What the Social Security and Medicare trust funds hold, and depletion day

Two questions about Social Security and Medicare get asked more than any others, and they get answered worst. The first is "Where is my money?" The second is "What happens when the trust fund runs out?" The standard answers are a vault in West Virginia on one side and "the whole thing goes bankrupt and checks stop" on the other. Both are wrong. The truth is more interesting, and in some ways more sobering.

I will answer both questions with numbers, show the arithmetic, and say plainly which parts are facts and which are my judgment. The facts come from the annual Trustees Reports, which are public and which you can read yourself. The most recent report I can vouch for in detail is the 2026 report, released in June of that year, so verify the figures against whichever report is newest when you read this. Trustees' projections shift by a year or two from report to report, and they will keep shifting.

What a trust fund is

Let me define the terms.

Social Security is really two programs, run by the Social Security Administration and funded by separate trust funds. Old-Age and Survivors Insurance (OASI) pays retirement and survivor benefits. Disability Insurance (DI) pays benefits to disabled workers. The two are often combined and called OASDI.

Medicare Hospital Insurance (HI), also called Part A, pays for inpatient hospital care and related services. It has its own trust fund. The other parts of Medicare, which cover doctors and prescription drugs, are financed mainly through general revenue and premiums and are not what we are discussing here.

A trust fund, in this context, is an accounting device. It is a ledger at the Treasury Department that records two things: the dedicated tax revenue that came in, and the benefits and costs that went out. When income exceeds spending, the surplus is credited to the fund. When spending exceeds income, the fund is drawn down.

Now the part that surprises people. When a surplus comes in, the law requires the Treasury to invest it in special securities. These are not stocks, not gold, not a diversified portfolio. They are special-issue Treasury bonds, issued only to the trust funds, not tradeable on the market, and carrying interest set by a statutory formula. They are real obligations of the United States government, and the government does pay them. But they are IOUs from one part of the government to another.

The IOU problem, explained step by step

Follow the money with a simple example. This is a hypothetical, with round numbers, to show the mechanism.

1. In a year when payroll taxes exceed benefits by $100 billion, the Social Security trust fund has $100 billion more than it needs that year.
2. The law says the surplus must be lent to the Treasury. The Treasury gets $100 billion in cash and gives the trust fund $100 billion in special bonds.
3. The Treasury does not lock the cash away. It spends it, on defence, roads, interest, anything in the budget. Spending the cash reduces how much the government has to borrow from the public. In effect, the surplus financed other spending.
4. Years later, when benefits exceed payroll taxes, the trust fund redeems some of its bonds. The Treasury has to produce cash.
5. Where does the cash come from? The Treasury cannot pull it from a savings account, because there is none. It must come from the same two sources as any other federal spending: taxes collected that year, or new borrowing from the public.

That is what I mean when I say the special bonds are a claim on future taxpayers. They are a legal promise that the Treasury will pay, and a promise by the government to itself is only as good as the government's willingness to tax or borrow. Nobody is hiding a vault. The trust fund balance is not an invested nest egg; it is a record of how much the Treasury already borrowed from the program and spent.

I want to be fair about this. It is not a swindle, and the bonds are not worthless. If you buy a Treasury bond, you hold a claim on future taxpayers too; that is what all government debt is. The difference is that a private investor holds the bond as an asset, whereas for the government as a whole, the program's asset is the Treasury's liability. Counting it as an asset and as a liability nets to zero. The practical effect of the trust fund balance is that the law gives Social Security a legal right to be paid from general revenue after payroll taxes fall short, up to the amount of the balance. Whether Congress would repay in full is a political question, not an accounting one. You can see the size of the larger debt problem on the national debt clock, and in Seven myths about the national debt that Washington finds convenient.

Where the money comes from: the payroll tax

The funds are financed by payroll taxes under the Federal Insurance Contributions Act.

• Social Security (OASDI): 12.4 percent of covered wages, split evenly, 6.2 percent paid by the employee and 6.2 percent by the employer, up to an annual taxable maximum that adjusts with average wages. Self-employed workers pay both halves.
• Medicare Hospital Insurance: 2.9 percent of all covered wages, with no cap, split 1.45 percent from the employee and 1.45 percent from the employer. Higher earners owe an additional 0.9 percent above certain thresholds.

Together, a typical employee and employer pay 15.3 percent of wages, 12.4 plus 2.9. Economists generally agree that the employer's half is borne in practice by the worker in the form of lower wages, which is why the worker's cost is closer to the whole 15.3 percent than to half of it, though that is a matter of economic analysis rather than statute.

Compare the 1960 rate. In 1960 the combined Social Security rate was 6 percent, 3 percent each on employee and employer, and Medicare did not exist yet; it was enacted in 1965 and began in 1966. The rate that Social Security needs has more than doubled since 1960, and the reason is in the next section.

The arithmetic of workers per beneficiary

Social Security is mostly a pay-as-you-go system. Today's workers pay for today's beneficiaries. So the arithmetic that matters is how many workers there are for each person collecting benefits.

According to the Social Security Administration, in 1960 there were about 5.1 covered workers for each person receiving benefits. In the 1960s, that number was drifting down as the program matured and the population aged. By the 1970s it was closer to 3.2 workers per beneficiary. In recent years it has been about 2.7. The Trustees project that it will fall to roughly 2.3 by the middle of the 2030s.

Here is why that matters, with a hypothetical to show the arithmetic. Suppose the average benefit equals 40 percent of the average wage. That is a made-up round number for illustration, and not the actual replacement rate, which varies by income and is defined in several ways.

• With 5.1 workers per beneficiary, the payroll tax rate needed to pay the benefits is 40 percent divided by 5.1, which is about 7.8 percent of wages.
• With 2.7 workers per beneficiary, the rate needed is 40 divided by 2.7, which is about 14.8 percent of wages.
• With 2.3 workers per beneficiary, the rate needed is 40 divided by 2.3, which is about 17.4 percent of wages.

Same benefit. Different demographics. The burden per worker nearly doubles between 5.1 and 2.7 and rises by about another sixth after that. Nobody has to be wasteful for the math to work this way. It is the straightforward consequence of lower birth rates and longer lives, which are good things, because a system built on the idea that there would always be five workers behind every retiree was a system built on a demographic assumption that did not last.

Congress last made a major structural fix to Social Security in the Social Security Amendments of 1983, signed by President Reagan on April 20, 1983, after the report of the commission chaired by Alan Greenspan. The 1983 law raised payroll taxes, taxed a portion of benefits for higher-income retirees, and gradually raised the full retirement age. It was a bipartisan compromise that bought a generation of solvency. The solvency it bought is what we are now using up.

Depletion day

Now the second question: what happens when the trust fund runs out?

From the 2026 Trustees Reports:

• The OASI Trust Fund is projected to be depleted in the last quarter of 2032. At that point, incoming payroll taxes are projected to cover about 78 percent of scheduled benefits.
• The combined OASDI funds, if treated as one, are projected to be depleted in 2034, with about 83 percent of scheduled benefits payable.
• The Medicare Hospital Insurance (HI) trust fund is projected to be depleted in 2033, with about 89 percent of scheduled benefits payable from continuing income.

Please read those percentages carefully, because they are the answer to the second question. When a trust fund is depleted, benefits do not stop. The law says benefits can be paid only from the fund, which means that when the fund runs dry, the program can pay only what its ongoing tax income brings in. The result is an automatic cut to what the law has promised, of roughly 22 percent for retirement benefits in the case of OASI.

Let me show you what that cut looks like, with a hypothetical. A retiree is scheduled to receive $2,000 a month. If benefits are cut to 78 percent of the scheduled amount, the check is $2,000 times 0.78, which is $1,560. That is $440 less each month, or $5,280 a year. For a household that depends on that check, that is not a technicality.

I said depletion does not mean the end, and that is true. I also said the picture is sobering, and that is true as well. The cut is automatic. Nobody has to vote for it. It happens by default if Congress does nothing, and Congress has a long record of waiting until the last moment, and of using the 1983 playbook of taxes up and benefits down, in a divided arrangement where each party prefers that the other side propose the unpopular half.

I should also be careful about what the projections are. They are projections. They depend on assumptions about birth rates, immigration, wage growth, employment, life expectancy and the economy, and the Trustees themselves show a range of outcomes. The depletion date has moved by a year or two in either direction from one report to the next. I am confident of the structure: the fund is finite and the cut is automatic. I am less confident of the exact year, and you should not be either.

The steelman: it is only a financing gap, and a small one

The strongest defence of the status quo runs this way, and I want to state it as its defenders would.

The trust fund shortfall is real but manageable. It amounts to a projected gap that can be closed by modest adjustments: raise the cap on taxable earnings, raise the payroll rate by a few percentage points, or some combination. The program is not "going bankrupt" because it will still collect hundreds of billions a year. The bonds are backed by the full faith and credit of the United States, and the government has never defaulted on them. Talking about "IOUs" suggests that the money has been stolen. It was lent, with interest, to the Treasury, which is how every bond works.

Much of that is right. The bonds are valid. Benefits will not drop to zero. The gap is, in the grand scheme, a fraction of the federal budget and can be closed.

My answer is in three numbered points.

1. A claim on taxpayers is a tax in waiting. The existence of the bonds does not change who has to pay. It confirms that the payments will come from taxes or from more borrowing. That is a decision, and honest accounting says so.
2. The gap is closed only by choices people do not like. Every way of closing it costs somebody: higher taxes on workers, lower benefits for retirees, or both. Saying "it is only a gap" is not a plan. It avoids saying whose pocket it comes from.
3. Delay makes every option worse. The longer Congress waits, the larger the adjustment needed per year, because fewer years remain to spread it. Each year of waiting is a year in which the cost of a fix grows and the options narrow.

What a conservative should draw from this

I will mark a value judgment as a value judgment. I believe an honest government tells citizens the truth about what it has promised and what it can pay. That is not a matter of left or right, but it is a conservative instinct about fiscal responsibility and about respect for the people who plan their lives around the promise. A person who is 55 deserves to know what the law currently provides and when it changes, and a person who is 25 deserves to know that the bill is coming to him.

There are three reasons for urgency that have nothing to do with ideology. The first is that the numbers are published every year and every year they say the same thing. The second is that the people most affected by a late fix are those with the least ability to adapt. The third is that a trust fund that holds Treasury paper rather than assets makes the whole arrangement depend on the federal government's overall finances, which is why this belongs next to the debt, not apart from it. I covered that in the debt ceiling began in 1917 to make borrowing easier, not harder and in baseline budgeting, where a smaller increase gets reported as a cut, which explains how Washington's own way of counting hides the problem.

There is also the legal side. Americans are told they paid into an account. The Supreme Court held in 1960 that workers have no contractual right to benefits under the law, and that Congress can change the terms. I walked through that decision in Social Security is not your property, and Flemming v. Nestor said so. Put the two facts together, special bonds and no property right, and you see why the promise is a political one, not a legal one, and why it is more fragile than most people think.

You can look at the broader federal numbers on the numbers page, and at who bears the tax burden in Who actually pays the federal income tax, read straight from the IRS tables.

The case in brief

1. The Social Security and Medicare Hospital Insurance trust funds hold special-issue Treasury bonds, which are a legal claim on future taxes and not an invested nest egg.
2. The payroll tax is 12.4 percent for Social Security and 2.9 percent for Medicare Hospital Insurance, and the ratio of workers to beneficiaries has fallen from about 5.1 in 1960 to about 2.7, with roughly 2.3 projected.
3. In the 2026 Trustees Reports, the OASI fund was projected to be depleted in late 2032, with about 78 percent of benefits payable, and the Hospital Insurance fund in 2033, with about 89 percent payable.
4. Depletion does not stop benefits, it cuts them automatically to what current taxes can pay, and every delay makes the fix larger.
5. Honest government tells people which promises it can keep, and reform is easier and fairer the earlier it is done.

What you can do: read the summary of the newest Trustees Report on the Social Security Administration's website, find the depletion year and payable percentage, and write to your two senators and your representative asking what they propose.