Social Security is six years from insolvency. That’s not a projection buried in an actuarial footnote—it’s the opening finding of a new report from the Penn Wharton Budget Model (PWBM), released Thursday, which puts the program’s Old-Age and Survivors Insurance Trust Fund on track to run dry by 2032.
And the fix lawmakers will likely reach for first—raising taxes—may be precisely the wrong move.
That’s the stark, counterintuitive conclusion suggested by PWBM researchers Seul Ki “Sophie” Shin and Kent Smetters, who modeled five distinct reform packages ranging from all-tax to all-cuts and found the approach most conventional analysts dismiss as politically radioactive—deep benefit reductions—generates the strongest long-term economic growth.
The counterintuitive math
Run the numbers through a standard accounting lens and the tax-heavy plan, called Option A, looks like the winner. It delays insolvency from 2032 all the way to 2058 by raising the payroll tax rate one percentage point (to 13.4%), lifting the taxable earnings ceiling to $250,000 (up from $184,500 in 2026), and switching to a slower inflation index for cost-of-living adjustments.
Switch to dynamic economic modeling—the kind that tracks how people actually change their saving and working behavior in response to policy—and the picture flips. Option E, the most aggressive benefit-cut plan (no new taxes, deeper formula reductions, and a retirement age raised to 69), projects a 6.1% GDP boost and a 13.5% surge in private capital by 2060. Option A, the tax-heavy plan, delivers only a 2.4% GDP increase and a 4.4% rise in private capital over the same period.
The mechanism is straightforward: Tell Americans their Social Security checks will be smaller, and they’ll save more on their own. Smetters and Shin call this the “incentive to save.” More private savings means more capital available for productive investment, which drives up wages. By 2060, wages are projected to be 5.7% higher under Option E versus just 1.6% higher under Option A.
Smetters told Fortune his goal in this exercise isn’t to make recommendations, but to show a “range of options,” instead. If he had to guess, he added, most people would prefer Option C, somewhere in the middle, but he is leaving that to the political process. His job is to “show the tradeoffs across a wide range of options on a holistic basis without bias.”
For critics who argue the math in this analysis is cruel, though, he offered the perspective that the cruelest approach is likely the one on the books under current law, in which benefits would be cut immediately in just six years. This means a $2,500-$2,700 cut in benefits per year for a person retiring in seven years, versus PWBM’s Option E, the harshest scenario, which would cut benefits by $2,300 per year (for women) and $2,500 per year (for men).
Even that comparison hides a lot of pain headed for retirees under current law, Smetters said. Once the trust fund is depleted, current law would cut benefits for all retirees, even the proverbial 90-year-old grandmother. His Option E, on the other hand, would concentrate pain for newer retirees, in their 60s.
Why Washington gets this wrong
The disconnect, the researchers argue, comes down to a concept that rarely makes it into political debate: implicit debt. Under Social Security’s pay-as-you-go structure, today’s payroll taxes flow directly to today’s retirees—a transfer that carries the same economic drag as explicit Treasury borrowing but doesn’t show up on the federal balance sheet. PWBM estimates those implicit pay-as-you-go obligations are currently twice the size of the U.S.’s explicit national debt. If they were booked under standard accounting rules, America’s debt-to-GDP ratio would exceed 300%.
That’s why plans that look good on paper—Options A and B significantly reduce the official debt-to-GDP ratio—can underperform in the real economy. They cut the visible debt while leaving the hidden debt intact.
The generational trade-off nobody wants to talk about
None of this comes free. The gains from aggressive reform flow primarily to younger and future workers, while current retirees and near-retirees absorb the losses. Under Option A, a 60-year-old middle-income earner today loses $30,745 in lifetime value. Under Option E, that same person loses $60,970.
For someone born in 2051, those options produce lifetime gains of $42,025 and $81,932, respectively—in the same middle-income bracket.
But the PWBM report does offer one unexpected piece of good news on the fairness front: Achieving the best long-run outcomes for future generations doesn’t always require the worst short-term pain for current ones. Under Option C—a middle-ground package combining some tax adjustments with retirement age increases—most 60-year-olds today actually come out ahead on a lifetime basis, even while future generations gain more than they would under Option A.
Importantly, none of the five options would fully close Social Security’s long-term funding gap. They would reduce the bleeding, not stop it. And with the 2032 deadline now just one presidential term away, PWBM’s core message is methodological as much as political: Decisions made using conventional budget scoring will lead lawmakers to the wrong place. The math that drives political consensus isn’t the same math that determines economic outcomes.
