Taxes Per Person

Harvard economist Greg Mankiw argues that taxation as a percentage of GDP is a misleading way to compare national tax burdens and instead argues that we should consider taxes per person, which he calculates as Taxes/GDP x GDP/Person.   Using this metric, the United States is in the middle of the pack of major economies:

tax-shakedown-devilFrance = .461 x 33,744 = 15,556

Germany = .406 x 34,219 = 13,893

UK = .390 x 35,165 = 13,714

US = .282 x 46,443 = 13,097

Canada = .334 x 38,290 = 12,789

Italy = .426 x 29,290 = 12,478

Spain = .373 x 29,527 = 11,014

Japan = .274 x 32,817 = 8,992

I’m not sure that taxes as a percentage of income wouldn’t be more useful than taxes as a percentage of GDP, since what most of us are interested in with respect to taxes is impact on individual consumption rather than impact on aggregate production.  Then again, I’m not an economist.

It’s also worth noting that the GPP/Person figures are at PPP (purchasing power parity) rather than raw numbers.  I’m not sure what impact that conversion has on the comparison.

UPDATE: Brad DeLong points out that, by Mankiw’s measure, North Korea is a low-tax state.   In reality, it’s a high-tax state with a very low GDP.

Matt Yglesias extends that analysis to attack Mankiw’s premise.

Does Mankiw really think that Italy has more scope to increase taxes and the size of its public sector than does the United States? Or consider that in Slovakia per capita GDP is just $20,000. By Mankiw’s logic, Slovakia could raise taxes up to 65 percent of GDP and it would still count as a country with a below-average tax burden!

Common sense is that if you’re worried about the impact of taxes on growth, then when you’re worried about is the scope of taxation relative to the total amount of economic activity taking place. For Slovakia to try to raise as much revenue as we have in the United States would involve potentially ruinous levels of taxation. Conversely, for the American government to raise as much revenue per person as they have in France would be relatively easy.

Both fair points.  And, as a commenter points out, Mankiw’s per capita formula spreads the tax burden out to include children and others who don’t pay taxes.

If one’s goal is to determine individual tax burdens, it makes sense to calculate total taxes/total taxpayers.  If it’s to determine the impact of taxation on national productivity, though, the traditional taxes/GDP is almost certainly a better measure than Mankiw’s slight-of-hand.

FILED UNDER: Economics and Business, , , , , , , , , , , ,
James Joyner
About James Joyner
James Joyner is Professor and Department Head of Security Studies at Marine Corps University's Command and Staff College. He's a former Army officer and Desert Storm veteran. Views expressed here are his own. Follow James on Twitter @DrJJoyner.

Comments

  1. kth says:

    I’m not really understanding what GDP does in the formula, since it appears in the numerator and the denominator and hence vanishes. Couldn’t he have just taken total taxes paid (fed + state + local) and divided by the population? Unless (total taxes)/GDP is a known quantity, but total taxes is by itself unknown, except as derived from the former statistic? But that seems unlikely.

  2. Actually, Mankiw does not argue it is a better measure. He just proposes it as an alternative measure because it is convenient for his ideology.

    I know he’s a Harvard economist… and supposedly a brilliant guy… but most of his blog posts are little more the partisan talking points. They are often ill-informed, and frankly some are just gibberish.

    This one is particular is analytically useless.

    Can we measure taxes per square foot of home? How about taxes per raw tonnage of population? How about taxes per total years of life? Then we can pick which ever one is higher/lowest and argue that we’re over/under-taxed.

  3. john personna says:

    I love numbers. It is a criticism I’ve heard hurled by Brits at Americans, that we love numbers too much.

    So sure, calculate taxes a few different ways. One caution on “/person” though … it disguises differences in age distribution. It isn’t “/worker” or “/household.” It includes kids.

    I suspect with bigger, younger, families, Mankiw’s “/person” is going to skew … as the other commentators suggest.

  4. James Joyner says:

    Actually, Mankiw does not argue it is a better measure. He just proposes it as an alternative measure because it is convenient for his ideology.

    But his rationale makes sense if one’s goal is to figure out tax burdens on people.

    One caution on “/person” though … it disguises differences in age distribution. It isn’t “/worker” or “/household.” It includes kids.

    This strikes me as a fair criticism. It may well distort that data in a way that illustrates the concept Mankiw is trying to capture poorly.

  5. steve says:

    Mankiw claims that higher taxes depress GDP, which in theory makes sense. However, when one looks at the actual numbers, there is not such strong evidence. Factors other than tax rates seem to dominate. Link is to wiki list of taxes as a percent of GDP for all countries. The countries with the lowest tax rates do not seem especially successful. While this is an interesting way to look at taxes, I do not find it especially useful. Probably better to figure out what you want to achieve with your tax money, then set tax rates at the lowest level to achieve your goals.

    http://en.wikipedia.org/wiki/List_of_countries_by_tax_revenue_as_percentage_of_GDP

    Steve

  6. Steve Plunk says:

    I assume the tax per person measure is for all people, kids and retirees as well as productive workers. If we’re going to look at how taxes depress GDP I wonder if we shouldn’t look at taxes per person on those who add to the GDP?

    Hey, here’s another idea. How about we judge whether taxes are too high not by comparison with other countries but why what citizens of this country consider too high.

  7. Wayne says:

    I for one am more concern than just what percentage they take out of my paycheck. Taxes that hamper or destroys the economy hurts all. Destroying businesses and the capitalist system does nothing to help the population’s financial well being.

  8. Wayne says:

    Pretty aggressive spam filter.

  9. john personna says:

    It may well distort that data in a way that illustrates the concept Mankiw is trying to capture poorly.

    Heh. I think Mankiw knows exactly what he is doing (using his economic power for evil ;-).

  10. john personna says:

    I assume the tax per person measure is for all people, kids and retirees as well as productive workers. If we’re going to look at how taxes depress GDP I wonder if we shouldn’t look at taxes per person on those who add to the GDP?

    One classic way to divide it is into income quintiles, the 5 buckets into which we fall. I’ve shared taxes in each quintile before … ah, so has Mankiw. Those are only Federal, but we could assume the “shape” is the same for total taxes.

    I don’t really like the greater “spread” (with lower income folk paying less as a percentage of income and higher income folk paying more). I suspect though that it isn’t a “soak the rich” thing. I think it the lower quintiles have a real problem living the “American middle class lifestyle” … and we aren’t willing to tell them to live poor (in an apartment, without a car, ride the bus).

  11. Ugh says:

    Interesting. One thing to point out is just how much richer the United States is than some of the other “largest developed nations”, as Mankiw put it. Closest in per person GDP is Canada, we earn 20% more per person in GDP than they do, with just under 10 times more population. Huge difference.

    Further, in terms of taxes as a % of GDP, we’re the second lowest, yet people constantly go around whining about how we’re a high tax country.

    James: I’m not sure that taxes as a percentage of income wouldn’t be more useful than taxes as a percentage of GDP, since what most of us are interested in with respect to taxes is impact on individual consumption rather than impact on aggregate production.

    yes, taxes as a % of income would be interesting, and what is most interesting is that the 400 people with the highest reported adjusted gross income in 2007 paid an effective tax rate of 16.6% on an average of $344,800,000 in income. My wife and I’s effective tax rate that year was closer to 25% than 16.6%, with less than half of one percent of that income.

    Class war?

  12. Steve Plunk says:

    I think it’s important to include all taxes. Federal income tax, Social Security and Medicare, state income tax, local taxes, sales tax, property tax, fuel tax, government fees, basically all revenue diverted into the public sector. Use all taxes and divide into quintiles or even smaller categories.

    No matter how much we might raise incomes we will always have a lower class by definition. Telling them to live poor makes no sense in a country where people can rise out of poverty.

  13. Dave Schuler says:

    I doubt that any metric other than historic norms within a particular country is particularly useful in this regard. It’s a cultural issue. What’s tyranny in Germany might be considered quite acceptable here and vice versa.

  14. Dr. Schuler, but America is always held up as deficient in some respect against these relative measures by public expenditures in health care, foreign aid, the arts, ad infinitium. Could it be that they have different cultural inclinations in those areas as well?

  15. Median tax/person would be more useful then mean tax/person as well.

  16. James Joyner says:

    what is most interesting is that the 400 people with the highest reported adjusted gross income in 2007 paid an effective tax rate of 16.6% on an average of $344,800,000 in income. My wife and I’s effective tax rate that year was closer to 25% than 16.6%, with less than half of one percent of that income.

    Presumably, that’s a function of there being a cap on FICA contributions. That really, really skews the data but it’s not unreasonable given that there’s also a cap on Social Security payouts.

  17. steve says:

    “Presumably, that’s a function of there being a cap on FICA contributions. That really, really skews the data but it’s not unreasonable given that there’s also a cap on Social Security payouts.”

    Partially. It is also a function of income at the high end coming from investments.

    Steve

  18. Ugh says:

    Presumably, that’s a function of there being a cap on FICA contributions. That really, really skews the data but it’s not unreasonable given that there’s also a cap on Social Security payouts.

    Uh, perhaps at the margin, but it’s much more a function of a capital gains/dividend top tax rate of 15%, whereas the top rate on so called “ordinary income” is 35%.

  19. Definitely capital gains since not many people have a salary of $300M and considering the year they may also have had significant investment losses to offset gains, so it is probably important to specify whether or not we are talking about gross income, adjusted gross income, or something else. There can be other reasons too such as tax-free investments, e.g., state or municipal bonds, or various perversions of the tax code for social engineering purposes that have something other than the intended effect.

  20. Steve Plunk says:

    It occurs to me all of these measures are useful and should taken together to form a more complete picture. Like buying a sports car we factor in more than just horsepower. We look at weight to horsepower, braking, cornering, and other factors. Using all of these measures can only help understand what works and what doesn’t.

  21. pete says:

    What would work is The Fair Tax. Then this discussion would not be happening.

  22. Fausta says:

    Greg ought to drop by Princeton Township and find out the per capita property tax burden exceeds his $13,097.
    Which puts little Princeton Township in the middle of the pack of major world economies. I’m sure that’ll warm Greg’s heart.