If we hike payroll taxes as much as they want to, billions in income-tax revenues are no longer needed. Social Security reform hasn’t seen this much activity since President Bush’s ill-fated efforts in the mid 2000s. But this time, the action is on the Democratic side, with near-unanimous support among House Democrats for the Social Security 2100 Act, which would make Social Security solvent by increasing payroll taxes and taxing high earners more heavily. But what most people don’t know is that the Social Security 2100 Act would actually make Social Security’s funding less progressive, because those payroll-tax increases would replace an even more progressive source of funding that the one the program now relies upon. This opens some opportunities for conservatives, who currently are rudderless on Social Security reform, to expand the field for coming to a deal.

Thanks to the aging of the population, Social Security faces a future with more beneficiaries and fewer workers to support them. Based on current trends, Social Security’s trustees project it will be unable to pay full promised benefits beginning in 2035. Social Security reform plans would address that looming insolvency by raising taxes, cutting benefits, or both. Social Security 2100 is on the far end of the progressive spectrum, in that it not merely fully funds promised benefits, but actually expands benefits across the board.

To pay for these benefits, the Social Security 2100 Act gradually raises the payroll-tax rate from 12.4 percent to 14.8 percent in annual increments of 0.1 points. At the same time, while Social Security taxes currently apply only to earnings up to $132,900 — and benefits are based solely on those capped earnings — Social Security 2100 would also tax earnings above $400,000, with an untaxed “doughnut hole” in between. Since the new $400,000 threshold is not indexed while the $132,900 current-law wage ceiling rises annually with wage growth, the gap would shrink over time, such that by 2048 all earnings would be taxed by Social Security.

On paper, this looks like the usual progressive playbook of pushing the tax burden upward, even for a program like Social Security that historically has sought not to appear like a “welfare program.”

But there’s another perspective, from which the Social Security 2100 Act would be a regressive change that would make Social Security’s funding less progressive. That perspective takes into account the massive amounts of income-tax revenues that will be funneled into Social Security in the years leading up to its insolvency.

From the mid 1980s through 2009, Social Security ran payroll-tax surpluses that were credited to the program’s trust fund. However, economic research concludes that these surpluses weren’t saved in a broad budgetary or economic sense. Instead, Social Security surpluses allowed the rest of the federal government to tax less or spend more than it otherwise would have, because borrowing from Social Security isn’t reflected in the unified budget deficit or the publicly held national debt. Nevertheless, those past Social Security surpluses were credited to the trust fund, and Social Security is now drawing those balances down. And since the trust fund doesn’t reflect an actual store of wealth, this money is coming from other taxes.

When the federal budget repays the Social Security trust fund using general tax revenues, it is effectively transferring income taxes to Social Security, because income taxes are by far the largest source of general tax revenues. Those transfers to Social Security are growing ever larger as Social Security’s own dedicated annual tax revenues increasingly fall short of benefit costs. By the year 2034, 19 percent of Social Security’s resources will come from general tax revenues. That $431 billion is equal to about 11 percent of the income-tax revenues expected to be collected in that year.

But in 2035, Social Security’s trust fund runs dry, and so will its legal right to make those claims on income-tax funds. In the absence of reform, Social Security would have no legal option but to cut benefits by roughly one-fifth — to the level payable through Social Security’s own dedicated taxes, which consist mostly of the 12.4 percent payroll tax plus a small amount derived from income taxes levied on Social Security benefits.

The Social Security 2100 Act prevents that outcome by levying new taxes that increase Social Security revenues in 2035 by $461 billion over current-law levels. In other words, payroll tax increases substitute for expiring income-tax transfers. Now, those new Social Security payroll taxes might be more progressive than the current Social Security payroll taxes. But they’re a lot less progressive than the income taxes currently being used to keep Social Security checks arriving in mailboxes.

According to IRS data, the top 10 percent of households pay nearly half of all income taxes, and the top 1 percent by themselves pay nearly one-fifth of income taxes. So the $431 billion in general revenues to be paid back to the Social Security trust fund in 2034 is heavily weighted toward the rich.

Versus that 2034 status quo, the Social Security 2100 Act will shift the Social Security tax burden downward. For instance, a Joint Committee on Taxation analysis found that by 2029, the Social Security 2100 Act would increase total federal taxes by 8.6 percent for taxpayers making from $20,000 to $30,000 and by 5 percent for taxpayers making $30,000 to $40,000. Taxes would rise by 7.4 percent for taxpayers making $500,000 to $1 million, while taxes would increase 9.4 percent for taxpayers making $1 million or more. By itself, this doesn’t seem massively progressive.

But there’s something even more important: These figures don’t account for the fact that if the Social Security trust funds are exhausted in 2035 and payroll taxes go up to replace the lost general-revenue funding, the hundreds of billions of dollars of income taxes currently flowing to Social Security will no longer be needed. Instead, over a one-year period, a potential 11 percent cut in income taxes comes onto the negotiating table. That potential tax cut could easily offset Social Security 2100’s payroll tax increases on high earners while giving little back to low earners who currently don’t pay much in income taxes. As a result, come 2035 the overall federal tax code could become less progressive.

Now, this gives conservatives — who are at present, at best, adrift on Social Security policy — some room to maneuver. We’re already putting a lot of revenues from high-income taxpayers into Social Security, they might argue. If progressives want to raise taxes on those same taxpayers to keep the system solvent thereafter, that’s fine. But they can’t do that and retain a half-trillion dollars per year of income taxes used to repay the Social Security trust fund once those revenues are no longer needed. With the trust fund repaid, there is no reason those income taxes should be redirected to new spending. Instead, income taxes could be cut which would keep higher Social Security tax rates from pushing U.S. top marginal rates to Scandinavian levels. Alternatively, the now-unneeded income-tax revenues could be used to reduce the budget deficit and debt.

Alternatively, if Republicans are truly frightened of Social Security reform and want as much of a do-nothing policy as possible, they might simply propose that we continue the 2034-level income-tax allocations to Social Security, with some marginal changes to adapt to the relatively modest Social Security cost increases that occur post-2035. That would retain continuity with the current tax code while avoiding the across-the-board benefit increases embedded in the Social Security 2100 Act.

Ideally, of course, conservatives — along with moderates, liberals, and everyone else — would put their own best Social Security policy on the table. “Policy” means going beyond simply how we finance currently promised Social Security benefits, or how much we cut back Social Security benefits to the level we’re willing to pay for. It means rethinking, in a 21st-century way, how an old-age and disability-insurance program should be designed.

I believe an updated program would look very different from one designed in 1935, when demographics, work patterns, and financial markets were very different. I’ve proposed gradually transitioning Social Security to a flat-dollar benefit that guarantees beneficiaries against poverty in a way the current program doesn’t, and that helps Americans to save to provide themselves benefits over and above that guaranteed government payment. This approach, which resembles the retirement programs in New Zealand and the United Kingdom, would strengthen poverty protections, increase saving, and improve incentives to work and delay retirement.

But however one thinks Social Security should be fixed, the expiration of massive income-tax transfers to Social Security come 2035 throw another option into the policy mix. Congress and the Trump administration should think over the opportunities for compromise it might offer.

Andrew G. Biggs is a resident scholar at the American Enterprise Institute. In 2005 he worked on Social Security reform at the White House National Economic Council, and in 2007–2008 served as principal deputy commissioner of the Social Security Administration.