The 2033 Shortfall: Who Would Pay for Social Security Changes

Near the bottom of every Social Security statement sits a sentence most people skim past: without legislative change, the program will be able to pay only about three-quarters of scheduled benefits once its trust fund reserves run out. That footnote is the entire policy debate compressed into one line, and it is the reason the current round of proposed Social Security changes focuses on taxing things that have never been taxed for Social Security before — stocks, estates, and employee benefits among them.

The useful way to read that footnote is not as a prediction about whether your check arrives. It is a description of a mechanism. Social Security has a defined funding formula, a defined reserve, and a defined legal consequence when the reserve hits zero. Take the mechanism apart one step at a time and the proposals stop sounding like political noise and start looking like arithmetic with distributional choices attached.

Step one: the money is current, not saved

Social Security is financed almost entirely by a dedicated payroll tax on current workers, levied at a combined 12.4% of covered wages — 6.2% withheld from the employee and 6.2% paid by the employer, with the self-employed owing the full amount. That revenue flows out to beneficiaries almost immediately. There is no funded account with your name on it.

Two features of that tax base matter more than anything else in the reform debate. First, it applies only to wage and self-employment income; investment income, capital gains, rents, and inherited wealth are outside the base entirely. Second, it stops at the annual contribution and benefit base — $176,100 in 2025, indexed each year to average wage growth. Earnings above that line pay nothing further into Social Security and earn no additional benefit credit.

That ceiling is the hinge. Because the cap is indexed to average wages while top-end incomes have grown faster than average wages, a shrinking share of total compensation sits inside the taxable base over time. Income growth concentrated above the cap generates no Social Security revenue at all. The funding problem is therefore not only demographic; it is partly a definition problem about which dollars count.

Sizing the gap: one number does most of the work

The 2025 report of the Social Security Trustees projected that reserves in the Old-Age and Survivors Insurance trust fund would be depleted in 2033, after which continuing payroll tax revenue would cover roughly 77% of scheduled benefits. On a theoretical combined basis with the disability fund, the Trustees put depletion at 2034 with about 81% payable. The same report put the 75-year actuarial deficit at 3.82% of taxable payroll.

Interpret that last figure carefully, because it is the number every proposal is implicitly priced against. A deficit of 3.82% of taxable payroll means that closing the whole gap with the payroll tax alone would require roughly a 3.8-percentage-point increase, taking the combined rate from 12.4% to somewhere near 16%, split between worker and employer. That is the benchmark. Any proposal that does not raise something equivalent closes part of the gap, not all of it.

Note also what depletion does and does not mean. Reserve exhaustion is not insolvency in the corporate sense; the payroll tax keeps arriving. The legal problem is that the program cannot spend money it does not have, which is why the consequence takes the shape of an across-the-board shortfall — the distance between 100% of scheduled benefits and the 77% the Trustees project as payable — rather than a missed payment.

💰 The levers, and who they reach first

Reform proposals cluster into four mechanisms, and they differ less in ambition than in whose income they newly touch. A recent MarketWatch survey of the proposals on the table — a secondary summary, not an official score — groups the newest ideas around extending Social Security taxation to stock-based compensation, estates, and employee benefits currently excluded from the wage base.

The pattern is worth stating plainly: revenue levers bite now and benefit levers bite later. That asymmetry explains why the political conversation keeps drifting toward taxation. With depletion inside the 2030s, formula changes phased in over twenty years arrive too late to prevent the shortfall the footnote describes.

Stocks, estates, and benefits: the definitional frontier

The newer proposals attack the base rather than the rate, and their appeal is that they reach income the payroll tax has never seen. Equity compensation is the clearest example. Restricted stock that vests is already wages for income tax purposes, but for a high earner it typically vests above the Social Security cap, so it generates no Social Security revenue. Removing the cap captures it automatically; targeted base-broadening captures it directly.

Estates sit further out. Social Security has no claim on transferred wealth, and the federal estate and gift tax basic exclusion is $15 million per person for 2026 — a threshold high enough that very few estates are taxable at all. Routing estate tax revenue to the trust funds would be a structural break: the program would stop being purely contributory and start drawing on general wealth transfers. That is a defensible design, but it changes what Social Security is, not just how much it collects.

Precedent exists for taxing capital to fund social insurance. The 3.8% net investment income tax and the 0.9% additional Medicare tax, both applying above $200,000 for single filers and $250,000 for joint filers, already do exactly this for Medicare — and because those thresholds have never been indexed, they reach steadily further down the income distribution each year. Any new Social Security surcharge written with fixed thresholds would behave the same way over time.

⚠️ Where this explanation stops holding

Three caveats limit how far the mechanism carries. First, projections move. The depletion date and payable percentage shift with each annual Trustees Report as fertility, mortality, immigration, wage growth, and legislation are re-estimated, and recent legislation has pushed costs upward — the Social Security Fairness Act, signed in January 2025, repealed the Windfall Elimination Provision and Government Pension Offset, raising benefits for affected public-sector retirees.

Second, lifting the cap is not pure revenue. If newly taxed earnings also earn benefit credit, part of the money comes back out as higher future benefits, which is why the gross tax increase and the net solvency gain differ. Proposals that credit nothing above the cap raise more and sever the contributory link more sharply.

Third, the taxation of benefits already functions as a quiet, unlegislated change. The thresholds at which Social Security benefits become taxable income — $25,000 for single filers and $32,000 for joint filers — were set decades ago and have never been indexed. Every year of nominal income growth pulls more retirees across them. For a household planning retirement income, that drift is more certain than any bill in Congress.

Reducing dependence on a single check

The planning conclusion follows from the structure, not from a forecast. Social Security was designed to replace part of pre-retirement earnings, not all of it, which already implies the remainder must come from savings, pensions, or continued work. The reform debate adds a second point: both the size and the tax treatment of that partial replacement are subject to legislative revision on a timetable no household controls.

Model the 77% case, not the zero case

Catastrophe framing produces bad plans. Running a retirement projection at zero Social Security overstates required savings dramatically and can push households into excessive deferral. Running it at the Trustees’ own payable figure is the defensible stress test: if roughly three-quarters of scheduled benefits still clears your fixed costs, the exposure is manageable; if it does not, the gap is quantified and specific rather than vague.

Diversify the tax character of retirement income

Because benefit taxability hinges on thresholds that do not move, the mix of taxable, tax-deferred, and Roth-type income determines how much of a benefit a household actually keeps. Two retirees with identical gross income can face materially different taxation depending on which accounts their withdrawals come from. That lever is entirely within household control and does not depend on which reform, if any, passes.

Treat the claiming decision as the largest single variable

Delayed claiming raises the monthly benefit for life and is inflation-indexed, which makes it one of the few ways to increase lifetime income from a statutory source rather than a market return. It is also the decision most likely to be distorted by headlines, since fear of legislative change pushes people to claim early and permanently lock in a smaller base.

The open variable is timing, not arithmetic

The arithmetic is settled and public: a 3.82% of payroll gap, a 2033 depletion date, and roughly 77% of scheduled benefits payable thereafter. What is unknown is when Congress acts and which base it chooses, and history suggests the answer is late — the 1983 amendments, which raised the full retirement age and first subjected benefits to income tax, were enacted only shortly before the trust fund would have been exhausted.

What would change this reading? A Trustees Report that moves the depletion date materially in either direction, or an enacted bill that fixes a definition — whether equity compensation, investment income, or transferred wealth enters the base, and whether new thresholds are indexed. Until one of those lands, the only assumption that stays valid under every proposal is that your plan should not require any single income source to perform exactly as currently scheduled.