This is a beta. We are publishing it early, unfinished, and open to being wrong, because we would rather be corrected now than after it is printed. Household dollar figures are illustrative until our calculator is live; cost estimates are ranges, and we say so. If you think a number here is wrong, or a piece of the design would fail in practice, tell us: that is what this draft is for.

America does not have a health insurance problem. It has a health care shortage sold at monopoly prices.

Every plan of the last 30 years (Clinton's, the Affordable Care Act, Medicare for All) is about who writes the check. None of them builds a single hospital, trains a single doctor, or opens a single clinic. That is why the ACA insured 20 million people and rural hospitals kept closing anyway.

A card is not care.

The fact that should open every argument

The American government already spends $2.56 trillion a year on health care: $7,526 per person.

That is more than the entire health system, public and private combined, of 27 of the 37 advanced nations. More than Japan spends in total, for the best health outcomes on Earth. More than Spain, Italy, South Korea, Israel, or Portugal. More than the United Kingdom spends in total.

American taxpayers already pay for a universal health system. They just don't get one.

Then they pay for it a second time, roughly $2.12 trillion more in premiums, deductibles, copays, and cash, and get the worst outcomes in the rich world.

So whatever this plan is, it is not "government taking over health care." The government already funds the majority of it. What it has never done is get anything for the money.

What America buys for the highest price in the world

	United States	Rank among 35 advanced nations
Spending per person	$13,473	35th of 35 — the most expensive
Life expectancy	78.9 years	30th of 35
Healthy life expectancy	63.9 years	34th of 35
Years lived in poor health	12.5	34th of 35
Infant mortality	5.5 per 1,000	33rd of 35
Maternal mortality	17 per 100,000	34th of 35
Hospital beds per 1,000	2.7	vs 12.6 in Japan and South Korea
Hospitals per million people	17.9	vs 65.8 Japan, 81.9 South Korea, 43.1 France
Health score per $1,000 spent	—	64th of 64 — last of every country that achieves good health

Americans do not merely die younger. They spend more of a shorter life sick. That double penalty is the finding, and reversing it is the job.

Nobody quotes the two capacity rows. Japan and South Korea run the best-performing health systems in the world at roughly 40% of American cost. They have four and a half times as many hospital beds per person as the United States, and three to four times as many hospitals.

America is not an over-medicalized country drowning in excess capacity. It is one of the thinnest-provisioned health systems in the developed world, and it charges the most in the world for access to the little it has.

That is why this plan leads with building.

The plan on one page

Four things, in this order:

1. Build the capacity. Ten years of federal capital, a new Hill-Burton, to build and reopen hospitals, clinics, and long-term care where people actually live, including the places the market abandoned.

2. Open the professions. End the artificial cap on doctors with more residencies, free medical school for service, licenses for qualified foreign-trained physicians, and bigger schools, and clear the nursing bottleneck of faculty and clinical placements. The shortage was manufactured: Medicare has held residency funding at 1996 levels since 1997.

3. One payer, one price. A single public payer with an all-payer fee schedule negotiated in the open, hospitals paid by global budget rather than by claim, drugs bought at a public price. That is where the savings are: America overspends on prices and billing, not on overuse.

4. Ban the dividend, not the owner. Providers may stay private, as they do in Japan and South Korea, the two best systems on Earth. But an institutional provider that bills the public payer pays no shareholder dividends, takes public patients, and charges the public price. Break up the roll-ups. Build public pharmaceutical manufacturing.

Everyone is covered, from birth, automatically, with no premium, no deductible, no network, and no bill. You choose any doctor and any hospital in the country. What the public pays for is decided in the open and published. What it doesn't pay for, you may always buy: nobody is ever forbidden care.

And the insurance industry is wound down with its workers landed rather than dropped: full wage replacement, paid retraining, and first claim on the biggest health hiring drive the country has seen. Software is taking those jobs either way; the plan decides whether the people in them land or get laid off.

Every part of the plan

What follows is the whole plan in brief: every element, so that nothing is a surprise. The full plan argues each of these properly.

Build the capacity

A ten-year federal capital program: a Hill-Burton for the 21st century. Hospitals, clinics, maternity units, psychiatric beds, dialysis, dental, and long-term care, placed where people actually live rather than where margins are best. Capital gets its own budget, decided publicly, so hospitals stop expanding by chasing profitable service lines. Reopenings and renovations can happen in years two and three; genuinely new construction takes four to seven.

Open the professions

The shortage was manufactured: Medicare's funding of residency slots has been frozen at 1996 levels since the Balanced Budget Act of 1997. The United States has roughly 2.6 practicing physicians per 1,000 people against an OECD average near 3.7, a gap of something like 370,000 doctors.

So: uncap residencies, make medical school free in exchange for service, license qualified foreign-trained physicians, and expand the schools. Break the nursing bottleneck, which is not a shortage of applicants but of faculty and clinical placements. Pay the preceptor, because clinical placements are the binding constraint. And the fast half of supply moves in months, not years: telehealth at payment parity with licensure preemption, full practice authority, pharmacist prescribing for minor conditions, and re-entry pathways for retired clinicians.

One payer, one price

A single public payer with an all-payer fee schedule negotiated in the open. Hospitals paid by global budget rather than by claim, which is what ends the volume treadmill and what makes a rural hospital viable. Physicians on a negotiated schedule. Prices set in daylight, published, and argued about in public rather than by a private committee.

This is where the money comes from. America's excess spending is prices and billing, not overuse: Americans see doctors less often than people in peer countries, not more.

Drugs: pay for the invention, then free the medicine

Buy at a public price, negotiate like Germany's AMNOG, and use § 1498 (the government's existing right to use any patent and pay compensation) as the backstop. Add a prize fund: pay a bounty for genuinely new medicines and put the patent in the public domain, free to the whole world. And build a public pharmaceutical manufacturer for essential generics, sterile injectables, insulin, and vaccines, because America has none and the periodic shortages are a market failure with a known fix.

One sequencing rule: separate the pricing reform from the plumbing. The pharmacy counter is where this fails first. A 180-day universal transition fill is the cheapest insurance policy in the plan.

Ban the dividend, not the owner

Providers may stay private. That is what Japan and South Korea do, and they run the two best systems on Earth. The OECD data is clear that private ownership is not what America gets wrong.

But as a condition of billing the public payer, institutional providers may not distribute profits to shareholders, refuse public patients, or set their own prices. Break up the roll-ups, freeze acquisitions during the transition, and let the public share of capacity grow by building rather than by confiscating.

What's covered — and what you can always buy

Everything medically established is covered. What the public purse pays for is decided in the open by an independent body and published, not left to a Secretary's discretion, because a vacuum gets filled badly and invisibly.

Anything the schedule excludes, you may always buy, with cash or with private supplemental insurance. Nobody is ever forbidden care. The one rule that keeps this safe: supplementary, never substitutive. Everyone is in the public system and pays for it regardless of what else they buy.

And one deliberate transitional exception: the ban on paying to skip the queue does not start at enactment. It starts, service line by service line, when the public waiting-time guarantee is actually being met, because the people who would use a fast lane are using one today, and forcing them into the public queue in year one would add demand in the worst possible year.

The insurance industry, and the people who work in it

They are not insurance companies any more. UnitedHealth alone employs or contracts roughly one in 10 American doctors. So the plan splits them four ways: underwriting dissolves, the pharmacy benefit managers are abolished, the doctors and clinics they own are converted and kept running rather than closed, and the ancillary businesses evaporate.

Between 1.5 and 2.5 million people lose their jobs. They get a funded guarantee: full wage replacement starting before the job ends, wage insurance, paid retraining, and first claim on the largest health workforce expansion in American history, which this same bill pays for. Those jobs are going to software anyway. The only question is whether the workers get a landing or a layoff notice.

The systems we already run

The VA is not abolished. It is the model, and it is expanded. The Indian Health Service is a treaty obligation funded at under a third of need. The territories (3.5 million citizens) face a funding cliff. And the largest mental health facilities in America are county jails, because of an exclusion written into the Social Security Act. Every national health plan of the last 30 years has left these out. This one doesn't.

The gaps every plan forgets

Long-term care: the biggest hole in American life, and the thing that actually bankrupts middle-class families. Covered in both settings, home and institutional, not split between them. Mental health and addiction, which is a capacity problem rather than a coverage problem. Dental, vision, and hearing. Public health, which is not the same thing as medical care. Medical debt, canceled, and refunded to the people who already paid theirs off. And malpractice: replace it with public indemnity for clinicians plus no-fault injury compensation for patients, on the New Zealand model. Injured patients get compensated faster and more often, clinicians stop practicing defensively, and it is one of the cheapest ways to move organized medicine from opposition to neutrality.

The record, the app, and the end of paperwork

One national health record you control, with a public access log showing every clinician who opened it: Estonia's design, and the reason a national record is politically survivable there.

And the territory nobody is standing on: in American health care the delay, the paperwork, and the opacity are not friction. They are the product. Every one of them moves money or risk from a company to a patient. So: publish coverage criteria in a form your doctor's software can read before anything is submitted, so most prior-authorization requests never happen. Answer the rest inside the federal deadline or they are approved automatically. Auto-refill chronic prescriptions. Pre-authorize permanent conditions permanently. One national search that tells you where the short wait actually is. And if you are past the waiting-time guarantee, the system finds you a slot elsewhere, books it, and pays your travel.

What we are borrowing, and from whom

Nothing in this plan is untested. Germany's AMNOG drug pricing and volume corridors. Taiwan's single-payer build, the one major system designed rather than inherited. Japan's fee schedule and its care-manager workforce, built three years before its long-term care benefit went live. Costa Rica's EBAIS teams and Brazil's Family Health Strategy for primary care. Sweden's quality registries. Dutch GP cooperatives for after-hours care. New Zealand's no-fault injury compensation. The strongest evidence for every contested piece of this plan is that some other country already runs it.

The money arrives first. The care arrives second.

This is the part most plans get wrong. It is a design decision, not an apology.

Building hospitals takes years. Training a doctor takes a decade. But almost everything that makes an insured middle-class American hate this system needs nothing built. Not being bankrupted. Not fighting a denial. Not getting four envelopes from four companies for one hospital stay. Not choosing between a prescription and a car payment.

So the plan runs three clocks, not one:

	What it delivers	How fast	Limited by
Relief	rules and money applied to the coverage you already have	months	drafting
Enrollment	people move onto the public payer, by cohort	~4 years	administration
Building	hospitals, clinics, doctors, nurses, dentists	~10 years	physics

The decisive move is that the relief clock does not run through the new payer at all. Nothing in the first year requires anyone to change plans, change doctors, hold a new card, or be enrolled in anything. It is a set of rules imposed on the insurance Americans already hold, plus a federal check that pays first.

Which gives the rule that governs the whole transition:

Nobody loses anything before they gain something.

What actually happens, and when

On the day it is signed:

• Medical debt is canceled. Collection suits, wage garnishment, and liens are prohibited. About 41% of American adults currently carry medical or dental debt. And because canceling debt would otherwise punish the people who scraped and paid theirs off, anyone who made payments on medical debt in the previous seven years is automatically refunded what they paid, up to $10,000, wherever the records already exist.
• Your insurance premium may never rise again. From that day forward no plan may price above the previous year, and each year it must fall by the value of what the federal government has taken off its books. This is backed by federal reinsurance, the same instrument that cut Alaska's individual-market premiums by 38.5% and Minnesota's by about 20%.

Within three months, all of it by rule, with nothing built:

• Prior authorization stops working the way it does now. Any request nobody answers inside the existing federal deadline (72 hours urgent, 7 days standard) is approved automatically. Denials require a named same-specialty physician's signature and a published reason. The clock is not new: those are the deadlines federal rules have imposed on Medicare Advantage, Medicaid, and exchange plans since January 2026. They simply do not apply to employer coverage. The question is only: why shouldn't your plan follow the rule Medicare Advantage already follows?
• Appeals are decided fast, by an independent clinician rather than by the insurer. And where your own doctor certifies that even a three-day wait is dangerous, the care happens immediately, at the insurer's expense, while the appeal runs.
• A permanent condition is permanently authorized. No more re-authorizing a drug every January for a disease that is not going to go away. No more being made to fail on a cheaper drug first when your own doctor has said not to.
• Surprise billing ends completely, including the ambulance. And that is not mainly a fairness measure. People price the ambulance in the driveway: at two in the morning, with crushing chest pain, deciding whether the ride is worth a thousand dollars. That calculation is made constantly, it is recorded nowhere, and it is paid for in dead heart muscle.
• A monthly allowance for family caregivers, paid to the person already doing the care, because that is a check and not a workforce.

And the largest single reduction in paperwork is not a faster answer: it is no question at all. Insurers publish their coverage rules in a form your doctor's software can read, so the criteria get checked at the moment of ordering. You find out then, not three days later, and most prior-authorization requests simply stop happening. This is a small ask: federal rules already require insurers to build the interface by January 2027 and already recommend the standard that makes it work in the exam room. We would only be requiring what is already recommended, and extending it to every plan.

By month six: chronic-disease drugs (insulin, inhalers, statins, blood-pressure medication, metformin, antidepressants, thyroid medication, blood thinners, naloxone, EpiPens) cost nothing at the counter, for everyone.

This one is not speculative. A federal program that pays pharmacies directly, at the counter, for people who are enrolled in nothing, already exists and is running today: the Medicare Limited Income NET program has done exactly this since 2010, and the Medicare GLP-1 Bridge launched in July 2026 using the same machinery. The rail does not need to be invented. It needs to be widened, and the copay set to zero.

By month nine: a single national search for the shortest wait. Every provider publishes real-time appointment availability. This creates no new capacity: it reveals the capacity that already exists and that no patient has ever had a way to find.

By month 12: primary care is free. Out-of-pocket costs are capped at $1,000 per person, $2,000 per family, for anything your plan covers. Glasses and hearing aids arrive for children and for adults over 65. The home-care waiting lists start clearing.

By month 18: the bill and the explanation-of-benefits are abolished as documents. One plain-language statement, from one place. Nobody else may send you anything.

And then the slow half. Staffing before buildings. Reopenings and renovations in years two and three. New construction in years four to 10. The first physicians from uncapped residencies reach independent practice around year eight.

We say the slow part out loud, first, unprompted, because a plan that promises transformation in 18 months is judged a failure in month 19. A plan that promises money in 12 months and care in six years, and then delivers the money in 12 months, is judged a success in both.

What it costs

This is not a spending plan. It is a plan to spend less. The comparison is not "cost of plan versus zero." It is $6 trillion a year versus the same care, and much more of it, for less.

We are not going to claim it saves a trillion dollars a year. That claim is common and it does not survive scrutiny: it presents gross savings as net, against the new spending on the other side of the ledger. Our claim is weaker in dollars and much easier to defend:

At maturity this plan is roughly spending-neutral. That is still an extraordinary claim, because of what the same money buys. Everyone covered, from birth. Nobody bankrupted. Long-term care included. Dental, vision, hearing, and mental health included. A hospital or clinic within reach in every county. For about what America already spends.

Run it on the share of GDP, not on dollars saved, because that is how every country that controlled health costs did it: by holding health spending growth below GDP growth for a sustained period, not by cutting spending.

	Health spending	Share of GDP
Today (2026)	$6.0 trillion	19.0%
Year 10 on the current trend, as CMS projects it	~$9.9 trillion	~21.1%
Year 10 under this plan, health growing ~2%/yr against GDP at ~4%	~$7.4 trillion	~15.6%

That is about $2.6 trillion a year less than the country is currently on course to spend, without ever cutting a nominal dollar. It does not reach Germany's level in 10 years — at this rate that takes about 25. It does stop the thing that is actually happening, which is health care taking a larger share of the economy every year with nothing to show for it.

Most plans leave this out: the savings are back-loaded and the spending is front-loaded. The buildout, the training pipelines, the worker transition, and the new benefits all start in years one to three; price convergence and administrative savings arrive in years five to 10. Years one through roughly seven are cash-negative. The buildout is best understood as deficit-financed capital investment, the way every other piece of national infrastructure is. But that has to be argued deliberately rather than discovered by the Congressional Budget Office, and we would rather say it here than be caught with it later.

Where this is going

This document is a beta and there is work we have not finished. The year-by-year fiscal path needs a named financing bridge. The household calculator, which will let any family look up its own number and which we consider the actual deliverable, is not built yet. Several cost estimates are ranges built on public data rather than on a scorer's model, and we have marked them as such rather than dressing them up.

We would rather publish it in this state and be corrected. The plan has already been improved twice by people who set out to attack it and found real problems. That is the process working.

If you want to tell us where it is wrong, we would like to hear it.