Medicare's Two Trust Funds and the 45 Percent Warning Congress Keeps Ignoring

Washington has a talent for building a smoke detector, wiring it to the ceiling, and then arguing about whether smoke is really a problem. Medicare has such a detector. Congress installed it in 2003, wrote the trigger into statute, and has watched it go off for most of the years since. The detector is a number: 45 percent. Almost nobody outside the Trustees' staff knows it exists.

Let me explain what the number measures, why it matters to anyone who pays a payroll tax, and how you can read the evidence yourself without trusting a pundit, including this one.

Medicare is not one fund

Start with a definition, because most of the confusion in this debate is a confusion of terms. When politicians say "the Medicare trust fund," they are naming one thing and implying that it is everything. Medicare has two trust funds, and they are financed on different principles.

The Hospital Insurance trust fund, usually called HI, pays for Part A: inpatient hospital stays, skilled nursing care after a hospital stay, hospice, and some home health. It is financed mainly by the payroll tax. The rate is 2.9 percent of wages, split evenly, 1.45 percent from the worker and 1.45 percent from the employer. Self-employed people pay the whole 2.9 percent themselves. Since the Affordable Care Act, workers with wages above $200,000 (single filers) or $250,000 (joint filers) owe an additional 0.9 percent on the wages above those thresholds. HI also receives a share of the income tax that some higher-income Social Security recipients pay on their benefits, and it earns interest on Treasury securities it holds.

The Supplementary Medical Insurance trust fund, SMI, pays for Part B (physician visits, outpatient care, many drugs given in a clinic) and Part D (the prescription drug benefit that began in 2006). Here the financing is different in kind. SMI is not funded by a payroll tax at all. It is funded mainly by two things: the premiums that beneficiaries pay, and transfers of general revenue from the Treasury. General revenue means income taxes, corporate taxes and, when those fall short, borrowed money.

Notice what that means. The program people describe as "something you paid into" has a half that was designed from the start to be paid mostly by whoever happens to be funding the federal government that year. Part B was set up in 1965 so that premiums would cover roughly half the cost. Today the standard Part B premium is set to cover about a quarter of the cost for the typical beneficiary, which is the statutory design of the current law, and the Treasury covers the rest. You can check both statements in the annual Trustees Report.

There is also a bookkeeping feature unique to SMI. Because Congress guaranteed that general revenue would cover whatever premiums did not, the SMI fund can never run dry the way HI can. Each year the premium and the general revenue transfer are set to match expected spending. People who say "the SMI fund is solvent" are saying something true and very nearly meaningless. Of course it is solvent. A fund that can send a bill to the taxpayer is never short of cash. The real question is how big the bill is.

The arithmetic of who pays

Take a worker who earns $50,000 in wages. The Medicare payroll tax on those wages is 2.9 percent, so the total is 0.029 times 50,000, which is $1,450. The worker's share is $725 and the employer's share is $725. Take a worker who earns $100,000: the total is $2,900, with $1,450 from each side. Economists argue about who really bears the employer's half, and the usual view is that it comes out of the wages the worker would otherwise have been paid, but for this arithmetic the point is simpler. Those dollars flow to HI, and only to HI.

Now take a single filer earning $300,000. The regular 2.9 percent applies to the whole $300,000, which is $8,700. The extra 0.9 percent applies only to the amount above $200,000, which is $100,000, so it adds $900. The total comes to $9,600. That extra 0.9 percent is paid by the worker alone; there is no employer match on it. It goes to HI as well.

Not one cent of those payments goes to Part B or Part D. If you have ever heard someone say, "I paid into Medicare my whole working life, so I earned every dollar," the first thing worth asking is which Medicare. The payroll tax paid into the hospital fund. The doctor-visit and drug coverage was mostly paid by someone else's taxes, by the premium, and by the national credit card.

I'll flag where fact ends and value begins. That Medicare is financed this way is fact, and it is on the face of the Social Security Act and the Trustees Report. That this is a bad design is my judgment, and I will defend it below.

The warning of 2003

In 2003 Congress passed, and President George W. Bush signed, the Medicare Prescription Drug, Improvement, and Modernization Act, the law that created Part D and renamed Medicare+Choice as Medicare Advantage. Buried in that law was a provision, section 801, that almost nobody quotes: a funding warning.

Here is how it works. Every year the Medicare Trustees issue a report. The 2003 law told them to measure how much of total Medicare spending is being paid out of general revenue rather than from dedicated sources such as payroll taxes and premiums. If the Trustees find, in two consecutive reports, that general revenue is projected to exceed 45 percent of total Medicare spending within the first seven years of the projection, they must issue what the statute calls a Medicare funding warning. That triggers a chain of requirements. The President must submit to Congress, within 15 days of submitting his next budget, a proposal to respond to the warning. Congress then gets expedited procedures to consider that legislation.

In plain English, Congress built a fire alarm and attached a procedure that forces a vote. The first warning was triggered by the Trustees reports of 2006 and 2007, and by the Trustees' own accounting the warning has been issued again in a long run of years since. I will not give you a table of the years from memory. The place to check is the Trustees Report itself, which lists whether a warning has been triggered; the Centers for Medicare & Medicaid Services posts every report. What I can say with confidence is that the threshold was never meant to be a goal. It was a tripwire.

And what has Congress done each time the tripwire was crossed? It has not enacted a plan that brings general revenue below the threshold. The expedited procedures exist; the political will to use them to actually change the arithmetic does not. A warning that nobody must obey is an advisory. Congress wrote itself a very polite letter and then treated the letter the way the rest of us treat the terms and conditions of a software update.

How to read the Trustees Report

You do not need an economics degree. You need to know where to look, and I would like every reader to try it once, because the thing about primary documents is that they do not have a talk-radio contract.

The Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, to give the official name, consists of six people: the Secretary of the Treasury, who serves as managing trustee, the Secretary of Labor, the Secretary of Health and Human Services, the Commissioner of Social Security, and two public trustees appointed by the President and confirmed by the Senate. They report to Congress each year. The statutory due date is April 1, and in practice the reports often arrive later.

Four things to look for:

1. The HI depletion date. The report projects the year in which the HI trust fund will no longer be able to pay all scheduled benefits from its income and reserves. When that happens the law does not allow the fund to borrow; it can pay only what comes in. The report states what share of scheduled benefits could still be paid. Read the figures for the report year you are looking at, and notice that the date moves from year to year as the economy and the spending estimates change.
2. The general revenue percentage. This is the section that applies the 45 percent test. Look for the phrase "Medicare funding warning" or "excess general revenue Medicare funding."
3. The split between HI and SMI spending. Compare the dollar amount for Part A with the dollars for Parts B and D. The SMI side is now the larger part of Medicare outlays.
4. The sensitivity and alternative scenarios. Trustees now publish alternative projections, because the official current-law projection assumes things Congress itself may not do. For example, current law calls for certain payment updates to physicians and a slowdown in provider payment growth that the actuaries themselves have flagged as unlikely to happen as written. They say so. It is the most honest sentence in Washington.

The Congressional Budget Office publishes its own Medicare baseline, which is a useful cross-check; the two do not always agree. When two nonpartisan agencies produce different numbers, that is not a scandal. It is a reminder that projections are projections.

The best case for the other side

The best argument against my concern goes like this. Medicare is an insurance program for older Americans and for certain disabled Americans, funded by a mix of dedicated taxes, premiums and the general fund by design. Congress chose that mix. Choosing to fund medical coverage for seniors with general revenue is a policy choice a democratic government is entitled to make, and the 45 percent threshold is an arbitrary line that the 2003 Congress picked, not a law of nature. A program that serves a growing share of the population should expect to claim a growing share of the budget. The aging of the population is a demographic fact, not a failure of discipline. And Medicare has far lower administrative overhead than private insurance, so spending money on it is spending it well.

That is a fair statement of the view, and a few parts of it are right. The 45 percent line is a judgment, and the demographics are real. Where I part company is on four points.

One. The line was not picked by an enemy of the program. It was chosen in a bipartisan bill by the same Congress that expanded the benefit. They knew they were adding a large new obligation in Part D and wanted a trigger that would force a conversation if general revenue climbed. The warning is the program's own architects telling you where they thought the guardrail should be.

Two. "A democratic government may choose it" is not an answer to "can it afford it." Of course Congress may choose it. The question is who pays and when. The national debt clock shows what the choices add up to; the public debt was $40.56 trillion on November 5, 2026, up from $38.05 trillion a year earlier, per the Treasury's Debt to the Penny. That is a difference of $2.51 trillion in twelve months, and it is arithmetic: 40.56 minus 38.05 equals 2.51. Medicare is not the only reason, but it is one of the large ones, and a growing share of Medicare is the kind that runs on borrowing when the taxes fall short. For more on how the debt story is told and mistold, see Seven myths about the national debt that Washington finds convenient.

Three. Demography is the reason to act, not an excuse for not acting. If the number of beneficiaries is rising and the number of workers per beneficiary is falling, then financing the program from payroll taxes alone gets harder every year. That is exactly the situation in which a fiscal tripwire should be taken seriously. A smoke detector is not an argument that the house cannot burn.

Four. Low overhead is not the same as low cost. Medicare's administrative expense as a share of spending is small. That tells you the claims office is lean. It tells you nothing about whether the spending being administered is wise, or whether the price Medicare pays per service is the right one. A cheap checkout line does not make the groceries cheap. The same reasoning is at work in Six Things People Believe About Health Insurance That Are Simply Not So, where the gap between how a program feels and how it is financed is the whole story.

Why the design matters morally, not just financially

This is the Founding-era part of the argument. When the Constitution's authors gave Congress the power to tax and spend in Article I, Section 8, they attached to it a duty of accountability that rests on the voter being able to see the bill. James Madison's famous lines in Federalist 47 and 51 are about separation of powers, but the logic that runs through the whole document is that power should be visible and answerable. A payroll tax is visible on a pay stub. A general revenue transfer is invisible to everyone except an actuary. The more of a program that is paid from the invisible column, the less a voter can tell what it costs.

That is why I care about the trust fund distinction. A tax you can see is a tax you can argue about. The visible half of Medicare is the half that has a reserve and a depletion date and a rate you can find on your pay stub. The invisible half has no limit and no date, and nobody gets a pay stub for it.

I'll also say what this is not. It is not a case against caring for the elderly. Honoring the old is a moral inheritance older than the Republic, and the Judeo-Christian tradition makes it a command. The question has never been whether we care for our parents. The question is whether we do it in a way that tells the truth about what it costs, and whether we leave our children a program or a bill.

What would answer the warning

I said I would defend the view that the design is bad, so I owe you alternatives. Serious conservatives disagree about the right response, and I am not going to pretend there is a single plan. The options on the table, in outline:

• Raise the dedicated revenue. Raise the payroll tax or the income thresholds. This keeps the program as designed and makes the cost visible. The objection is that it taxes work and makes hiring more expensive.
• Change the benefit design. Combine the separate deductibles of Parts A and B, add an out-of-pocket cap, and give beneficiaries more reason to ask what a service costs. Traditional Medicare's Parts A and B have no cap on out-of-pocket costs; that is why so many people buy supplemental coverage. A modernized benefit could reduce both risk and waste.
• Means-test more of the benefit. The premium for higher-income beneficiaries already rises under what the law calls income-related adjustments. Expanding that idea shifts costs toward those most able to bear them.
• Move toward premium support. Instead of the government paying whatever the bills are, the government would pay a defined contribution toward a plan the beneficiary chooses, with plans competing. The case for it is in Medicare Advantage Proves Seniors Will Choose Private Plans When Allowed.
• Change the payment side. Price transparency and site-neutral payment, so that the same service costs the same in a hospital outpatient department and a doctor's office, attack the price rather than the patient. The related argument about the broader price problem is in Hospitals Went to Court to Keep Their Prices Secret, and They Lost.

Each of these has costs and each has critics, and you may reject some. But each one at least confronts the number. The one option that has been tried repeatedly is the one I consider worst: hold a hearing, express concern, and leave the warning in the drawer.

It is worth remembering how this program began and how its own founders sold it. The way a promise was made shapes the way it is kept, and I tell that story in What the Medicare Debate of 1965 Promised and What the Program Became.

The case in brief

1. Medicare has two trust funds. HI is financed mainly by a 2.9 percent payroll tax (1.45 percent each from worker and employer), plus 0.9 percent on high earners. SMI is financed by premiums and general revenue, and it cannot run dry because the Treasury covers the gap.
2. The 2003 Modernization Act wrote a funding warning into law: when general revenue is projected to pass 45 percent of Medicare spending in two consecutive Trustees reports, the President must propose a response and Congress gets expedited procedures to act.
3. The warning has been triggered repeatedly since the 2006 and 2007 reports, and Congress has not enacted a response that resolves it. A tripwire nobody obeys is only a suggestion.
4. The answer is not to abandon seniors but to tell the truth about the cost: make the bill visible, redesign the benefit, and fix prices, not just premiums.

What you can do: read the summary of the latest Medicare Trustees Report on the CMS website, find the general revenue percentage and the HI depletion date, and then ask your representative's office, using the representative lookup, what the member intends to do about the warning.