Liveblogging World War II: November 13, 1942
Jérémie Cohen-Setton: Empirical and theoretical multiplier uncertainty

Tax Shares Paid by the Rich: Does the U.S. Have the Most Progressive Tax--Not Tax and Transfer, But Tax--System in the OECD?: For the File Weblogging

We know that the U.S. has one of the least progressive tax-and-transfer systems in the OECD.

Suppose we don't like that answer, and seek the answer to a different--and less interesting, and less relevant--question: suppose we ignore transfers, and just look at taxes? Could Veronique de Rugy be right in her extremely narrow question, and the U.S. have the most (or one of the most) progressive tax systems in the OECD?

Note that the thumb is heavily on the scale in this formulation of the question.

U.S. data on the share of taxes paid by the top 10% does not mean what it means for other countries. As a result of historical accident--the fact that Sen. Russell Long (D-LA) was on the Finance Committee--the Earned Income Tax Credit and the Child Tax Credit parts of our welfare system are run through the Treasury and the IRS rather than through HHS. These large "negative taxes" reduce the denominator of total net taxes collected, and so inflate measurements of the share of taxes paid by the top 10% of Americans.

Suppose we have an economy in which the top 90% pay a 20% tax rate and the top 10% pay a 40% tax rate. If the top 10% have 20% of the income, they pay 20% x 40% / (20% x 40% + 80% x 20%) = 33% of the taxes. If the top 10% have 40% of the income, they pay 40% x 40% / (40% x 40% + 60% x 20%) = 57% of the taxes. Dividing tax shares by income shares gives us a ratio of 61% for the first economy and 70% for the second. Even though the tax systems are exactly the same, the one with more inequality has a greater ratio of taxes paid/income at the top than the one with less inequality. The ratio of taxes paid/income at the top is a bad measure of tax progressivity--not as bad a measure as the share of taxes paid at the top, but a bad measure.

What is a better measure of progressivity? A better measure would be to ask whether the rich are paying a greater share of taxes than one would expect given the fact that income is more unequal. And that leads us to look at:

Workbook1  version 1 xlsb   Recovered

The U.S. looks like what a normal OECD country with its extraordinarily unequal income distribution would look like--if, that is, you accept that Italy and Poland are not normal OECD countries but outliers. Lumping the U.S. in with the rest of the OECD, the American share of taxes paid by the top 10% is larger than the share paid in other countries by what you would expect given the pattern between income and tax shares seen in the (non-Italy, non-Poland) OECD.

Does the U.S. have the most progressive tax--not tax-and-transfer, but tax, system; and pretending that the data for the U.S. mean the same thing as the data for other countries by ignoring that if the U.S. were like other countries the program-equivalents of the EITC and CTC would be not in Treasury but in HHS?

Looks to me like the answer is: "no". Korea, France, Finland, the Netherlands, and Australia all look to have more progressive tax systems than the U.S. (And, of course, nearly everybody has a more progressive tax-and-transfer system. And, of course, more than half have clearly more progressive tax systems if the denominator is gross rather than net of refundable credits.)

Comments