The Social Security Tax: It's a Candy Mint, and a Breath Mint

-- Posted by Neil H. Buchanan

In my Verdict column this week, I discuss one of the new talking points on the political right in the United States, an argument with its roots in the false debate over "who pays for government." The claim -- which is not really new, given that I first encountered it over four years ago, but is new in the sense that it has gained prominence for the first time as a major political argument -- is that Social Security and Medicare taxes are not really taxes, because they entitle taxpayers to receive benefits in the future. Thus, the argument goes, what appear to be taxes are actually merely insurance premiums.

One might think that the anti-tax party would be anxious to label everything government does a tax. Calling something "not a tax" might, after all, result in the non-tax being increased, perhaps even progressively. (Proposals to extend the SS and Medicare taxes to all income, not just labor income, come to mind. Also, lifting the cap on earnings subject to Social Security.) The political payoff from this argument, however, is that it feeds into the claim that roughly half the country is getting a "free ride" from the wealthy people. Taking SS and Medicare taxes out of the mix significantly slants the distribution of tax payments toward the rich, because the overall federal tax system is designed to impose income taxes mostly on the higher end, with payroll taxes being the tax that everyone else pays. This allows the apologists for the rich to claim that the rich are paying more than enough, and that everyone else should be grateful, rather than clamoring for more redistribution.

My basic argument in the column is that -- even putting aside the rank hypocrisy of right-wingers describing entitlement spending as guaranteed payouts, even as they claim that the trust funds are meaningless and that the government should cut those benefits -- this attempt to relabel Social Security and Medicare taxes is simply a ham-fisted effort to measure all of the costs but only some of the benefits of government. That is, somehow Social Security and Medicare taxes are not taxes, because paying them results in the government doing something good for the payors. If that is the standard, however, then obviously we should not limit the analysis merely to the benefits that people receive from Social Security and Medicare.

The appropriate approach is to determine all of the costs paid, and all of the benefits received. And if one looks at all of the costs and benefits, not only is government a net plus -- making it possible to have a $14 trillion economy, rather than roving bands of hunters and gatherers fighting over berries -- but the biggest gainers are the wealthy, even after taking into account their taxes. Isolating SS and Medicare benefits is tempting only because they are provided in dollars and are nominally tied to a unique funding source. But the aggregate benefits of government are felt by the wealthy in ways that go far beyond checks from the Treasury (although they receive plenty of those, too).

Even this analysis, however, is incomplete. The essence of the argument in my column is: "Government benefits include more things than entitlement benefits." What should be added is: "Government costs include more things than just taxes paid." Suppose, for example, that it were actually true that the wealthy paid all taxes, while everyone else received direct benefits from the government. It would still be true that the rich in the United States (and in every other prosperous nation) gain more than everyone else does, due to the existence of the government. Even so, we would seriously mismeasure the distribution of costs of the government if we did not note that government policies allow high concentrations of income and wealth to exist, in part by determining the level of wages. Changes in various government policies could easily change the distribution of wealth, increasing the earnings of the non-rich. (The following analysis, by the way, is not affected by the possibility that such redistributive changes might lower overall GDP, except under the most extreme assumptions.)

If the government's policies allow employers to pay their workers less than they would under alternative sets of policies, then the "cost" of government to the non-rich must include the reduction in their wages caused by the current set of policies. For example (using purely illustrative numbers), if one set of policies would give the average worker a pre-tax income of $60,000 and someone in the top 1% an income of $2 million, while the current set of policies results in incomes of $50,000 and $3 million, respectively, then the current government policies have resulted in lower before-tax incomes for the non-rich. In the latter case, you could levy a 0% tax rate on the non-rich, and collect taxes directly only from the rich, and you would be tempted to say that the rich have "paid for government" -- even if their after-tax incomes were well in excess of $2 million, and even though the non-rich have 16.7% less in pre-tax incomes.

What all of this tells us is that the effort to measure separately the costs and benefits of government -- and to attribute those costs and benefits to particular groups -- is ultimately a fool's errand. I am hardly the first person to have reached this bottom line, but it is worth repeating that the truly important distributional questions are all about after-tax-and-benefit standards of living.

What about the title of this post? When I argue that it is false to say that the Social Security and Medicare taxes are not taxes, that does not mean that it is never appropriate to think about those programs on their own. In a blog post earlier this week, for example, Bruce Bartlett pointed out that it is sometimes sensible to ask whether people's behavior is affected by their perceptions of what their Social Security and Medicare taxes mean: "To the extent that workers perceive a linkage between the Social Security taxes they pay and the benefits they receive, the Social Security system reinforces work incentives rather than being a tax on work, as is commonly believed." He raises this possibility in the context of a discussion of whether the temporary payroll tax reduction is stimulative.

There is nothing inconsistent between my argument and Bartlett's. I am arguing that, to the extent that it even means something to measure "who pays for government," it is dishonest to single out Social Security and Medicare taxes as "not really taxes." He, on the other hand, is offering an empirically testable claim about workers' responses to this year's reduction in their payroll taxes. I can be completely agnostic about Bartlett's claim, without undermining mine. Thus, even though he begins his discussion of that issue by saying, "Another issue is whether the Social Security tax is really a tax at all," he is only arguing that the Social Security system's design might reasonably lead people to change their behavior. The Social Security tax can be both "a tax" in the broader sense in which I am using it (and in which the people to whom I am responding must use it, in the context of their argument), and "not a tax" in the sense that Bartlett uses it.

Along somewhat similar lines, a new Urban Institute policy brief by Gene Steuerle and Stephanie Rennane examines the distribution of net lifetime benefits paid out by Medicare and Social Security. Even though they find that Medicare pays out more than it takes in from dedicated taxes (while Social Security is generally balanced in the aggregate), that can only tell us whether or not those programs are partially funded by other taxes, in the long run. While that can tell us something useful about the programs' finances, it says nothing about the "Who pays for the government" question, in light of the appropriately broad framework described above.

Similarly, a new working paper by tax prof Erik M. Jensen, "A Tax or Not a Tax: That is the Question," includes the following claim (in the paper's abstract): "The Social Security levy might be understood as payment for a specific benefit, a retirement and disability plan, not a tax." The very next sentence, however, reads: "If so, constitutionality would not depend on the taxing power." The analysis is a legal one, parsing whether the taxing power under the constitution extends to certain ways in which money is collected and spent. That analysis has nothing to do with my analysis above.

As I argue toward the end of my Verdict column: "While there are sensible reasons to structure certain government programs as stand-alone entities, the proper way to summarize the overall benefits and burdens of our governments’ activities is to take the whole of what we spend, and compare it to the whole of what we receive." Regular readers of this blog know that I take very seriously the legal. economic, and political implications of Social Security's status as a nominally self-financed entity. That does not mean that it is appropriate to treat Social Security and Medicare separately when we are defining "taxes" for the purpose of measuring something so broad as the costs and benefits of government.

It is not mere wordplay -- nor arm-chair quantum physics -- to say that something can be both a tax and not a tax at the same time. As always, context matters.