Yesterday, Jack critically analyzed Senator Bernie Sanders’s latest Social Security pitch. He left very little of it standing. Sanders wants to scrap the payroll-tax cap, raise benefits by at least $2,400 a year, and extend the program’s solvency “for generations.” Jack correctly replies that the tax, as it is before the increase in benefits proposed by Sanders, closes only part of the shortfall (the difference between the future benefits and the payroll-tax revenue), but that the proposed new benefits devour the additional revenue, while top marginal rates in California and New York climb past 60 percent. The cherry on top is that nothing is left over for Medicare, which has a gap that’s far larger than the one now plaguing Social Security.
I agree with all of Jack’s analysis. But I want to go further.
Look at how the argument is scored. Sanders promises decades of solvency. Jack answers that the plan buys far fewer years than advertised. Both are counting the same thing, which is the number of years until the Social Security trust fund reads zero.
The thing is that we have to stop measuring a plan’s validity based on the idea of a Trust Fund balance. It doesn’t mean at all what people think it means and, more importantly, as long as we use it, plans like Sanders’s will keep sounding better than they are.
What is in the OASI Trust Fund
From 1983 to 2009, Social Security collected about $3 trillion more in payroll taxes than it paid out in benefits. The money went into the Treasury’s general account and paid for everything else the government does, from buying aircraft carriers to doling out farm subsidies. In exchange, the Treasury handed the Social Security Administration a stack of special bonds, which were stored somewhere in Parkersburg, West Virginia.
Those bonds are a claim that one federal agency, the Social Security Administration, holds against another, the Department of the Treasury. In other words, what Social Security counts as an asset, the Treasury owes as a liability of exactly the same size. Add them up, and you get zero.
Private employers are not allowed to do such convenient accounting. Yet that’s the way Washington structures its largest retirement program.
The bill is now arriving. Since 2010, Social Security has been running a permanent cash flow deficit, meaning it has spent more than it collects in taxes every year. In 2025, the gap was about $200 billion, accounting for the interest the Treasury credited to the Trust Fund (i.e., future borrowing). Reserves fell from $2.5 trillion in 2024 to $2.3 trillion in 2025. Each bond that gets redeemed is paid for with borrowing.
The Trust Fund does have one real function, and it is a legal one. While the balance is positive, the program may pay full benefits. When it reaches zero, which the Trustees expect for the retirement fund in late 2032, the law limits benefit payments to incoming revenue. That means a cut of about 22 percent in benefit payments. So the date matters. It marks the end of a borrowing authority.
Doubling down
When Sanders says his plan makes Social Security solvent, he means the trust fund balance stays above zero. The Chief Actuary’s 2023 score of his Social Security Expansion Act, the one he still cites, shows that the Trust Fund will collect more than the program spends for fourteen years, from 2024 through 2037. Then it will cash its IOUs every year from 2038 through 2096.
This is literally what the government has done since 2010. I can see why Sanders would love that solution. It lets him and his big-government friends on the right and the left spend all that extra cash collected as soon as the cap is lifted, until the program starts running a cash-flow deficit. He is saddling future generations with both higher taxes and more borrowing.
Indeed, in present-value terms, workers would pay an additional $33.8 trillion in taxes, more than half from new levies on investment and business income that the slogan “scrap the cap” never mentions. They would be handed $12.8 trillion in new benefit promises. And from 2038 on, the U.S. Treasury would again begin borrowing to pay for all of that, at which point workers or their children would have to pay a second time through higher taxes, inflation, or both.
This is what the left’s expansion agenda amounts to. Its definition of success is a positive number in an account that holds only the government’s IOUs, paid on the backs of American workers and payable to itself through more borrowing.
Jack showed that the Sanders plan fails on its own terms. My amendment is that the terms are the problem.
The better approach
Forget about the Trust Fund. We should instead reduce the gap, each year, between what the program owes in benefits and what it takes in revenue. It should do it without raising taxes further on the American people. It also doesn’t have to.
Start immediately by capping the largest benefits, some of which exceed $60,000 a year for top earners. Tie the growth of initial benefits to prices instead of wages. Use a more accurate inflation index for cost-of-living adjustments. Then move over time toward a flat benefit that keeps every retiree out of poverty, as Cato Institute’s Romina Boccia and Ivane Nachkebia propose in Reimagining Social Security (a must-read book for anyone interested in our fiscal future), and let people save the rest in private accounts that they own.