Sunday, November 28, 2010
WSJ: Tax Revenues = 19% of GDP, Regardless of Tax Rates
Tax revenues as a share of GDP have averaged just under 19%, whether tax rates are cut or raised. Better to cut rates and get 19% of a larger pie.
Even amoebas learn by trial and error, but some economists and politicians do not. The Obama administration's budget projections claim that raising taxes on the top 2% of taxpayers, those individuals earning more than $200,000 and couples earning $250,000 or more, will increase revenues to the U.S. Treasury. The empirical evidence suggests otherwise. None of the personal income tax or capital gains tax increases enacted in the post-World War II period has raised the projected tax revenues.
Over the past six decades, tax revenues as a percentage of GDP have averaged just under 19% regardless of the top marginal personal income tax rate. The top marginal rate has been as high as 92% (1952-53) and as low as 28% (1988-90). This observation was first reported in an op-ed I wrote for this newspaper in March 1993. A wit later dubbed this "Hauser's Law."
Over this period there have been more than 30 major changes in the tax code including personal income tax rates, corporate tax rates, capital gains taxes, dividend taxes, investment tax credits, depreciation schedules, Social Security taxes, and the number of tax brackets among others. Yet during this period, federal government tax collections as a share of GDP have moved within a narrow band of just under 19% of GDP.
Why? Higher taxes discourage the "animal spirits" of entrepreneurship. When tax rates are raised, taxpayers are encouraged to shift, hide and underreport income. Taxpayers divert their effort from pro-growth productive investments to seeking tax shelters, tax havens and tax exempt investments. This behavior tends to dampen economic growth and job creation. Lower taxes increase the incentives to work, produce, save and invest, thereby encouraging capital formation and jobs. Taxpayers have less incentive to shelter and shift income.
Here are the data from American Thinker:
https://taxprof.typepad.com/taxprof_blog/2010/11/wsj-hausers-law.html
Comments
I want my economic analysis from people like myself who are educated, trained and experienced in the field,.
I chuckled over that. I'll take Margeret Thatcher's good sense, sound judgment and accomplishmnets over teh best and brightest economists, like, say, Implications of the New Fannie Mae and Freddie Mac Risk-Based Capital Standard, 2002, Joseph E. Stiglitz, Jonathan M. Orszag and Peter R. Orszag:
The paper concludes that the probability of default by the GSEs is extremely small. Given this, the expected monetary costs of exposure to GSE insolvency are relatively small — even given very large levels of outstanding GSE debt and even assuming that the government would bear the cost of all GSE debt in the case of insolvency. For example, if the probability of the stress test conditions occurring is less than one in 500,000, and if the GSEs hold sufficient capital to withstand the stress test, the implication is that the expected cost to the government of providing an explicit government guarantee on $1 trillion in GSE debt is less than $2 million.
Posted by: guy in the veal calf office | Dec 1, 2010 10:39:23 AM
I'm surprised nobody has done the basic research on this. It is explained very succinctly here
Basically, if you take out social security taxes which have increased by a huge amount in the last 50 years, you see a decline in tax revenues just as you would expect. So this graph really is like comparing taxes and storks, there is alot more going on than this lets on.
Posted by: Paul | Dec 1, 2010 10:27:17 AM
Sid, My take away from the article and the comments that follow is that all the other variables that are not accounted for in the graph are so important that they overwhelm any change in Income tax receipts that a change in the top marginal rate makes. In other words, raise the rates if you wish, but you will see a slowing of productive economic activity that depresses other sources of revenue. Or, lower the rates and see an increase in productive economic activity that will increase other sources of revenue. It is precisely the complexity of a system as large as this economy, all the uncontrolled variables, that makes the prediction "increasing the top bracket income tax rates will increase government revenue" so hard to correlate to actual historical data. Without a strong correlation from history it seems foolhardy to raise the top income tax rates in an effort to increase total revenues from all sources.
Posted by: Nathan | Nov 30, 2010 10:07:10 PM
Looks like a pretty good proof that the Laffer curve is mainly true. Not completely true, since lowering rates does not bring in more revenue, but it also clearly does not bring in less. And since it is clear that lower rates help economic growth, then lower rates are obviously good. They cost nothing in revenue, and help growth. Mind you, there might be a low point where further lower rates do cost revenue, but the low 30% rate in 1989 did not reach that point, so tax rates could at least be lowered to that level, with no cost at all in revenue. So dem claims that renewing the Bush tax cuts will cost revenue are clearly disproved.
It is also pretty clear from this graph that any gov plan for sustainable deficits must cap total spending at less than 19-20% of GDP, since higher taxes will not bring in more revenue than that, just harm growth. Reducing spending is the onoly answer to the deficit, tax hikes do nothing but harm growth.
Posted by: richard40 | Nov 30, 2010 3:37:17 PM
Per Sid: "First of all, lets use income tax revenues rather than all revenues as a percent of GDP, since we are talking about income tax rates. Ok, makes more sense. (The conclusions below do not change if we use total revenues and not income tax revenues, just trying to get a little more integrity in the analysis)"
No, no increase in integrity. Just a different question.
If we take Hauser and you at your word, then there were offsetting tax revenues that changed total receipts.
"To quote your post, I think this proves that receipts go up when the top marginal rate is raised and that they go down when the top marginal rate is lowered."
Only in the revised context you have presented. Receipts (ie the total amount of revenue) stay the same as a percent of GDP. Since revenue normalized to GDP stays constant, a change in GDP is the only way to increase or decrease absolute revenue.
There is only one situation that is consistent with both you and Hauser: the relative contributions of income tax vs. other taxes changes when the top rate changes. This would be consistent with the top rate payers acting to reduce their capital gains exposure, for example, in the face of higher income taxes. When confronted with confiscatory tax rates, people find ways - legal or not - to shelter their income.
It's a purely practical question. Taxation requires the consent of the taxed, and this makes the maximum percentage that can be peaceably extracted from the citizens self-limiting.
We could ignore history, or we could believe that human nature has changed. If not, then we must agree that raising the top rate will do nothing to erase the deficit.
Posted by: timgr | Nov 30, 2010 12:02:05 PM
In re Ken Royall: "WHATEVER the government has done to tweak our ridiculous tax code, we end up with 19% of GDP in revenues"
It is worse than that. Whatever the govt does, THEY end up with 19% of GDP in revenues. What the private sector ends up with (i.e. did the changing tax policies lead to increased income (good) or increased wealth protection / tax avoidance (bad)) is another question altogether, and is not addressed here.
Posted by: Troll Feeder | Nov 30, 2010 5:31:20 AM
In re Sid: "Am I the only one who understands that simple(one dependent variable, one independent variable) models as opposed to multi-variate, multi-equation models are pretty much useless and represent the kindergarten level of empirical analysis?"
Yes. It's not us, Sid; it's you.
Do you feel vindictivicated now?
Posted by: Troll Feeder | Nov 30, 2010 5:23:49 AM
Those who are trying to dispute the conclusions of Hauser seem to be ignoring the obvious in an effort to cloud the issue. If you acknowledge that government realizes about 19% of GDP in revenues historically, the case is already made. Effective tax rates up and down the income scale really don't matter. NOTHING matters, that is the entire point. The answer is simple, WHATEVER the government has done to tweak our ridiculous tax code, we end up with 19% of GDP in revenues. Whatever they do in the future will yield the same result, end of story.
Of course higher taxes discourage economic activity and capital formation, to deny that is sheer idiocy. What you penalize you get less of, this is a very basic principal of economics. If the private sector is robbed to fund government waste, investors have less funds to invest and less incentive to do so. This is an indisputable fact, especially at the margins. When evaluating an investment, the amount you give up in taxes will have a direct impact on whether you will risk the capital and effort to go forward. Better that we promote growth through lower taxes and increase the size of the economic pie.
Posted by: Ken Royall | Nov 29, 2010 8:33:50 PM
Ok Timgr, let try it on your terms, a two variable model.
First of all, lets use income tax revenues rather than all revenues as a percent of GDP, since we are talking about income tax rates. Ok, makes more sense. (The conclusions below do not change if we use total revenues and not income tax revenues, just trying to get a little more integrity in the analysis)
In the first year of the Clinton administration, 1993 the top marginal tax rate was increased. Income tax revenues as a percent of GDP went from 7.7% of GDP in 1993 to 10.2% in 2000. This is about a 30% increase.
In 2001 the top marginal rate was decreased. Income tax revenues as a percent of GDP went from 10.2% in 2000 to 7.5% in 2005. This is about a 25% decrease(source: 2011 Federal Budget, Historical Tables).
To quote your post, I think this proves that receipts go up when the top marginal rate is raised and that they go down when the top marginal rate is lowered. So I can from the data, if you allow other variables are irrelevant, which you do.
Personally I would reject this type of casual empiricism for the same reason I reject the original article. At least two other variable are in play here, what happened to rates below the top rate and what happened to the distribution of income. However, if you want to play by your rules, I kinda think you have to accept the outcome even if you disagree with its policy implications.
Posted by: Sid | Nov 29, 2010 6:47:58 PM
Socialists would rather have 100% of nothing than 10% of something big.
Posted by: Eagle | Nov 29, 2010 6:09:34 PM
Why can't we just experiment with zero capital gains taxes for 5 years, and see what happens.
That would be a better experiment to run than any other.
Posted by: Eagle | Nov 29, 2010 2:35:02 PM
Sid, it appears your inclination for more taxation is skewing your ability to read what Hauser, JFK and others are saying. The gov't revenue as 19% of GDP is not a theory, it is straightforward empirical fact. The chart disproves (the null hypothesis) the theory that higher tax rates produce more government revenue. Of course there are many more factors than just tax rates that determine the reason gov't revenue has been more or less a constant percentage of GDP since WWII. That is the point conservatives generally try to point out but that liberals ignore.
Posted by: TexEcon | Nov 29, 2010 2:17:50 PM
Sid again "Until and unless these other variables can be accounted for the presence of zero correlation between the two variables is just that, zero correlation and that has no relevance to the hypothesis".
Not true. You're turning the hypothesis on its head, or willfully misinterpreting the question. Prove that receipts go up (down) when the top tax rate is raised (lowered). You can't from the data. Other variables are irrelevant.
"I think from a methodological point of view no serious academic or highly educated professional would disagree that the methodology is insufficiently rigorous to support the conclusion."
Wrong again. Wow, I wonder what your mental picture of me and the rest of the discussion participants is. Credentials at 20 paces. The conclusion is not rigorous enough for some disciplines, but in the context of my dabbling interest in economics, the data is very convincing.
Posted by: timgr | Nov 29, 2010 2:10:36 PM
I reeeaaaaallly want globalization to enter a phase where nations have to compete on capital friendliness, which forces nations to engage in a race to zero on the capital gains tax rates.
That will force all major economies to a 0% capital gains tax rate (as they should be).
Posted by: Eagle | Nov 29, 2010 12:19:18 PM
Wow, interesting comments and not as vindictive as I was expecting.
Let’s see. The hypothesis of the article seems to be that revenue as a percent of GDP is independent of the top marginal tax bracket, and that a lack of correlation of these two variables over an 80 year period is proof, or at least evidence that this is the case.
My point is that over that time frame there have been a huge number of other events which influence revenue as a percent of GDP, including the tax system itself (size of brackets, deductions, capital gains rates etc.), inflation and the brackets, the government policy (as JFK alluded to), economic growth, demographics, and on and on. Until and unless these other variables can be accounted for the presence of zero correlation between the two variables is just that, zero correlation and that has no relevance to the hypothesis. I think from a methodological point of view no serious academic or highly educated professional would disagree that the methodology is insufficiently rigorous to support the conclusion. Idealogues, however, with pre-formed opinions independent of facts or analysis would have no trouble with the methodology.
Now to address some specific points:
That you can jack the top marginal tax rate up and down and not change tax revenues as a percent of GDP, all other things being constant, seems just plain silly. I know of no laws of economics or finance or public policy that says this, but perhaps some respondent to this Blog can point out a legitimate academic or professional study using accepted methodology that show revenues must be 19% of GDP regardless of the top marginal rate (Please no WSJ or Forbes articles)
I don’t really want economic analysis from a person with the background, education and experience like Margaret Thatcher, despite her other accomplishments. I want my economic analysis from people like myself who are educated, trained and experienced in the field, just as I want my medical diagnosis from trained physicians, my legal opinion from lawyers, and my bridge building advice from civil engineers. Multiple correlation analysis is not a new mthodology, it is decade old, generally accepted and widely used.
For those of you who commented that lack of correlation was determinant, I would say the following. Lack of simple correlation between two variables, all else constant certainly implies independence; lack of correlation between two variables, other things not constant, implies, well implies nothing.
I disparage the opinion of the article because it is based on flawed analysis. However, note that I said I have an open mind, that the conclusion may or may not be valid, but that it could not be supported by the simplistic data analysis. However, let’s recognize reality. The editorial pages of the WSJ have an agenda, and it is not even a hidden agenda, and they print articles that support that agenda as they have every right to do. We, however, have every right to demand the analysis to be somewhat rigorous before we support the conclusions.
Posted by: Sid | Nov 29, 2010 12:17:53 PM
So from this chart, we can conclude :
A simple reform of :
a) Tax simplification, so that the billions of hours spent by taxpayers and the IRS on processing and compliance are no longer wasted..
b) Elimination of capital gains tax..
Will quickly increase the size of the US economy by 10%.
So what are we waiting for?
Posted by: Eagle | Nov 29, 2010 12:10:37 PM
This pretty much shows that the 'tax the rich' crowd is more driven by envy, rather than any basis in sound economic policy.
If the US were to take the bold action of making capital gains taxes the same as China/India/Russia (i.e., 0%), then the US economy would very quickly surge in size, and tax revenues would INCREASE.
Millions of jobs would be created.
Posted by: Eagle | Nov 29, 2010 12:03:23 PM
Hey, I got laid off in 2009 and haven’t had a job since. Let’s just confiscate all the money of anyone who makes over 100k a year and redistribute it to poor folks like me. My computer is out of date and I need an upgrade.
Posted by: Low On Prozac | Nov 29, 2010 11:57:32 AM
I agree.
Russia, China, and India all have a 0% interest rate on long-term capital gains. ST gains also vary from 0% to 15% among these three countries.
Wait, I thought we won the Cold War....
....or is that the wrong statement, with a more accurate statement being "where has the traveling disease of socialism traveled to now?".
Posted by: Eagle | Nov 29, 2010 11:39:54 AM
Paleveda, I think Hauser is actually arguing the contrapositive, namely that raising tax rates on the wealthy is unlikely to solve our economic woes. However, assuming the string of letters after your name is indicative of career achievement as well...you could always help to prove Hauser wrong by giving the government an extra 8-10% of your income in 2010...I mean, if you think raising the tax rates on the wealthy is both morally and financially sound policy, why wait until next year when the rates go up? Do your part now and stop being selfish before the other wealthy people stampede towards virtue.
Posted by: JFK | Nov 29, 2010 10:57:56 AM
Quoth Nick "I love the argument that if rates go up rich people will cheat to keep more. What about poor people where the funds mean whether or not they can feed their familes as opposed to whether they should settle for a Mercedes as opposed to a Rolls".
So your solution is to impose a feel-good punitive tax on the wealthy that will make everyone poorer? Something about two wrongs...
Posted by: timgr | Nov 29, 2010 10:02:00 AM
"Idealogues with a particular agenda focus on marginal tax rates, because they are higher than effective tax rates and give the distorted implication that taxes in the U. S. are higher than they actually are. The average tax rate for the top .1% in 2007 was about 21.5% (IRS, SOI Table 5). A typical couple with $1 million in income has an effective tax rate of 26% in 2010, a typical retired couple with $60,000 of income has an effective tax rate of 1.3% (Tax Foundation, Fiscal Fact 251)."
You are, perhaps unwittingly, explaining the entire point of the article; push up tax rates and you will cause more activity to avoid the taxman. Funny how your 21.5% effective top tax rate is close as spit to the 19% average stated in the article.
Posted by: CR | Nov 29, 2010 9:22:33 AM
Sid, you want to have your cake and eat it too. You point to a bad single variant example as evidence that the single variant data in the graph is similarly flawed. What you neglect is that the graph demonstrates a lack of correlation, as opposed to the stork-baby example you give. That's a very different thing. Only very rarely is a _lack_ of correlation hiding a causal relationship, however complicated.
The only way you can get the result in the graph is by having some other unaccounted-for variable that exactly compensates for the lack of correlation the graph expresses. Statistically, that's a much higher hurdle than finding an unexplained correlation. I'm sure you understand that in history and economics there is no opportunity to do a controlled study... however, imagine that you were doing a controlled study of this question, and you carried out the experimental part of the experiment first. You got the results above. Would you need to complete the control to demonstrate that there was also no correlation in the control group?
It seems that you disparage the opinion content of the linked article more than the history that it tries to explain. I'm sure that if you include enough variables in your model, that you can find a causal path that does not include the "animal spirits" of entrepreneurship that Hauser points to. However, that does not diminish two significant assertions 1) that Hauser's law is well supported by the historical evidence, and that the null hypothesis (changing the top tax rate _does_not_ significantly affect federal revenues) is supported, and 2) that in light of this history, in the context of the current debate over raising the top rates back to pre-Bush levels, motivating these tax increases with the expectation of raising revenue, is a specious argument.
Posted by: timgr | Nov 29, 2010 8:57:04 AM
Interesting argument. Laffer and now Hauser are attempting to make a name for themselves in the guise of lower taxes for the wealthy will solve all our economic problems. Why not just come out and say it! I have Paleveda's law. "Lower taxes for wealthy individuals=solution to economic problems." It is a proven fact that wealthy people are more generous and give more to worthy causes,cite the Gates foundation. Also the study is flawed if they believe the highest marginalr rates were in 52-53 when the highest rates occurred in 1944-45 of 94%.
I love the argument that if rates go up rich people will cheat to keep more. What about poor people where the funds mean whether or not they can feed their familes as opposed to whether they should settle for a Mercedes as opposed to a Rolls. Finally, they invoke the "Standford" name. Is this "Stanford University" or the Standford Financial which is involved in embezzlement?
Posted by: Nick Paleveda MBA J.D. LL.M | Nov 29, 2010 8:48:03 AM
Margaret Thatcher used to drive people like Sid wild with rage. She would explain world economics in terms that housewives understood in totting up their household budgets. The Sids of the world insist that it's really very complex and takes intellectual heft and statistical analysis to comprehend.
Tax a thing, you get less of it. Subsidize a thing, you get more of it. It takes years of indoctrination to lose sight of those facts.
Posted by: S. Weasel | Nov 29, 2010 8:35:57 AM
This makes all the sense in the world to me. Like Paul, though, I would be interested to see something that shows the correlation between top marginal rates and GDP growth.
Posted by: Matt Tanner | Nov 29, 2010 8:26:58 AM
Ummm, Sid, the proposition under discussion is a "simple model". The proposition is, "If we raise tax rates, we will raise tax revenues." Tax rates going up is nigh on a certainty at this point...the question is will that also increase tax revenues. Perhaps you missed the author's point...which was that this particular simple model doesn't work...that you can jack the tax rate both up and down and the fraction of the GDP collected by the government doesn't change all that much...that, perhaps the government has to do something else to increase revenues like set a tax rate AND change some policies to increase growth in the GDP (which would be the multi-variate model you mentioned...after the author alluded to it).
Posted by: JFK | Nov 29, 2010 8:25:43 AM
Jimmy/Jimmy/Jimmy/Jimmy -
One huge fallacy in your string of losers, the government. As we have all seen, the government will continue to thrive and flourish. Salaries, benefits, sick days, vacations, retirements continue to go on and on and on even if tax receipts are down. The government (we) just borrow money from the other countries of the world to afford a nice lifestyle for the public employees.
Posted by: DeductionSeeker | Nov 29, 2010 8:23:32 AM
"When tax rates are raised, taxpayers are encouraged to shift, hide and underreport income."
Actually it's less nefarious than this. When 1/2 of the second earner's income in a professional couple's earnings goes to taxes of one form or another there becomes much leverage in playing defense with the family budget rather than offense. They fire the babysitter, the house-keeper... they clip coupons and eat in more... they pack their spouse's lunch and wash their own car rather than opt for detailing at the local car wash... they do their own car repairs instead of hiring them done... they build their own deck and do other house repairs in work that should go to unskilled and semi-skilled workers.
This is pent-up productivity what wastes the scarce resources of an educated society reducing the key component of specialization, both in skills and in wasteful redundancy of the tools of trade.
When you have to earn $1.50 for every $1.00 you spend (especially in the 2nd income professional households) it is unreasonable to expect people not to choose the leverage of thrift over the tax penalty of work to make their lives better. Who loses? Well… the housekeeper… the servers in the restaurants that are no longer frequented… the babysitter whose services are no longer needed… the government who no longer receives a portion of the higher earners income (and lower earners alike) because it over-reached.
As is almost always the case with well intentioned policies of redistribution… the cure becomes the cause. In this case it is the cause of further class separation and a dampened economic system that hurts the people it purports to help the most.
Posted by: JimmyNashville | Nov 29, 2010 7:54:24 AM
Am I the only one who understands that simple(one dependent variable, one independent variable) models as opposed to multi-variate, multi-equation models are pretty much useless and represent the kindergarten level of empirical analysis?
My point was that the simple correlation analysis between two variables in a very complex world is pretty much meaningless, and valid only if all other potential independent variables are held constant, which I don't think has happened over the 80 year period of the graph.
Idealogues with a particular agenda focus on marginal tax rates, because they are higher than effective tax rates and give the distorted implication that taxes in the U. S. are higher than they actually are. The average tax rate for the top .1% in 2007 was about 21.5% (IRS, SOI Table 5). A typical couple with $1 million in income has an effective tax rate of 26% in 2010, a typical retired couple with $60,000 of income has an effective tax rate of 1.3% (Tax Foundation, Fiscal Fact 251). Not the stuff of revolutions, is it?
My point: The conclusions of the article may or may not be correct, but they cannot be supported by the simple chart. Sorry, I did not think I would have had to explain all this to the readers of this Blog, whom I assumed were fairly familiar with the concepts of statistical modeling.
Posted by: Sid | Nov 29, 2010 7:53:54 AM
So, how about a Fair Tax with a 19% tax rate? Does that sound crazy?
Posted by: Matt Groom | Nov 29, 2010 7:49:01 AM
Is there similar information available on other industrialized nations?
Posted by: Mike Landfair | Nov 29, 2010 5:55:34 AM
Make the top rates high enough, and the rich will simply leave. They're rich. They can afford to.
This is why Exile on Main Street was made in France.
Posted by: S. Weasel | Nov 29, 2010 5:42:20 AM
Quoth Sid "The article reminded me of a famous research project that found that the birth rate was highly correlated with the number of storks in an area."
Yes, one could explain the correlation of births-to-storks with the type of houses or communities the storks like to nest on or in. Not many births in industrial areas, for example. Correlation is a necessary but not sufficient condition for causation.
In this case, we see a distinct lack of correlation between top tax rates and tax revenues. This is a much more powerful result than an unexplained correlation between possibly unrelated data. Here the data proves the null hypothesis (ie, tax rates do not affect revenues), which must be accepted lacking an alternative.
Posted by: timgr | Nov 29, 2010 5:17:07 AM
The chart gives gross revenues regardless of source and base thereby resulting in a simplification. As income tax rates were dropped in the mid 1980s the base was broadened under the 1986 Act. That is a key missing ingredient. Also, let us not forget that as rates were dropped at the top, wage earners took a big hit as FICA was raised at about the same time, thereby raising more tax from wage earners.
Posted by: Bill | Dec 1, 2010 6:44:19 PM