wakalix

Here is the text of my email to Alan Reynolds:

Dear Mr. Reynolds,

I just read your article entitled "Has U.S. Income Inequality Really Increased?." I found some of the results very thought-provoking, particularly your point about the amount of dollars presumably that is presumably building up in tax-sheltered retirement accounts. Clearly, tax returns have a quite a few flaws as data points for measuring inequality (although perhaps IRA/401k contribution data is available on tax returns and could be included in a better analysis).

I do have a question about one of you calculations and would very much welcome a response. My question is about whether it is correct to add transfer payments to the denominator of your top 1% income ratio. In my understanding, these are just transfers from one income group to another, so they should already be in the denominator. If the transfers were actually drawn from the top 1%, then this wouldn't significantly affect the results. However, I believe these payments are mostly drawn from the first dollars of salaries (and not at all from capital gains), and hence would mostly come from outside the top 1%. One could argue that some fraction (maybe half) of the transfer payments are provided externally from corporations in the form of their half of payroll taxes, but we know those taxes are really just taxes on the earnings of the same individuals paying the other half of the taxes. This, of course, illuminates a serious flaw in looking at pre-tax income at all. It seems like any accurate income inequality study should look at disposable income.

I would also be interested in knowing whether you think you could repeat your analysis to show income stability for the top 0.1%.

Sincerely,

Andrew Brown

So I read Alan Reynolds piece and Piketty and Saez's response (which was disappointing to me). Here is my question for him?

Why does Alan Reynolds think he can add transfer payments in the tax ratio denominator as if they just materialized out of thin air? Aren't Social Security, Medicare and unemployment insurance paid for by the first dollars of our salaries? They are in fact "Transfer Payments" -- just the movement of money around in the denominator -- and they do not even represent transfers from the top 1% to the other 99%. I think he has a point about tax sheltered retirement plans hiding a significant amount of the income of the bottom 99%, but this other part seems like a remarkable oversight for a professional economist.

Also, there is some interesting analysis here from the IRS that I found on Krugman's blog:

http://www.irs.gov/pub/irs-soi/04asastr.pdf

Alan Reynolds of the Cato Institute published a critique of the Piketty and Saez data in the Wall Street Journal here. Reynolds argues that one must be careful when using income tax data because if the marginal tax rate is high, people with the means to shift the wealth to a non-taxable form will do so. If the marginal tax rate decreases, they will not, which will make it look like their income increases.

The authors respond here.

What I find curious is how the New York Times graph refers to the "nation's income" and "share of the United States individual income." To me, this type of language implies that "income" belongs to the nation (indeed, an apostrophe indicates possession). Where does income come from? Where does wealth come from? Does "the nation" create it and the "distribute" it to people in a manner that is seen as "fair" or "unfair" to certain people? Or do people create wealth?

Thomas Sowell has pointed out the that people in the top 1% change from year to year, and that income differs from wealth.

Here's a graph of the income of the top 1%: [link].

So I just read this piece and I'm trying to figure out whether it is deliberately misleading. I do thing the steady effective tax rate is an interesting finding, but I had two concerns:

1) Why doesn't the author actually tell us the effective average tax rate for the top 1%? If it doesn't change at all, why can't he just tell us the number. It just struck me as sort of odd that it was missing.

2) The author does write:

Using recent data, in other words, it would appear on its face that the Democratic proposal to raise taxes on the upper-income earners, and lower taxes on the middle- and lower- income earners, will result in huge revenue losses on both accounts. But some academic advisers to Democratic candidates have a hard time understanding the obvious, devising outlandish theories as to why things are different now. Well they aren't!

So I believe it to be the case that income reported by the top 1% of taxpaying households has been steadily increasing since 1980 (in real dollars and also as a percentage of GDP). Assuming it continues to increase, and the effective tax rate continues to be constant, then tax revenues should keep going up, all while the tax paid by the low- and middle-income households (who are becoming less and less a source of tax revenue) continue to decline.