Tax Avoidance

Warren B’s call for higher taxes on the rich should delight their accountants.   We need job growth and a growth in the number of accountants would count!  His OP-ED reminded me that I have written an unknown paper on this topic. 

Bill Gentry and I had a smart thought.  For each U.S zip code, you can collect data on how much income is reported to the IRS.    For the same zip codes, using year 2000 Census data, you can collect data on how much income is reported to the Census interviewers.  For each zip code, we took a ratio of these numbers.  My intuition was that people would use legal and illegal means to shade down their taxable income reported to the IRS but would tell the truth to the Census.   In this case, this “batting average” of  (IRS Reported Income/Census Reported Income) is an interesting zip code level variable. We then study in which areas of the nation is this ratio high?   These are “patriotic Buffet” places that are paying their taxes.  If you look at our Table 4, we find that in states where the tax rate is high that this ratio is lower. 

Switching subjects, Michael O’Hare and I are having lunch today.   The RBC is a community of scholars.


Author: Matthew E. Kahn

Professor of Economics at UCLA.

11 thoughts on “Tax Avoidance”

  1. The Census estimate of household income also could be influenced by a social status component. People might over-report income on the census interview to appear richer than they are.

    Are the high ratio areas prestige seeking?

  2. Not a post I expected to read here. I do have a question. What about (equivalantly)

    1) raising tax rates but capping deductions so that for every dollar of allowed deductions families with income over $250,000 pay 35 cents less in taxes.
    2) adding a tax on gross income plus capital gains (0 up to say 300,000 per family then say 6%) to the income tax.

    I think tax avoidance consists mostly of legal reduction of tax liability and not out and out lying (I think your data show the rate of plain lying). In any case, one doesn’t need an accountant to lie about one’s income (nor would one help if it’s on one’s w2 say).

    I think it is possible to take more from the rich without increasing incentives to find deductions or disguise income as capital gains. It seems to me to be a simple problem. I am proposing something radically new. Maybe it could be called an alternative minimum tax.

  3. I find the analysis intriguing, but not very conclusive. One could easily imagine systematic variations between IRS and Census income even for scrupulously honest people. For example, a lot of moderately wealthy people in the mid- to late-accumulating stage of life would not think of their substantial interest and dividends as income, because they automatically reinvest them and live only off their earned income. This might lead to underreporting to Census, as they might only be thinking of earned income when asked the question. But they would report the unearned income to the IRS, because tax forms force them to focus on unearned income.
    This wouldn’t explain Kahn’s data, because such people would be more prevalent in the wealthy states, which are also the states that tax income more highly. (Red states prefer regressive taxes.) And I agree with the basic intuition behind Kahn’s study: that higher taxes generate more avoidance activity on the margin. Not that this is a conclusive argument against higher taxes, for the same reason that the marginal pollution generated by marginal economic growth is not a conclusive argument against economic growth.

  4. “The RBC is a community of scholars.” I beg to differ. It includes bloggers like me and – much more important – a faithful body of commenters, who don’t have PhD’s to their names. Its general type is admittedly the academic popularisation blog, in which scholars talk and listen to a self-identifying educated public in a level coffee shop. A community that values evidence and scholarship, sure.

    Enjoy your lunch with Michael as much as I did.

  5. I’m inclined to think that a person who engages in questionable tax sheltering would also be inclined to make their IRS and Census reportings match. There may not be a rational basis for them to do this but if you’ve got something to hide you don’t do anything to discredit your lie. Plus you don’t trust the government to begin with so any promise to keep IRS and Census data separate is just a “promise” to them.

  6. “If you look at our Table 4, we find that in states where the tax rate is high that this ratio is lower.”
    Am I correct in understanding that a link is being proposed between tax rates and willingness to report taxes fairly, such that higher tax rates correlates with higher rates of avoidance?

    This is general claim from the right, right? If you raise rates too high you end up losing revenue through avoidance. On their face, these findings would seem to back that up.

    There’s a number of ways to go here. First, the interpretation might be wrong. Is there a qualitative difference between different tax payers in different states, in terms of level of income, etc.?

    Although assuming the claim is true, it is an interesting problem. My first notion would simply be better enforcement. I think people would generally see a lot of justice in going after high-class tax dodgers. I’d personally much rather see prisons filled with suits instead of potheads and meth abusers.

  7. Eli: the laffer curve claim is that you lose taxes through loss of economic activity (which is tax avoidance, but in a pretty drastic, not-usually-used sense). The weaker claims are that a) each percentage point of increase in tax rate yields less than a percentage point increase in revenue (which is mostly irrelevant, because it’s not as if there’s a direct cost to the government of setting tax rates at a particular point) and b) that you distort the economy because money gets diverted into paying tax consultants and into activities that would not be lucrative except for the tax consequences. Because, as we all know the free market unfettered by tax concerns is better at allocating resources than any other mechanism.

  8. “My intuition was that people would use legal and illegal means to shade down their taxable income reported to the IRS…” Oh, joy! We can be as Greece, too.

    The RBC is a “community of serious-minded people,” with the exception of one or two. 😉 All in all a very civil place. And not a little useful. I must have missed on thing, though. Has the proprietor yet explained to us how BHO nailed that debt ceiling negotiation (sic)?

  9. An example might be helpful, so I’ll use my own situation last year as an example of tax avoidance.

    My son was in the hospital for over a month, and I had only 3 months work. When I did my taxes, I used the standard deduction, even though almost certainly the itemized would have been higher given the hospital bill–because since I was out of work so long, my income, and hence my marginal rate, was low (15% IIRC).

    If I had had a high income and hence a high marginal rate (say 40%), it would have been worth the time and trouble to itemize deductions–a perfectly legal and above-board way of reducing my reported taxable income.

  10. “Because, as we all know the free market unfettered by tax concerns is better at allocating resources than any other mechanism.”

    Thing is, it doesn’t have to be better than any conceivable mechanism. (Such as a command economy controlled by hyper-intelligent altruists with perfect information.) It just has to be better than what you’re actually going to get if you abandon it. In reality, that turns out to not be much of a challenge…

Comments are closed.