The Laffer Curve and New Evidence that Taxes Stifle Economic Output

Many people may remember Christina Romer, who was the chair of President Obama’s Council of Economic Advisors from January 2009 to September 2010.  She is currently an economics professor at University of California Berkeley.

Approximately three months before she left the Council of Economic Advisors, the American Economic Review– a journal which many people consider the most prestigious peer-reviewed journal in economics – published an article that she co-wrote with her husband David Romer, also an economics professor at UC Berkeley.  The article, “The Macroeconomic Effects of Tax Changes: Estimates Based on a New Measure of Fiscal Shocks,” is in my view one of the most important economic articles of the last several years.

The Romers examined the effects of tax policy on GDP.  They found that the effects are very large.  Specifically, they found that for every 1% that taxes rise (as a percent of GDP), this causes GDP to fall about 3%.  The authors employed some clever methods to try to find what economists call “exogenous” changes in tax rates.  When they employed their methods, they found much higher effects than economists had previously found.

The article was something of a Nixon-goes-to-China phenomenon.  That is, while conservatives tend to claim that taxes strongly decrease GDP, liberals tend to claim that taxes have at best a weak influence on GDP.  When the Romer-Romer article reported a strong influence, one of the most interesting aspects of the finding was that it came from a very liberal quarter – namely, one of its authors was a senior member of the Obama administration.

The article, in my view, should have been big news.  However, it’s now been two years since it was published, and I am aware of no mainstream news outlet that has mentioned it. 

This is especially surprising given its implications for the Laffer Curve.  Specifically, the Laffer Curve specifies that there exists a “hump” tax rate – a rate that maximizes revenue to the government, and if the government raises taxes above the hump rate, then its revenue actually decreases.

Academic economists generally agree that the hump rate is very high, something like 70%.  However, although Romer-Romer article did not explicitly discuss the Laffer Curve, its results imply that the hump rate is much lower, something like 33%.

To see this consider the following example.  Suppose a country’s GDP is $100 billion, and suppose its tax rate is 33%.  Then its tax revenue will be 33% of $100 billion, or $33 billion.  Now suppose it raises taxes to 34%.  If the Romer-Romer result is accurate, then this will decrease GDP by 3% to $97 billion.  Tax revenue will be 34% of $97 billion, or $32.98 billion.  Note that this is slightly less than the revenue at the 33% rate.  If you experiment with other tax rates, you’ll see that revenue is maximized when the tax rate is 33 1/3 %.   Moreover, as the tax rate increases to rates higher and higher than 33 1/3 %, government revenue becomes smaller and smaller.

(Alternatively, one can show the above points more rigorously with calculus.  Namely, let Y(t) be the GDP of a country, which is a function of t, the tax rate it chooses.  Romer and Romer show that if a country increases its tax rate by 1%, then GDP falls by 3%.  That is, the decrease in GDP, dY, is approximately .03 Y(t).  Note that .03 is 3 times the change in the tax rate, .01.  Thus, when we define dt as the change in the tax rate, the Romer-Romer finding suggests that dY = -3*dt*Y(t).  Thus, the derivative of Y with respect to t, dY/dt, is -3Y(t).

Note that total revenue for the government is t*Y(t).  To maximize revenue, we take the derivative with respect to t and set the result equal to zero.  I.e. we want t to satisfy the following:  t*dY/dt + Y(t) = 0.   Now let us substitute the above expression for dY/dt.  We get t[-3Y(t)] +Y(t) =0.   Once we divide both sides of the latter equation by Y(t), we get -3t+1 = 0.  Some algebra shows the solution to the latter equation is t=1/3.  I.e. the tax rate that maximizes revenue is 33%.)

U.S. taxes are generally below 33%.  For instance, the total amount of federal taxes is about 17 or 18% of GDP.  If you add in state and local taxes, this raises the percentage to something in the low twenties, but still significantly below 33%.

However, the Romer-Romer result likely applies not just to the entire U.S. economy, but also to subsets of it – such as the subset of very rich taxpayers.  They currently face a top federal income tax rate of 35%.  And if you add in their state and local taxes, their rate reaches the high thirties, and the low forties in some states.  The Romer-Romer result suggests that they are above the hump rate.  If we increase their tax rates—as President Obama says he plans to do—then the government would actually receive less revenue.  President Obama claims that he can reduce the deficit by making the rich pay their “fair share” of taxes.  However, the Romer-Romer result suggests the opposite might happen – the deficit might actually increase.

As I mentioned, all this should be big news with the U.S. media, but it has not been.  Although I suspect that the reason is that liberal journalists want to squelch the Romer-Romer results and hide them from the public, perhaps the true reason is more innocent:  Perhaps U.S. journalists simply never learned about the Romer-Romer results. 

If so, over the next several days, we will get to observe something akin to an experiment.  On Monday, Prager University will post a new five-minute video to its web site describing many of the above findings.  In addition, Glenn Beck’s web site, “The Blaze,” is scheduled to publish an article about these results, and the Daily Caller is scheduled to publish an op-ed by me describing the results.

Perhaps the latter outlets, in addition to this Ricochet post, will help spread the word about the importance of the Romer-Romer result.  I suspect that it will but only to a small group of conservative intellectuals. 

On the other hand, maybe word will spread more widely, and the Romer-Romer result will receive the attention it deserves.  I will be watching with much interest.

  1. Aelreth

    Could we add to that, tax revenues to GDP regardless of tax rate fall at or around 19% of GDP?

  2. Herbert Woodbery

    Most high income people pay lower rates than middle/upper middle  income people.  That’s why the buffet rule has such appeal.    Mitt claims to have paid at least 13 percent ,  a long way from laffer curve influence.

  3. raycon and lindacon

    What has taxation got to do with revenue.  Its purpose is to manage behavior.  We are awash in money.  We can get all we need by mortgaging our future in bonds.  Taxation is now superfluous to spending, and by any measure not very important to the financial equation.

    But, as Maggie the Great said; Socialism will collapse as soon as they run out of other people’s money.

    The major difference between the US and Greece is order of magnitude, and an historic reputation for fiscal honesty.  How long before that cow runs dry?

  4. Herbert Woodbery

    I disagree raycom,   It’s not spending per se that is the main problem.  It’s spending what you don’t have that caused the Greece fiasco and will eventually cause financial issues in the u.s.  If it continues.

  5. drlorentz

    Much as I would like to agree with the conclusions, one important distinction is glossed over:

    • The analysis conflates marginal and effective tax rates. The overall tax rate is t, but people respond to marginal tax rates. For example, Mr. Romney’s effective tax rate may be 13%, but I’m quite sure his marginal rate is the top one – closer to 40% federal plus whatever his state income tax rate is.

    This simple analysis only works when there is a perfectly flat tax with no deductions or exemptions. Beyond this, I question whether the curve is necessarily constant over time. The location of the maximum likely changes as a society becomes wealthier: the less you have, the more painful a 33% rate is. We have, on the average, become significantly wealthier over the last 30 years. The curve may also have hysteresis; the shape might depend on whether the tax rate is rising or falling. I suspect it does.

    One last concern: if this applies to effective federal tax rates, aren’t we below 33%? If so, then isn’t this an argument for raising taxes?

    I’m sure these thoughts aren’t original, but they deserve responses.

  6. Rodin

    The importance of the “tax as percentage of GDP” is that it gets at the layering of taxes. ( This is like fat within the muscle tissue, not just under the skin. ) It demonstrates the drag affect of taxes as a cumulation of millions of transactions. Converting that to the expected behavior of any individual taxpayer is problematic.  The question is when is a taxpayer group large enough that a “hump” can be calculated for that group alone?

    The argument against raising taxes is that taxpayers have feet. The GDP measurement reflects that the country as a whole does not have feet. Similarly individual taxpayers only have feet if, considering all other factors, they have someplace to go. 

  7. BrentB67
    Great post.  I think the idea that taxes are below 33% should be re-visited. Federal rates are below that, but add on state income taxes, sales tax, property tax, utility/communication/fees, gasoline taxes, assorted fees and levies and the ultimate tax rate is probably north of 50% depending on where you live.

    Academics do agree that the hump point in the Laffer curve is high, but those of us in the real world earning a living know that it is much lower. 

  8. drlorentz

    The weak link in the argument is the extrapolation of effective rate to marginal rates for individuals at various income levels. Though I’m no friend of ‘progressive’ tax rates, I do recognize that the more you have the less of a sacrifice a 33% rate is. Taxes are like a luxury good, except that they’re compulsory.

    To make this case, you’d have to analyze the effect of marginal tax rates on taxpayers in various brackets. I’d be surprised if this hasn’t already been done. Mr. Groseclose, what is the state of knowledge about this?

    Edit: The argument also presupposes that the only reason people work is for money. If you like your work, you’d be inclined to work more at a higher marginal tax rate than a person who hates his job. As jobs have become less arduous and more satisfying, the hump of the curve shifts up. In short, the disutility of work is not constant.

  9. Fake John Galt

    The purpose of government is not to maximize its tax revenues.  The purpose of government is to maximize the prosperity of its citizens. 

  10. drlorentz
    Fake John Galt: The purpose of government is not to maximize its tax revenues.  The purpose of government is to maximize the prosperity of its citizens.  · 0 minutes ago

    Agreed. The Laffer Curve argument is a pragmatic, not a moral one. The economic conservative view is better served by moral defenses of capitalism.

  11. Larry3435

    “As I mentioned, all this should be big news with the U.S. media, but it has not been.  Although I suspect that the reason is that liberal journalists want to squelch the Romer-Romer results and hide them from the public, perhaps the true reason is more innocent:  Perhaps U.S. journalists simply never learned about the Romer-Romer results.”

    Perhaps the true reason is more innocent yet.  Perhaps U.S. journalists simply never learned about economics.

  12. Larry3435
    Herbert Woodbery: Most high income people pay lower rates than middle/upper middle  income people.  That’s why the buffet rule has such appeal.    Mitt claims to have paid at least 13 percent ,  a long way from laffer curve influence. · 2 hours ago

    Seriously?  On what planet?  Only high income people who are also retired and living on capital gains off their savings pay lower rates.  Of course, they already paid top rates on that money when they saved it in the first place.

  13. raycon and lindacon
    Herbert Woodbery: I disagree raycom,   It’s not spending per se that is the main problem.  It’s spending what you don’t have that caused the Greece fiasco and will eventually cause financial issues in the u.s.  If it continues. · 1 hour ago

    On what basis do we disagree????

  14. Vectorman
    BrentB67: Great post.  I think the idea that taxes are below 33% should be re-visited. Federal rates are below that, but add on state income taxes, sales tax, property tax, utility/communication/fees, gasoline taxes, assorted fees and levies and the ultimate tax rate is probably north of 50% depending on where you live.

    The ancient Israelites set the tax rate (tithe) at 10%.  Recent surveys (Reader’s Digest) say that the maximum rate should be 25% in all (Federal,   State, and Local) taxes.  Even at 19%, Fed taxes are still too high.   To get to an average 19% in a progressive system, the upper marginal rate needs to be at least 33% with a base rate near 12%, which is close to the present social security (both halves) tax.  Until Fed spending gets below 15%, you cannot approach a fair tax system.

  15. Essgee

    Liberal response:  Are you attempting to confuse us with facts?

  16. Joseph Eagar

    I suspect Romer was making a Keynesian point.  Remember that supply-side economics concerns itself with the growth rate of potential output, not GDP per se.  One-time falls in GDP do not reduce potential growth in the future.

  17. ParisParamus

    Are such ‘impositions’ as fees and sales taxes considered in these studies/calculations?

  18. Steven Jones
    drlorentz: The weak link in the argument is the extrapolation of effective rate to marginal rates for individuals at various income levels. Though I’m no friend of ‘progressive’ tax rates, I do recognize that the more you have the less of a sacrifice a 33% rate is. Taxes are like a luxury good, except that they’re compulsory.

    · 4 hours ago

    Edited 4 hours ago

    A fair assertion. Bear in mind, however: Very few individuals willingly pay more than required by law – and the more you have, the more leeway in finding ways to shelter your earnings from rapacious taxation.

  19. Steven Jones

    I perceive a fundamental difference between conservative and progressive economic theory.

    Conservative economists observe the way systems actually function, and formulate theories to explain them.

    Progressive economists visualize systems as they believe they should be, and formulate theories to arrive at the desired result. Human nature, however, invariably declines to cooperate with the the best-laid theories.

  20. Larry3435
    Essgee: Liberal response:  Are you attempting to confuse us with facts? · 2 hours ago

    Nah, the liberal response is:  ”You are a stupid, racist Repugnantant who got all that stuff off of Faux News and say whatever the Koch brothers tell you to say.”  Trust me.  I live in California.