The new Congressional Budget Office baseline is not a forecast. It is an arrival. Spending from the Old-Age and Survivors Insurance trust fund — the part of Social Security that actually pays America’s retirees — will exceed its income every year through the end of the decade, and by 2032 the fund will be empty. That is one year sooner than the previous baseline said. One year is not a rounding error: it is the difference between a problem the next president can defer and a problem the next president cannot.
The headline cut, on paper, is brutal: from 2033 onward, scheduled benefits would automatically drop by an average of 28%, with a smaller 7% reduction in the first year of depletion. For the average retired worker — currently receiving $2,071 a month, per the Social Security Administration’s January 2026 fact sheet — that is a fall to roughly $1,491. For a married couple living on the SSA’s typical joint benefit of $3,208, the CBO estimates an annual cut of about $18,400.
What the CBO actually said
The numbers come from CBO’s February 2026 baseline projections. The OASI fund alone empties in 2032; the smaller Disability Insurance fund is in better shape; if Congress chose to combine them — which would require a vote — exhaustion slides to 2033. The illustrative scenario assumes the law remains as currently written: when the trust fund hits zero, the program can pay only what it collects in payroll taxes that month, and benefits get reduced automatically to fit. This is not a policy choice. It is a statutory cliff.
The reason the date keeps moving forward is structural. CBO projects that OASI’s annual cash deficit will rise from about $207 billion this fiscal year to $525 billion in 2032; spending grows from $1.5 trillion to more than $2.5 trillion by 2036, while payroll-tax receipts grow more slowly. The denominator is shrinking — the U.S. fertility rate has been below replacement since 2007 — and the numerator is the cohort born between 1946 and 1964.
Why the inflation that helps you also accelerates the cliff
The 2.8% cost-of-living adjustment that landed in January, applied across roughly 70 million beneficiaries, raised total program outlays by about $80 billion a year. COLAs are good for retirees in the next year and bad for the system over the long run, because they compound inside a fund that no longer has the contributor base to keep pace. Add the Bureau of Labor Statistics’ reading that services inflation continues to run above headline, and the fund’s liabilities grow faster than the wages on which payroll taxes are levied.
Medicare quietly piles on. The Part B premium climbed to $202.90 a month in 2026 — the first time it has ever crossed $200 — and is deducted directly from every Social Security check. Even before any 2032 cliff, the net amount landing in retirees’ bank accounts is already eroding: the headline benefit is, in the technical sense, an «advertised price».
1984 is the most useful date in this debate
Anyone who wants to know what Congress will probably do should look at 1984.
That year, two months after the Greenspan Commission’s reforms went into effect, Congress made Social Security benefits taxable for the first time and set the income thresholds at $25,000 for a single filer and $32,000 for a married couple — designed to hit roughly the top 10% of beneficiaries. Then Congress wrote the law without indexing them to inflation, and walked away.
Forty-two years later, those thresholds still have not moved. They now capture roughly half of all Social Security recipients. A tax on the rich became, through inertia alone, a regressive levy on the median retiree. This is what can-kicking actually looks like in U.S. fiscal policy: not loud refusal, but a vote on a number, no escalator clause, and forty years of silence. The system behaves like a ruminant, chewing the same fiscal cud rather than swallowing the politically painful bite.
What Washington will probably do — and what it should
The technically obvious answers are well known: lift or remove the cap on payroll-taxable wages (currently $176,100 for 2026), gradually raise the full retirement age beyond 67, swap the COLA index from CPI-W to a chained CPI. Each carries a constituency that defeats it on its own. Combinations close the gap, but require a coalition.
The probable path is the path of 1984. Congress will pass something — likely two or three years before the cliff, when the political theater becomes unbearable — that closes the gap by a familiar mix: a partial wage-cap lift, a phased retirement-age tweak, and a quiet change to the COLA formula that the press will understate. It will not be sold as a benefit cut. It will function as one. Promised parity is preserved on paper, eroded in cash flow.
The colder lesson is the one for Americans planning retirement. The idea that Social Security alone is enough is already obsolete, even at today’s $2,071. By the early 2030s it will be untenable. That is why the private retirement architecture — Roth IRA, 401(k), HSA — has stopped being optional for any worker under fifty-five, and why the same demographic squeeze has already been quietly reshaping how Americans plan for retirement: longer working years, larger private buffers, a closer eye on what actually lands in the bank account each month.
If 2032 arrives without legislation, the country will not face a crisis. It will face the answer to a question Washington has been avoiding for forty years: whether the floor is for everyone, or only for the half the 1984 vote forgot to inflation-protect.
Sources: Congressional Budget Office, “The Budget and Economic Outlook: 2026 to 2036” and “Social Security Trust Funds Baseline” (February 2026); Social Security Administration, “Fact Sheet on the Old-Age, Survivors, and Disability Insurance Program” (January 2026); SSA Monthly Statistical Snapshot (March 2026); Bureau of Labor Statistics CPI-W data; Centers for Medicare and Medicaid Services, “2026 Medicare Parts A & B Premiums and Deductibles” Fact Sheet (November 2025); Congressional Research Service R47040, “Social Security: Trust Fund Status in the Early 1980s and Today and the 1980s Greenspan Commission”.