Four funds, four clocks, and the deadline that arrives first

Three dates are in circulation for the same event, and every one of them is correct.

The Social Security and Medicare Boards of Trustees released their 2026 reports on 9 June. Read the first page and the retirement fund empties in the fourth quarter of 2032. Read a paragraph further and Medicare's hospital fund empties in the second quarter of 2033. Read the table after that and the reserves run out in the third quarter of 2034.

These are not revisions of one another. They are four separate legal entities on four separate clocks, and the reason the internet cannot agree on when Social Security runs out of money is that the question, asked that way, has no single answer.

There is also a fourth date, which almost nobody has reported. It is 2031, it appears in a letter the Trustees sent to Congress on the same morning, and it is the first one that carries a statutory obligation.

When will the Social Security trust fund be depleted?

The Old-Age and Survivors Insurance Trust Fund, which pays retirement and survivor benefits, is projected to deplete its reserves in the fourth quarter of 2032. At that point continuing tax income covers 78 percent of scheduled benefits. The Disability Insurance Trust Fund is projected to pay full benefits through 2100. The 2034 figure describes a combined fund that does not exist in law.

That last sentence is the one doing the work, and the Trustees say it themselves. The summary sets out the combined projection and then adds, in parentheses, that "the two funds could not actually be combined unless there were a change in the law."

So the 2034 date, which is the one that reached the largest audience and which the Social Security Administration's own press release led with, is a projection of a hypothetical. It assumes Congress passes a law allowing money to move between two funds that currently cannot lend to each other. That assumption is reasonable, because Congress has authorized exactly such a transfer before. It is still an assumption, and it buys the retirement fund an extra twenty-one months by borrowing them from a fund that is in surplus.

Why do published dates for the same event differ by two years?

Because four trust funds are involved and the reporting rarely says which. OASI depletes in the fourth quarter of 2032. Hospital Insurance, which is Medicare Part A, depletes in the second quarter of 2033. The hypothetical combined OASDI fund depletes in the third quarter of 2034. Disability Insurance and Supplementary Medical Insurance do not deplete at all.

The confusion is not confined to secondary coverage. The Social Security Administration's own Trustees Report Summary is the highest-ranked government page for the phrase "social security trust fund depletion", and the sentence Google chooses to display from it reads: "Annual HI deficits are projected to return in 2027, requiring redemption of securities in the trust fund reserves until the trust fund's depletion in 2033." Every word is accurate. It is about Medicare.

What is the difference between OASI, DI and OASDI?

OASI and DI are separate trust funds established under separate provisions of the Social Security Act, holding separate reserves, funded by separate slices of the payroll tax. OASDI is not a fund. It is the arithmetic sum of the two, published because it is a convenient summary of the program, and it appears in the reports with a standing caveat attached.

The split is visible on every pay stub in the country, although it is not printed there. Of the 6.20 percent an employee pays, 5.30 points go to OASI and 0.90 points go to DI. The employer matches it. A self-employed worker pays the whole 12.40 percent, split 10.60 and 1.80 the same way. All of it applies only to earnings up to the taxable maximum, which is $184,500 in 2026.

Those two funds are in opposite conditions. In 2025 the OASI fund spent $1,448.8bn against income of $1,248.8bn and its reserves fell by $200.0bn. The DI fund took in $200.5bn against costs of $160.7bn and its reserves rose by $39.8bn. Over the 75-year projection period OASI carries an actuarial deficit of 4.55 percent of taxable payroll while DI carries a surplus of 0.13 percent. DI passes both of the tests the Trustees apply for financial adequacy. OASI fails both.

Combining them produces a number that is true of neither. It is the average of a fund in trouble and a fund that is not.

What happens to benefits when the reserves run out?

Payments continue, at a reduced level, without any further act of Congress. The funds cannot borrow, so once reserves reach zero the program can only pay out what is coming in. For OASI in 2032 that is 78 percent of scheduled benefits, falling to 62 percent by 2100. On the combined basis it is 83 percent, falling to 65 percent.

The mechanism is worth stating precisely, because "Social Security runs out" implies a cliff and the arithmetic describes a step. Payroll taxes do not stop. Roughly 91 percent of OASI income in 2025 came from payroll taxes, 5 percent from income tax on benefits and 5 percent from interest on the reserves. The reserves are the part that runs out. The tax is still there the next morning.

What no document specifies is how the reduction would be administered. Current law provides for benefits it does not provide the money for, and it does not say whether payments would be cut proportionally, delayed, or handled some other way. The Trustees model a uniform reduction because they have to model something. That is a modelling convention and not a description of what would happen.

What changed in the 2026 report, and why?

The long-range deficit widened from 3.82 percent of taxable payroll to 4.42 percent, an increase of 0.60 percentage points in a single year, and demographic assumptions account for 0.44 of it. The Trustees lowered the ultimate fertility rate from 1.90 children per woman to 1.75 and lowered assumed net immigration. Legislation contributed a further 0.16.

The legislative line is the One Big Beautiful Bill Act, enacted 4 July 2025. It made the 2017 rate schedule and the larger standard deduction permanent and added a temporary additional standard deduction for taxpayers over 65. Less income tax on benefits means less revenue routed back into the trust funds, and the reports book that as a permanent reduction in income.

The economic line moved the other way. Assumed labour productivity growth and real earnings growth over the next decade were both raised, which improved the balance by 0.10 percentage points. It was not close to enough to offset the birth rate.

This is where the projection deserves scepticism, and the honest place to put it is here rather than at the end. A fertility assumption is a forecast about behaviour seventy-five years out, and it just moved by 0.15 children per woman on one year's evidence. That single revision did more damage to the reported deficit than every economic and legislative change combined. It could move back. The reports carry low-cost and high-cost alternatives precisely because the intermediate set is a best estimate rather than a measurement, and the assumptions in this year's report were locked in February 2026.

What does not depend on the fertility assumption is the near-term reserve arithmetic. The OASI fund is spending more than it collects now, by a margin that is already recorded rather than projected, and the depletion date in the early 2030s survives a wide range of demographic outcomes.

Which document sets the earliest deadline?

Not the Trustees Report. On the same day, the Board sent Congress a separate notice under section 709 of the Social Security Act, triggered when a fund's reserves are projected to fall below 20 percent of a year's cost. The OASI balance ratio crosses that line during 2031, reaching 17 percent by the beginning of 2032.

The letter, addressed to the President of the Senate, exists because the statute requires the Trustees to tell Congress "specifically the extent to which benefits would have to be reduced, taxes would have to be increased, or a combination thereof" once that threshold comes into view. An identical letter went to the Speaker of the House.

It is the least covered document of the release and the most concrete. It carries a table of what it would take, year by year, simply to keep the fund above the 20 percent line through the beginning of 2036.

Two things stand out. The first is the shape: $55.9bn in 2031 and then $447.3bn the following year, an eightfold step in twelve months, because the first year only has to prop up the ratio and every year after has to fund a fund that has stopped funding itself. The second is that the letter says more will be needed after 2035, in increasing amounts.

A 2031 date is not a prediction that anything happens in 2031. It is a statutory tripwire, and the Board describes issuing such a report as reasonable advance notice.

What would it cost to close the gap?

Acting in January 2026 would require raising the payroll tax from 12.40 percent to 16.65 percent, or reducing scheduled benefits by 25.2 percent for everyone, or reducing them by 30.3 percent for people who become eligible in 2026 and later while leaving current recipients untouched. Waiting until reserves are gone in 2034 raises the tax option to 17.30 percent and the benefit option to 28.5 percent.

Those are the Trustees' own illustrations, and they are useful precisely because they are blunt. Nobody is proposing a 25 percent across-the-board cut. The point of the figures is to size the hole: $29.3 trillion in present value over 2026 to 2100, up from $25.1 trillion in last year's report, or about 1.5 percent of gross domestic product over the period.

The cost of delay is legible in the difference between the two sets. Eight years of waiting moves the required tax increase from 4.25 percentage points to 4.90, which does not sound like much until it is read as what it is: the same bill, paid by fewer people, over fewer years. Every year of delay narrows the population that shares it.

Is Medicare on the same clock?

No, and its numbers are better than Social Security's on every measure except the one that matters most to a hospital. Medicare's Hospital Insurance fund depletes in the second quarter of 2033 with 89 percent of costs payable, and that share rises to 93 percent by 2100 rather than falling. Its 75-year deficit is 0.56 percent of payroll against Social Security's 4.42.

The Medicare Trustees put the fix at raising the 2.90 percent hospital payroll tax to 3.46 percent, or a 12.0 percent reduction in benefits. Both are small next to the Social Security equivalents. The HI fund is nonetheless in weaker short-term shape than OASI: its trust fund ratio is 53 percent at the start of 2026, against 153 percent for OASI, and it has failed the short-range adequacy test every year since 2003.

Medicare Parts B and D cannot deplete, because the law resets premiums and general revenue contributions each year to cover the next year's costs. That guarantee is why they are solvent and also why they are expensive. Government contributions now finance about 75 percent of Part B and Part D costs, and the Trustees have determined for the tenth consecutive report that general revenue funding will exceed the 45 percent statutory threshold, which for the ninth consecutive year produces a formal Medicare funding warning.

Frequently asked questions

When will Social Security run out of money? It will not run out of money. The OASI trust fund reserves are projected to be depleted in the fourth quarter of 2032, after which payroll tax revenue still covers 78 percent of scheduled benefits. The widely quoted 2034 date applies to a combined OASI and DI fund that would require new legislation to exist.

Is the Social Security depletion date 2032, 2033 or 2034? All three, for different funds. 2032 is the retirement fund. 2033 is Medicare's hospital fund. 2034 is the hypothetical combined Social Security fund. The disability fund is projected to remain solvent through 2100.

Did the depletion date move up by a year? By a quarter. The 2025 report put OASI depletion in the first quarter of 2033 and the 2026 report puts it in the fourth quarter of 2032. The three-month shift crosses a calendar year boundary, which is why the change is sometimes described as a year.

What percentage of benefits would still be paid? 78 percent of scheduled OASI benefits at depletion in 2032, declining to 62 percent by 2100. On the combined basis, 83 percent at depletion in 2034 and 65 percent by 2100. For Medicare Part A, 89 percent in 2033, rising to 93 percent by 2100.

Why did the shortfall get worse this year? Mostly the birth rate. The Trustees lowered the ultimate fertility assumption from 1.90 to 1.75 children per woman and lowered assumed immigration, which together account for 0.44 of the 0.60 percentage point deterioration. The One Big Beautiful Bill Act accounts for a further 0.16 by reducing income tax collected on benefits.

How much would fixing it cost? Starting in 2026, a payroll tax rise from 12.40 percent to 16.65 percent, or a 25.2 percent cut to all benefits. Starting in 2034, 17.30 percent or 28.5 percent. The 75-year unfunded obligation is $29.3 trillion in present value.

What is the section 709 letter? A statutory notice the Trustees must send Congress when a trust fund's reserves are projected to fall below 20 percent of annual cost. The 2026 letter, sent 9 June, says the OASI ratio crosses that line during 2031 and puts the cost of holding it above the line at $1,935.9bn over 2031 to 2035.

The cost-of-living adjustment that determines the size of each payment is set by a different mechanism entirely, and we have set out how that calculation works.