U.S. Taxes Really Are Unusually Progressive

If you ask me, Jonathan Chait, a writer I respect, has made an ass of himself in a fight he picked with Veronique de Rugy over taxes and progressivity. She offended him by saying that America's income taxes are more progressive than those of other rich countries. Chait assailed her "completely idiotic" reasoning, called her an "inequality denier", "a ubiquitous right-wing misinformation recirculator" and asked if it was really any wonder he cast insults now and then at such "lesser lights of the intellectual world". (Paul Krugman said he sympathises. With Chait, obviously. The only danger here is in being too forgiving, Krugman advises. Chait may think the de Rugys of this world are only lazy and incompetent, but we know them to be liars as well.)

Just one problem. On the topic in question, De Rugy is right and Chait is wrong.

Income taxes in America are more progressive than in other rich countries--according to an authoritiative official study which, to my knowledge, has not been contradicted. The OECD's report "Growing Unequal", on poverty and inequality in industrial countries, includes a table that provides two measures of income tax progressivity in 2005. This is evidently the source of de Rugy's numbers. Here they are in an excel file. According to one measure, America's income taxes were the most progressive of the 24 countries in the sample, except for Ireland. According to the other, they were the most progressive full stop. (A more recent OECD report, "Divided We Stand", uses different data, a smaller sample of countries and a different measure of progressivity: the results are similar.)

Before you ask, this ranking takes account of employee-side payroll tax as well as the federal income tax.

Chait first objected to de Rugy's claim about progressivity because he thought she was inferring it from the fact that the US collects the biggest share of income taxes--45 percent of the total, col B1 in the table--from the top income decile. That would be a false inference, as Chait says, because it could be true of a country with a very unequal income distribution even if its taxes were not especially progressive. But look at the table. There was no need for de Rugy to draw any such inference, let alone try to mislead readers. All she needed to do--and all, I'm sure, she did--was glance over to the last column, which actually gives the measure of progressivity, showing the US to have the highest score.

The measure of progressivity is hard to explain, so I can see why de Rugy quoted the tax share instead. But she could have chosen a much more dramatic number if she was seeking merely to bamboozle her readers. Exclude payroll tax, and the top 1 percent of taxpayers, not the top 10 percent, have lately accounted for nearly 40 percent of income tax receipts, the top 5 percent for nearly 60 percent, and the top decile for roughly 70 percent. (Here are the IRS data, excel file.)

For the reason I just gave, this does not prove that the US tax system is more progressive than anybody else's--but it surely has some relevance to the question, "Are the rich paying their fair share of income tax?" If this isn't fair, what would be?

When Chait, with all the authority of a leading light of the intellectual world, says "Rich Americans pay a bigger share of the tax burden because they earn a bigger share of the income, not because the U.S. tax code is more progressive," he is making the same kind of sloppy bias-driven error he falsely accuses de Rugy of making. (I'll refrain from wondering whether he made the mistake deliberately.) According to the OECD, rich Americans bear a bigger share of the tax burden because they earn a bigger share of the income and because the US income tax system is more progressive.

There's a lot more to say on this subject.

Is measuring progressivity straightforward? No. It's difficult, because the underlying data are very complicated and hard to compare across countries. Another problem: expressing progressivity across the whole income range as a single number, so that one can say A is more or less progressive than B, can be misleading. Unfortunately, we all want to be able to say, A is more or less progressive than B.

Why, according to the OECD, is the US system so progressive? Not because the rich face unusually high average tax rates, but because middle-income US households face unusually low tax rates--an important point which de Rugy mentions and Chait ignores.

How does the picture change if you take indirect taxation into account? That would make the US system look even more progressive, because the US doesn't rely on a flat consumption tax like most other governments.

What does the tax-rate schedule tell you about progressivity? Very little, until you factor in deductions, thresholds and opportunities for avoidance. High top rates are meaningless if nobody is paying them.

What happens to tax progressivity at the very top of the US income distribution--inside the top one percent, or 0.1 or 0.01 percent? Good question. When it comes to international comparisons, I don't know of any good answers. But note that the US isn't the only country to tax capital at preferential rates.

Should the US system be even more progressive than it already is, bearing in mind the skewness of the pre-tax distribution, especially at the very top? Maybe it should--but a steeper tax-band schedule is a dumb way to improve fairness and raise more revenue.

I'll take some of these questions up in a bit more detail later. To be getting on with, here's a column I did that touches on the subject.

For Chait at his excellent best, try this. Meanwhile, I'd say he owes de Rugy an apology.