Social Security and the National Debt

By James Kwak

In this season of fiscal brinksmanship, the topic of Social Security has once again come to the fore. Republicans are generally in favor of cutting benefits, although they are bit afraid to say so after the demise of George W. Bush’s privatization “plan”; Democrats are generally in favor of not cutting benefits. But many liberals have another argument: Social Security is irrelevant to the whole issue of deficits and the debt, since the program cannot have any impact on either.

I generally count myself as a liberal, but I think this is a misleading argument. This could take some time to explain, as I’ll try to go through it carefully.

The standard liberal argument goes like this: Social Security has its own funding scheme that is walled off from the rest of the federal government. Employees pay payroll taxes into the Social Security trust funds, and benefits are paid out of those trust funds. The crux of the argument is that Social Security, by law, may not spend money that it does not have in its trust funds.  It is impossible for Social Security to incur a deficit over the long term, since it can only spend money it already collected. In the words of Dean Baker:

“Social Security is prohibited from spending any money beyond what it has in its trust fund. This means that it cannot lawfully contribute to the federal budget deficit, since every penny that it pays out must have come from taxes raised through the program or the interest garnered from the bonds held by the trust fund.”

To begin, let’s clear a distracting issue out of the way. The annual federal budget balance is measured two different ways. The “unified budget” balance includes cash flows associated with Social Security (that is, payroll taxes in and benefit payments out). The “on-budget” balance excludes those cash flows (and the Postal Service). Obviously, Social Security does not contribute to the “on-budget” balance but it does contribute to the “unified budget” balance. But that’s all economically irrelevant, since in this case how you measure the thing doesn’t change the thing. (This isn’t quantum physics, after all.) The important number is the amount of money that the federal government has to borrow from the public, and that number is not affected by which budget balance you look at.*

Moving on to the main course: Dean Baker is absolutely right about current law. So for illustrative purposes, let’s posit an alternate universe, in which current law includes the following General Revenues Clause (GRC): “If the Social Security trust funds are unable to pay scheduled benefits to beneficiaries, those benefits will be paid out of general revenues.”

How would the existence of the GRC affect anything? It can’t have any impact on the national debt until the GRC begins to affect actual Social Security cash flows. And that won’t happen for as long as Social Security is able to pay benefits out of its trust funds.

The real question is what happens when the trust funds run out of money, which is currently expected sometime in the 2030s. In the current law universe, benefits automatically get cut to the level of incoming payroll taxes, which will be about a 25 percent cut; Social Security has no impact on government borrowing and hence the national debt. In my alternate universe, Social Security continues to pay full scheduled benefits, which requires additional government borrowing and increases the national debt.

The first question is: Which one do you think is more likely in the real world? Do you really think that Congress will sit by and let Social Security benefits be cut by 25% overnight? Or do you think that Congress will amend the law, essentially inserting the GRC, to preserve full benefits for seniors?

Of course we can’t predict the future with certainty, but I would say that if Congress in twenty-five years is anything like Congress today, it will find a way to pay full benefits. So when we talk about the impact of Social Security on the national debt, the most likely scenario is that it will increase the national debt—exactly as if the GRC existed today. That’s the main reason why I think it makes sense to take Social Security into account when projecting future national debt levels.

But let’s say I’m wrong and that Congress will stand by as seniors’ Social Security checks get cut by 25 percent. That’s not a good thing.

In my universe (where the GRC exists), when the trust funds run out money, Congress will have choices. It could: (a) increase taxes to pay scheduled benefits; (b) reduce other spending (like on aircraft carriers) to pay scheduled benefits; (c) borrow more money to pay scheduled benefits; or (d) cut benefits by 25 percent. In the current law universe, Congress will have to choose (d). It’s not hard to see that my universe is better than the current law universe. The two are equivalent if (d) is a better choice than (a), (b), and (c), but otherwise my universe is preferable. This is especially true for liberals, who in ordinary circumstances would prefer (a), (b), and (c) to (d).

So let’s say we know with certainty that the GRC will never be inserted and benefits will have to be cut. In that case, Social Security cannot increase the national debt. But we would actually be better off if the GRC existed and Social Security could increase the national debt, because of the options that would give Congress in the mid-2030s—including the options to raise taxes on rich people or cut defense spending.

In other words, if you are right that Social Security cannot increase the national debt, you should acknowledge that that is actually a bad thing. If you care about preserving Social Security benefits, you should prefer a world with the GRC.

Put another way, if you make the argument that Social Security cannot increase the national debt, you are conceding that there should be an across-the-board benefit cut the moment that the trust funds run out of money. Is that what you want?

On its face, the statement that Social Security cannot increase the national debt, and therefore should be off the table, seems like a classical liberal position. On inspection, though, I think it is misleading, since Congress is more likely to add the GRC than not. And for traditional liberals, I think it’s actually counterproductive, since it is premised on the unavoidability of a draconian benefit cut in twenty-five years.

* There is a separate, equally confusing, debate over the implications of the fact that the Social Security trust funds are invested in a special kind of Treasury bonds. The short answer is that because Social Security exists, the Treasury Department had to borrow less money from the public in the past few decades, when Social Security was running surpluses; but in the future, the Treasury will have to borrow more money from the public, because the Social Security trust funds will be redeeming some of those special Treasury bonds.