Social Security is back at the center of the political debate. Senator Bernie Sanders wants to expand benefits by at least $2,400 a year and eliminate the payroll-tax cap. The promise is familiar: make the wealthy pay more, give retirees more, and secure the program for generations.
My colleague Jack has explained why raising taxes while expanding benefits fails to confront the program’s underlying financing problem. Veronique goes further: the trust fund is the wrong measure of success. A larger balance of government promises does not tell us whether we have made those promises easier for future taxpayers to honor.
Both arguments point toward a larger problem. Social Security changes what Americans do with their income. It takes money that could partly be saved and invested, uses it to finance current benefits, and promises contributors that future workers will support them in turn. It also reduces their incentive to accumulate retirement assets of their own.
The cost extends beyond the federal budget. By displacing saving without building a corresponding stock of productive assets (assets used to produce future goods and services), Social Security weakens the investment on which future incomes depend. Any discussion about expanding it should begin by confronting that cost.
A retirement promise is not retirement saving
Social Security is a pay-as-you-go system. Today’s workers largely finance today’s beneficiaries. Their contributions are not accumulated in personal accounts and invested over their working lives. Instead, contributors acquire an expectation of future benefits financed by the workers who come after them.
The scale is enormous. According to the 2026 Social Security Trustees Report, the program collected approximately $1.32 trillion in payroll taxes in 2025 and paid $1.60 trillion in benefits. About 185 million people paid into the system during the year, and roughly 70 million received benefits in December. Those totals cover retirement, survivor, and disability benefits.
The $1.32 trillion is not a measure of lost saving. Workers would spend some of that income themselves, and beneficiaries can save part of their payments. But it shows the scale of the income being directed through a system that finances current benefits rather than accumulating investments for contributors.
To understand the difference, consider someone setting aside money each month for retirement. Savings invested in businesses help finance the equipment, technology, and expansion that make future production possible. The saver acquires a claim on income generated by productive activity. Over decades, that investment can support both the saver’s retirement and higher output across the economy.
Now consider money collected from that same worker and paid immediately to a retiree. To the extent that the payment finances consumption, those resources are used today. The worker is promised future income, but the transaction has not created an asset capable of producing it. The promise must be met by collecting income from somebody else.
This difference also changes behavior. A worker expecting Social Security to cover part of retirement has less reason to save privately for that purpose. Payroll taxes reduce the income available to save, while anticipated benefits reduce the amount the worker believes needs to be saved. The program substitutes a claim on future taxpayers for some of the assets households would otherwise accumulate.
Martin Feldstein put this mechanism at the center of his 1974 study of Social Security and capital accumulation. His original estimate suggested that Social Security reduced personal saving by 30% to 50%. A more recent U.S. study points to substantial displacement. In a 2016 analysis, economist David Blau estimated that each dollar of expected Social Security wealth crowds out approximately 56 cents of private wealth accumulation. His model captures the central problem: workers accumulate fewer assets of their own when they expect the government to provide retirement income.
A benefit promise can make a household feel better prepared for retirement without making the country better equipped to provide for it.
The trust fund does not resolve the problem
What about the money Social Security accumulated when it collected more than it paid out?
As Veronique explains, the surpluses were exchanged for special Treasury securities. The Treasury used the cash as part of its financing, while Social Security recorded an asset. Those securities are obligations of the federal government, and redeeming them requires the Treasury to obtain cash through taxes, borrowing, or reductions in other spending.
This is different from a pension fund accumulating ownership claims on productive businesses. One part of the government holds a claim against another. The transaction does not, by itself, add a factory, a machine, or a new technology to the economy. If the social security surplus permits additional spending elsewhere, it displaces investment with government spending.
Crediting a trust-fund account alone does not establish that additional saving occurred. The relevant question is what happened to the resources. Were they used to increase productive capacity, or consumed while a financial claim remained on the books?
Leonidas Zelmanovitz and Thomas Lanzi make this distinction in their discussion of the destruction of savings. A financial asset can remain valuable to its holder even when the resources originally lent have been consumed. The borrower may repay from other income, but the existence of the claim does not mean the original resources financed productive capital.
That is the sense in which savings can be destroyed. Money has not vanished, and the recipient has received something of value. But resources that could have financed investment have instead financed consumption. A promise of repayment survives; the productive investment that might have helped finance it does not.
Applied to Social Security, the problem is the saving displaced by an unfunded benefit promise. Calling the promise retirement security does not make it accumulated capital.
The cost reaches beyond retirees
Why should this matter to someone whose principal concern is protecting retirement incomes?
Because retirement incomes ultimately depend on what the economy can produce. A larger stock of productive capital allows workers to produce more per hour and supports higher wages. Greater output makes it easier to support both the working population and those who have retired. Displacing investment leaves future workers with fewer productive resources than they would otherwise have inherited.
This creates a particularly damaging combination. Workers inherit the obligation to finance earlier generations’ benefits, while the system discourages some of the investment that would make that obligation easier to bear. The cost appears through lower incomes and a smaller tax base, as well as through the payroll deductions everyone can see.
The fact that retirees spend their benefits does not answer this objection. Of course that spending supports the businesses receiving it. But the money had an alternative use. Financing consumption today and financing productive capacity for tomorrow have different consequences. Pointing to the spending generated by benefit payments tells us nothing about the investment forgone.
Private retirement saving eventually finances consumption too. Its advantage is that, during the intervening working years, it can help build the productive assets that support that consumption. The distinction lies in what happens before retirement, not in whether retirees should consume afterward.
Expanding Social Security strengthens the mechanism that displaces those savings. More income is collected to finance current benefits, and larger promised benefits give households less reason to provide for themselves through accumulated assets. Raising enough taxes to balance the program’s accounts would not, by itself, remove this economic cost.
Sound social security reform should address the problem of displaced investment. Feldstein’s later work on rethinking social insurance emphasized incorporating investment-based personal accounts. The relevant principle is that retirement provision should help build the assets that finance future income. A transition would still have to honor most existing commitments; moving contributions into accounts cannot make today’s beneficiaries disappear. But that financing problem is a reason to confront the obligations already incurred, not to keep expanding them.
Sanders’s proposal focuses on how much more the program can collect and distribute. We should also ask what future workers will have available to produce the income being promised. A retirement system that draws down the incentive and capacity to invest makes its own promises harder to sustain.
Social Security’s visible achievement is the benefit check. Its less visible cost is the saving displaced, the investment forgone, and the future income never produced. Expanding the program expands that cost. We should be building the capital that makes retirement affordable, not making ever larger claims on the future workers expected to pay for it.