture large administrative savings, and those payment rates are themselves the central political fight, which is part of why this platform carries a bridge and not just a destination. Single payer is a historic shift of costs from premiums and deductibles onto the federal ledger, not a historic increase in what the country actually spends on care. Families would stop paying insurers and start paying taxes, and the binding design requirement is that for most working families the second number comes out smaller. The household tables below begin testing that requirement; they do not demonstrate it has been met. Version zero shows an illustrative 7 percent contribution clearing the bar for the example family and failing it for the example single worker, which establishes the need for a progressive rate structure or a wage exemption. Whether the final design satisfies the requirement remains to be demonstrated through official scoring and independently reviewed household tables, and the platform's commitment is that it does not pass without them. It is also why the platform accepts a Swiss-style universal system as a bridge: the destination is non-negotiable, the financing politics are hard, and pretending otherwise is exactly the kind of dishonesty this document exists to rule out.

The household tables, v0

A financing promise that never shows you a household is a slogan. These tables are version zero of the discipline: the left side is what representative households pay today, from named sources; the right side is the binding rule any financing design must satisfy before this platform supports it. One scenario is worked as an illustration and labeled that way, because the platform deliberately does not commit to a rate before scoring; the current-cost figures are the verified part.

What v0 is and is not: the current-cost column is pinned to the KFF 2024 Employer Health Benefits Survey and CMS national out-of-pocket data. The illustrative column is arithmetic on a placeholder rate, not a scored design, and it is published mostly because it exposes the single-worker problem a real design must solve. And the single-worker case gets a commitment, not just a diagnosis: the platform's default financing structure is a progressive schedule with an exempted first band of wages, and no financing design advances that leaves below-median wage-earning households paying more than they pay today in premiums plus out-of-pocket costs. Version one of these tables is independently reviewed household incidence for a fully specified financing design, and the healthcare brief commits to publishing that before passage, with the rule that if the tables cannot be made to satisfy the family constraint, the design changes until they do.

The bounded scenario, v0: three revenue cases

The gap above does not close itself, so here is the platform's first bounded answer, built the only honest way available: every component is a Congressional Budget Office Budget Option with its own published score, stacked into low, central, and high cases. Where a case scales a CBO increment (a 7-point corporate change from the 1-point option), the scaling is labeled, because independently scored options interact and the true combined revenue runs somewhat below the arithmetic sum. These are bounds for planning, not point estimates, and none of the indicative rates is yet a scored platform bill.

Read with the table above: the general-fund rows already scored there (IRS first tranche plus drug negotiation, about $583B) are counted in every case's residual; the pollution price and the Social Security cap fix are excluded here for the reasons stated there. Phase-in: Indicative sequencing: restored premium credits, school meals, and the child allowance in year one (their scores assume rapid start); childcare and paid leave phase over years two to four as systems build; rate changes effective year two; corporate phased two points per year. Unscored rows launch only when scored. If revenue misses: Priority order if revenue underperforms: the floors hold (child allowance, school meals, SNAP restoration, premium credits, the healthcare coverage floor); what slows first is the build-out layer (new-places pilots, service expansion, tutoring scale-up), explicitly and in public, rather than quiet across-the-board erosion. The debt path: Honest statement rather than invented curve: the country enters this window on CBO's baseline plus Public Law 119-21's roughly $4.1 trillion (with debt service). The platform's scored package under the Central case is designed to be near self-financing excluding interactions and interest; the Low case adds roughly $1.1 trillion over the decade unless spending phases slower; a year-by-year debt-to-GDP path will be published when the components are scored together rather than stacked, because publishing a fake curve would violate this page's first rule.

Part Two: How We Count

The scoreboard problem

In 1934, the economist Simon Kuznets delivered to Congress the first national income accounts, the ancestor of GDP, and attached a warning: the welfare of a nation can scarcely be inferred from a measurement of national income. We kept the number and discarded the warning. Ninety years later, GDP counts the coal burned, the cleanup afterward, and the hospital admissions in between, all as growth. It counts a ton of carbon dioxide at zero, though EPA's peer-reviewed 2023 estimate put the damage at about $190, a figure the current administration has formally withdrawn from government use, which changes the politics and not the physics; the peer-reviewed damage literature that produced it stands, the estimate is carried here as a central point in a contested range rather than a settled price, and if a future scorekeeper sets it lower, the revenue and the damage math shrink with it. It counts the $1.01 trillion of care that 59 million family caregivers provided in 2024 at exactly nothing, by AARP's updated valuation, which nearly doubled from its ~$600 billion prior edition because the counted caregiver population grew from 38 to 59 million and the hourly replacement value rose with care wages, a revision, and labeled as one. And it cannot tell the difference between a country getting richer and a country selling off its topsoil, its water, and its children's climate and booking the proceeds as income.

This is not an abstract complaint. In a peer-reviewed study in the American Economic Review, economists Muller, Mendelsohn, and Nordhaus found that several American industries cause more air-pollution damage than the total value they add to the economy; coal-fired power's damages ran up to 5.6 times its value added. On GDP's scoreboard, every one of those industries looks like a contributor. On an honest ledger, they are net extraction. The IMF's global estimate makes the same point at scale, and the honest label matters: $7 trillion a year in fossil fuel underpricing, of which only about 18 percent is explicit subsidy, checks governments actually write, while 82 percent is the implicit kind, unpriced damage and forgone taxes the public absorbs. Call the explicit part a subsidy and the implicit part what it is, an externality bill, and the argument survives the standard rebuttal intact: either way, the public pays it.

What we propose instead

Not the abolition of GDP, and not degrowth. The proposal is honest bookkeeping, and most of it is already underway. In January 2023 the United States adopted a national strategy to build natural capital accounts, official statistics that put the nation's land, water, air, and resources on the books alongside its factories, with Bureau of Economic Analysis pilot accounts running now and full implementation planned through 2036. This platform's asks are simple: fund it, codify it so no administration can quietly delete the accounts, and publish alongside GDP a short dashboard of what the headline number misses, net growth after pollution damage, the value of unpaid care, and whether natural assets are growing or being drawn down. That is the direction the Stiglitz-Sen-Fitoussi commission recommended after the 2008 crisis, and it is the same principle as the keep-the-count plank elsewhere on this platform: a country that stops measuring a problem has not solved it.

This is not a war on growth. It is a war on fake growth. Nearly everything this platform proposes to build, homes, clean power, transmission, care, cures, shows up as growth on any honest ledger. What changes is that the ledger stops rewarding the industries whose profits depend on damage the accounts refuse to see, and stops telling the country it is getting richer in years when it is merely billing the future.

The standard we hold ourselves to

Every figure in this document traces to a primary source listed below, the same standard as the Evidence page. When a number is contested, we use the official scorekeeper's version. When a proposal costs money, we name the financing family it draws on. When the honest answer is that something is a real tax, we say the word. And when the evidence changes, the ledger changes with it, and the correction is public.

Here is the short version. The money is there: nearly $700 billion a year in taxes owed but unpaid, $2.2 trillion in tax breaks never asked to justify themselves, $186 billion in payments the government itself calls improper, and pollution priced at zero while it costs us on the order of $190 a ton by the peer-reviewed estimate the government produced and then withdrew. Collect what is owed, end what fails the test, price the damage, tell the truth about the rest, and fix the scoreboard so growth means what people think it means: life actually getting better, on a ledger that counts everything, including the things money never touched.