(This is the second in a series of follow-up posts to my TNR column. The first is here. The comments to which I'm responding include Kevin Drum's, Kieran Healy's, He Who Must Not Be Named's.)

The objection is: The problem with the Wall Street Journal "Lucky Duckies" editorial wasn't that it involved using taxes on the working class for partisan gain. It was that the editorial rested on a deliberate falsehood: that there are millions of Americans who don't pay taxes. In fact, taxes in America are barely progressive overall, and even the poorest Americans pay a significant share of their income in payroll, property, and consumption taxes. This means three things. First, The WSJ editorialists are a bunch of lying bastards, and I shouldn't have wasted my breath (sort of) defending them, as I was defending them against the wrong charge, a more "hifaluting" one than the one that was already made. Second, the WSJ's implicit substantive proposal-- avoid further tax cuts for the working class-- is immoral, since (in combination with the WSJ's favord tax cuts for those higher up) it would leave the American tax system overall actually regressive. Third, the WSJ's political story-- that the constituency in favor of lower taxes is shrinking to a dangerously low point, and too many Americans are becoming detached from the costs of government, thereby losing their incentive to vote to keep the size of government (and taxes) in check-- is wrong, because those people are paying taxes. (This is a shorthand summary, but I hope a fair one.)

My responses:

First and foremost: I agree that non-regressivity is a desirable characteristic of a tax system. (I've already alienated some of my libertarian readers here-- some, by conceding that there could be any such thing as a desirable characteristic of a tax system, others, because I've embraced a distributional-pattern norm about taxation.) I'm going to remain agnostic here as between proportionality and progressivity-- it doesn't affect the substance of the objection, which is that the WSJ's combination of proposals would tip us into regressivity and that's wrong.

Non-regressivity isn't the only desirable characteristic of a tax system, mind you. Besides obvious rule-of-law, non-arbitrariness standards (no differential tax rules based on race; tax rules that are knowable in advance, one is only punished for actually breaking an actual rule, etc (not that the obviousness of these standards means they're particularly well-satisfied under our current tax system)), let me suggest some others.

2) The tax system, taken as a whole, should be relatively transparent, such that increases or decreases in the expense of government are pretty directly and widely experienced as increases or decreases in the tax burden. (For many people much of the time, those increases or decreases will be offset by decreases or increases in state-provided benefits. That's compatible with one's tax bill bearing a rational and consistent relationship to the tax system overall.)
3) The tax system should have room for jurisdictional competition on the optimal balance of taxes and benefits. That means that it should be federal or confederal, and the component units should have independent taxing and spending authority. Ideally this extends to the municipal, not only the provincial, level.
4) The tax system should carry a light administrative burden-- both in the aggregate and across taxpayers widely. Note: this tends to favor consumption and, to a lesser degree, property taxes over income and payroll taxes-- not always, but in general. It's thus in direct tension with progressivity. I suspect it weighs particularly heavily against multiplying the number of income-taxing authorities at lower levels, i.e. cities and towns.
5) The tax system should be minimally distorting in the aggregate-- no sectoral favoritism. (Yes, this criterion weighs against the mortgage deduction.) In particular it should be minimally distorting of the labor market.
6) (Here I'm really going to antagonize some of my fellow libertarians:) The tax system should encourage a sense of shared connection to and ownership of the state. This is closely related to (2), but I've become convinced that it should be stated separately-- because of the "resource curse" problem. Taxation is undoubtedly a cost, not a good. But it turns out that states funded by magic pots of money dug out of the ground-- something that should be very appealing to non-anarchist libertarians, who want a state but to minimize or eliminate taxes-- have all sorts of dreadful characteristics. There seems to be a strong connection between a broad tax base on the one hand and public accountability, reduction of corruption, reduction of militarism, diffusion of power and authority throughout the society on the other.

I would also add (7) The tax system should be sufficiently transparent that the burdens it creates are widely noticed, and thus create political pressure for keeping them in check. (This was the argument against automatic witholding on the income tax.) But that's a matter of my political preferences in a more-obviously controversial way, and I don't think it's reasonable to demand that everyone agree with that as a desirable criterion of a tax code. I think that (1)-(6) are plausible candidates for widespread acceptance as desirable criteria, each taken ceteris paribus, though people will obviously disagree on the weight to be assigned to each one.

There are certainly others. But it seems to me that, already with these, we've got obvious difficulties of simultaneous satisfaction. With lots of taxing authorities but reason to not want them all to tax income, there's going to be a tendency toward regressivity because of the use of consumption and property taxes. This creates pressure to make the income tax, where used, more progressive-- which tends to make it both more complex and more distorting. Moreover, the existence of lots of separate taxing authorities that use at least three different tax bases yield a serious reduction in transparency. There may be no rational relationship between increases or decreases in the cost of government overall and any representative person's annual tax burden, because the cost increases or decreases may get concentrated in odd ways.

The WSJ, as I read it, was centrally concerned with (2) and (6)-- and I think those are reasonable concerns. They're not the only concerns. But they're real and legitimate ones. And sometimes one can't address one of the desirable characteristics without creating pressure in an undesirable direction on one of the others.

And I doubt that (2) and (6) are satisfied simply by pointing to how broad the base of overall taxation in America is. The payroll tax in particular (about which more below) is widely understood to be something other than a straightforward tax, and not to have a direct connection to the cost of government in a general way. Property taxes can fluctuate wildly for reasons that have little to do with the cost of government-- e.g. a major commercial landowner comes to town, or leaves it. Statewide sales and income taxes, and the federal income tax, are the ones that people associate with the cost of government in a general way, and that move up or down with that cost (except during the Bush years, when costs skyrocket and income taxes go down-- what a clever idea...) Insofar as millions of people cease paying any income taxes, it does seem that they become detached from the costs of the federal government (and of some state governments). The WSJ's political case was incompletely and incompetently stated; but rests on something plausible. (Of course, even in an editorial one ought to show some awareness of complexity. The WSJ's apparent willingness to just ignore the payroll and other taxes, rather than engaging in this kind of explanation, was deceptive and I do not defend it.)

The WSJ might still be wrong, of course, and I suspect that it is. The progressive income tax with witholding seems to be the most invisible of taxes, as far as criterion (1) is concerned. Property taxes spark tax revolts. Complexity in the income tax sparks some general resistance and grumbling that can be capitalized on. But small, year-in and year-out increases in income taxes, deducted every two weeks or every month from the paycheck, don't seem to do that. States with income taxes have much higher overall government spending and tax burdens than those with only sales taxes or with neither. Causation almost certainly runs both ways here, but the evidence suggests to me that simply being on the income tax rolls does not turn people into vigilant monitors of the cost of government, and indeed that income taxes do that less than other taxes do.

That said, what about morality? Recall: "the WSJ's implicit substantive proposal-- avoid further tax cuts for the working class-- is immoral, since (in combination with the WSJ's favord tax cuts for those higher up) it would leave the American tax system overall actually regressive."

Here I think a great deal turns on how we understand the payroll tax. (Note to the two dozen people who e-mailed this point to me; thanks, but I was already working on this post to begin with and I understand this point perfectly well.) If the payroll tax is just a tax, not something special like an insurance premium or forced savings-- that is, if it ought to be counted in a way that is detached from Social Security payouts at the other end-- then the tax system is barely progressive, and could be tipped into being regressive overall. If the payroll tax is not just a tax but something special that should be counted alongside the ostensible benefits it ostensibly purchases, then this is not true. The payroll tax is the largest share of the tax burden for (I believe this is right) most Americans. It is proportional-- but only on earned income, and only up to a certain income cap. So it makes a big difference in the overall calculations-- and how we count it makes a big difference.

I think that the payroll tax is just a tax, and that Social Security is just a spending program. But that's not the official position; and it's not ordinarily the position of those who support a continued state system of Social Security. Hence the desire to take Social Security off-budget, to treat its surplus as different from other state funds, and so on. Hence the unwillingness to means-test Social Security-- in order to preserve the illusion of social insurance "bought" with one's "premiums." Hence the unwillingness to lift the income cap on the payroll tax-- because then we'd either have to allow benefits paid to the richest retirees to skyrocket, or we'd have to admit that one's taxes don't really purchase one's own benefits. In other words, I want to concede that the payroll tax is just a tax, making the tax system as a whole borderline-regressive-- and then to insist on the consequences that follow from that. But I dislike any attempt to have it both ways-- to treat the payroll tax as just a tax for purposes of calculating regressivity, but to treat it as something quite different when it comes to discussing Social Security as a program.

Posted by Jacob Levy to The Volokh Conspiracy at 9/19/2003 04:32:54 PM

Powered by Blogger Pro

Reply via email to