Whaddya expect from the cocktail napkin school of economics?

Oh. My. God.

My biggest problem is which of the "Planet's Funniest Statistics" blogs I read to post this to first.

"Why their readership puts up with this obvious contempt for their intelligence is a question for another day."

It would be more accurate to describe it as contempt for their innumeracy, and this has the virtue of making it easier to understand why their readership puts up with it: they're actually too goddamned innumerate to realize they're being snowed.

"Why their readership puts up with this obvious contempt for their intelligence is a question for another day."

Why wait? It's because people like to be told that they are OK, that their way of life is the "right" one, and that they don't have to give up anything to be noble and "good citizens." By telling readers in effect that grabbing all you can and keeping it is "good policy" for the country, publications like these are massaging the egos of their readers. That's why the readers don't worry that they're being played for patsies -- they assume that it's the rest of us being played, and that's all right. If we were noble and good, we'd be rich, too.*

Simple, isn't it (and aren't they?).

Ed

* Yes, I know they're not rich (most of them), but they think they're in that upper class, where the truly noble and honest and good folks belong, and don't notice the picking of their pockets that the purveyors of such talking points perform. By telling folks what they want to hear, you keep them from asking embarrassing questions.

When you already have the answer, you need to draw some crazy-ass curves sometimes.

Whoever drew that lulu has a desk waiting for them in the OMB.

Ugh, you're utterly right Kevin. This is atrocious.

Apparently nobody there took any statistics about linear regression.

I think it's just echo-chamber effect.

It says what they want to hear, so the fact that the underlying "reasoning" is a bit specious is glossed over.

Commenter "theo" over at Mark Thoma's site says what I was about to say better than I was going to say it:

Did Kevin Hassett draw that curve using Microsoft Paint?

Even if you had extraordinarily strong a priori reasons to fit a convex down curve from that exact family (what are they, log parabolas?), you would NEVER get the curve shown as a least mean squares fit. Especially if you used robust regression, as you almost certainly should for a dataset with such outliers.

I would say that Hassett has embarrassed himself, but he works for AEI, so he's clearly incapable of embarrassment.

Effective rate, effective rate, effective rate.

Nominal rates are like the list price at a car dealer (Big 3, at least - imports don't play the incentive game as much).

Funny how you left this line out of Mark's "more plausible" line:

"I haven't actually run the regression"

In other words, he chose his arbitrary line because it better fit his argument than the arbitrary line he was criticizing.

What we can deduce from the chart is that it is possible that lower tax rates can coincide with higher revenue percentages. And that the U.S. with one of the highest corporate tax rates ranks among the lowest in terms of revenue percentages.

Mark criticizes the inclusion of Norway in the WSJ's graph as an outlying situation. However, he has no problem using UAE (with a 0% tax rate) as part of the data set for his "best fitting line." A more fair analysis, knocking out both Norway and UAE, would show that the highest revenue percentages are more likely to occur with tax rates of between 20 and 30%. Countries with tax rates higher than that (i.e. the US) see lower revenue percentages and might increase those percentages if they moved to a lower tax rate.

Finally, Mark criticizes the Laffer curve claiming that his own curve shows "revenues rise with tax rates." But the Laffer curve does not conflict with that. It only suggests that after a certain point, increasing tax rates will actual lead to a decline in revenues. Looking at the data in the chart, that certainly appears to be the case.

the fact that they publish an editorial page five times a week shows they have a high tolerancef or embarassment. after hearing their criticism of the scooter libby commutation (bush too wimpy, should have given him a pardon) i came to the conclusion intellectually dishonesty is a requirement for the board.

It would be more accurate to describe it as contempt for their innumeracy, and this has the virtue of making it easier to understand why their readership puts up with it: they're actually too goddamned innumerate to realize they're being snowed.

Considering the notional target audience of the Wall Street Journal (that is, investors), that's something of a frightening thought, though I guess it would explain a lot about corporate America.

But then, since it's the only way to get an upside-down U out of the data, the graph goes nearly vertical.

There's another constraint the WSJ folks are working under: the data point for the U.S. has to be above the curve on the right hand side, so that they can argue that reducing corporate tax rates will increase corporate tax revenue. If it were below the curve, then it would be an argument that we could raise rates and revenues.

That is one hilarious graph. If it's continuous for positive tax rates, and if it follows that steeply descending curve on the right, it suggests that revenue is predicted to be negative for rates above 35% or so.

Hacksaw,
Wait a minute, you are criticizing Mark for not running the regression, then you come up with your own arbitrary line without running any regression?

Pot, meet Kettle.

Amateurs. When Murdock takes over, they'll publish an elipse.

Hacksaw -- didn't you observe that the line from the
UAE to Norway does not have any other observations near it.

Moreover, Hacksaw the line you describe does not fit your conclusions. Try it and see.

Mark criticizes the inclusion of Norway in the WSJ's graph as an outlying situation. However, he has no problem using UAE (with a 0% tax rate) as part of the data set for his "best fitting line."

Um, there's a good pragmatic reason for this. If the y-intercept were above zero for UAE, it would mean the model predicts ceteris paribus that UAE is collecting tax revenue on a 0% tax rate; if it were below zero, that UAE is giving out tax refunds based on no revenue collected.

Very entertaining -- I guess understanding how to best fit a function to a data set is no longer required to be a freakin' writer for a FINANCIAL NEWS ORGANIZATION!!!

Hello, Kevin is right. A straight line alone is a better fit than this dreck. Heck, any bozo could use the rudimentary curve fitting routine in Excel to identify what line would fit best.

Also, I wonder about those unlabeled data points. What countries are they? Me thinks I smell some heavy-handed data manipulation....

"Why their readership puts up with this obvious contempt for their intelligence is a question for another day."

Same reason our trolls keep blathering on about how the world will end if capital gains are taxed as ordinary income.

Ah, Kevin.

Whose to say what's an outlier? I think it's pretty arrogant for anyone to assume they can eyeball a chart and immediately pick out the outliers. If you really want to establish a reputation of integrity, you leave the outlier in and then explain the variance. Maybe Iceland is the outlier. Whose really to say?

I think it's pretty compelling and am excited about it. If I were the Republican candidate, I would start off every speech my just holding this graph up to the crowd for about a minute and just let it sink in. Powerful.

I just noticed something. This study demonstrates that we could conceivably cut all taxes to just 10%, and we'd STILL be taking in more tax receipts than we are today.

Like I said, powerful.

DR,

Well I certainly didn't do a regression analysis (not least because I don't have the source data) but that still doesn't mean that leaving in the UAE is as much a skewing factor as using Norway. Moreover, I didn't propose a line but did suggest that when you look at the data without Norway and the UAE you can draw a reasonable conclusion that the sweet spot for a tax rate that maximizes revenues falls between 20 and 30% and that rates above and below this range lead to lower revenues.

Spencer, the curve I proposed in my analysis very much fits the data. The curve would start of similar to (if flatter than) Mark's line but would peak around 30% and then decline. Even if you throw out all the countries with revenues above 4% GDP (i.e. get rid of the data that best supports my conclusions) you still see a concentration of the highest revenue percentages at tax rates in this case between 25 and 30%. Yes there are some up at 35% (above France on the graph) but their revenues are mostly below or equal to the results from the 25 to 30% group.

"Why their readership puts up with this obvious contempt for their intelligence is a question for another day."

I can answer this from my observations of my father in law. It is a restatement of an old quote (from Upton Sinclair IIRC): "it is hard to make a man realize facts if those facts would in any way interfere with his feeling good about paying lower taxes."

My FIL is a Wall Street Journal editorial page fan. He actually tried the "lucky ducky" argument on me once!

A junior high school geometry student would be embarrassed to produce work like this. But not the Wall Street Journal editorial page. Or the American Enterprise Institute, which created it in the first place. They apparently think their readers are too dumb to see what they're doing. Why their readership puts up with this obvious contempt for their intelligence is a question for another day.
—Kevin Drum

A friend reports for the WSJ from Tokyo. He says the paper has two categories of readers - business and economics professionals, who read the news and business reporting and features pages, mostly ignoring the op-ed section, and wingnuts who couldn't care less about business and economics reporting and only read the disconnected-from-reality op-ed pages. Only occasionally do the twains meet.

As for the American Enterprise Institute . . .

Bill Clinton is murderer! Bill Clinton is a drug dealer! Bill Clinton is a rapist!

We found the WMDs!

-

And one more thing. Just for laughs, take a look at what the Journal's barmy graph drawing implies: Norway, with a corporate tax rate of about 29%, generates enormous amounts of corporate tax revenue. But then, since it's the only way to get an upside-down U out of the data, the graph goes nearly vertical.

The odd thing is, it isn't the only way to get an upside-down U out of the data: if you just plot the numbers they have (I did as best I could in Excel just from the graph, the numbers aren't exact but close enough to get the shape of various trends), and use Excel to do a quadratic trend, you get a nice gentle parabola that peaks at a slightly lower tax rate than the stupid "draw an unjustified curve connecting the low and high revenue extremes and then drop like a bomb to fit a preconceived assumption" like Hassett uses, suggesting a "Laffer Curve" with the US still on the right side of the peak (though under the curve rather than outside of it). So Hasset didn't even have to be nearly this big of a hack to find a "Laffer Curve" with the US past the peak, using the data he did, it was gratuitous hackery.

Of course, I should note that the curve thus projected is not a great fit, and you get a better fit with a straight line much like the one Thoma provides.

And, for that matter, the phrase "Laffer Curve" has to be used in quotes above because the numbers being used aren't even relevant to try to find an actual Laffer Curve, since even if the Laffer Curve hypothesis were true (that higher general tax rates eventually discouraged economic activity so much as to reduce total revenue), you'd still expect tax share of GDP to increase monotonically with tax rates, you'd just expect the decline in per capita GDP with increasing tax rates to, beyond the Laffer "peak", be so sharp that the per capita revenue declined while the per GDP revenue continued to approach unity.

Norway's obviously an outlier. For good reasons, since the corporate taxes collected are from the (surprise) booming oil sector.

However, it looks to me like you could get a pretty good upside-U curve just by ignoring Norway and drawing along Luxembourg and Australia. The "peak" would still be at about the same range of rates, about 30%.

Egbert, I can't tell if you're serious: Whose to say what's an outlier? I think it's pretty arrogant for anyone to assume they can eyeball a chart and immediately pick out the outliers.
The curve as drawn is simply ridiculous. It's as if it were drawn as a joke by the editorial writers or their graphic artist. Take away the curve and the data is suggestive, though I wouldn't trust the source about tax rate numbers without additional fact checking.

I brought my friday brain to work today. Can someone seriously dumb down what this discussion is about? I hate not being able to follow issues that amke republicans look like idiots.

Well I think we should give them some credit. After close to thirty years of hearing about the Laffer curve, this is the first attempt I have ever seen to plot it to any actual numbers.

Homonym indeed.

There's another constraint the WSJ folks are working under: the data point for the U.S. has to be above the curve on the right hand side, so that they can argue that reducing corporate tax rates will increase corporate tax revenue.

Well, no. The US being under the curve (assuming the curve was appropriate) does not suggest that the US could increase revenue (or share of GDP collected as revenue, which isn't the same thing but what the chart actually reflects) by increasing taxes, nor does it being above the curve suggest that it couldn't. Those suggest, respectively, that the comparable curve for the US is above or below the general curve for reasons of some combination of extraneous factors. If they aren't policy factors, that doesn't mean they can be voluntarily changed, and if policy factors press the US down but other intrinsic factors push it up, the US could be above the curve and still be able to increase revenues while increasing taxes or keeping them constant.

OTOH, that's a comparatively subtle point that anyone gullible enough to believe in the legitimacy of the "Laffer Curve" shown on that graph probably wouldn't grasp, so you may have a point as far as the motives for drawing the curve where it is.

cmdicely,

You and I are saying roughly the same things with regard to what a reasonable analysis of the data would show ("nice gentle parabola that peaks at a slightly lower tax rate"). I am curious though why you think a straight line is a better analytical tool than a curve?

I was, like you, confused about the use of revenues as a % of GDP. What I came up with is that, since GDP measures economic activity within a country (as opposed to GNP measuring the country's income as it were), you might have a case where at higher tax rates you still see economic activity within the country (i.e. part of GDP) but the firms are taking tax-avoidance steps which could explain the lower revenue percentages.

I would also add that the Laffer curve did not only posit that "higher general tax rates eventually discouraged economic activity" it also said higher rates could lead to great tax-avoidance strategies, which would also match my previous point.

Flunked math, didn't you, Eggie.

IANAEconomist, nor a Statistician, but...

On what basis can we correlate income tax rates and total revenue? Does this graph take into account various tax deductions, and other sources of government revenue?

Put another way, how is this graph any different than a graph correlating deaths by drowning with ice cream sales?

Thanks for the correction, cmdicely; on reflection, I think you're right.

The clear conclusion that can be drawn from this cluster of data points is that it is time to nuke Iran.

The clear conclusion that can be drawn from this cluster of data points is that we're making progress in Iraq.

It looks like Kevin Hassett created it -- you know, Mr. Dow 36,000. That particular boo boo resulted from (among other things) double counting earnings in his valuation methodology. So math is clearly not his strong suit.

The people who are making this argument know that they are full of shit. And most of the people who repeat the argument also know the argument is full of shit. Nobody with even a tiny familiarity with statistics would draw their curve through the most convenient outlier.

As a professional mathematician, I have to say this curve is utterly embarrassing.

Egbert: Norway is an outlier. We could push through the stats and verify this or you could just use yer freakin' eyes. Indeed, the concept of "outlier" was defined just so that, when a person looks at a data set like this, he could say "Norway is an outlier".

I just noticed something. This study demonstrates that we could conceivably cut all taxes to just 10%, and we'd STILL be taking in more tax receipts than we are today.

Like I said, powerful.

Faith-based statistics gains a new acolyte.

Thoma's line is obviously correct. If a real academic, not in a wingnut welfare tank but a professor at a university, drew that curve instead of the line that Thoma drew, that professor's reputation woud be destroyed forever. That professor would be revealed as innumerate, not having even a basic understanding of math. And the fact is, no reputable journal -- a class whose members do not include The Wall Street Journal -- would ever publish anyone who did make such a claim. It couldn't possibly survive peer review.

Why do I get the feeling that the data points that are unlabelled are for the numbers from Middle Earth, Narnia, Atlantis, Utopia, Cimmeria, and Lilliput?

"Funny how you left this line out of Mark's "more plausible" line: 'I haven't actually run the regression'. In other words, he chose his arbitrary line because it better fit his argument than the arbitrary line he was criticizing."

Don't be an idiot. If you run a linear regression you are by definition, going to have a line fitted to those points. It isn't hard at all to eyeball such a line and get pretty close to what it really is. Thoma's line is good.

I wonder who did their regressions -- Charles Murray?

Their operation is, I think: there are enough mentalities, especially those who see what they want to for ideological purposes, that such editorials bump up just enough support to help The Cause. And that little bit may be just enough to tip a few percent of the population to say, voting Republican. Given the tightness of the electorate (?), it is worth spending a bit of their time on.

Junius:

Neither I nor Mark specified that it has to be a linear regression. In fact, a linear regression would be useless if one were trying to assess whether or not increasing tax rates could at some point lead to lower revenue percentages (since a straight line by definition could never bend to show this change).

So yes, Thoma's line is good if he was approximating the results of a linear regression analysis of the data, but it is also irrelevant since NO linear regression analysis could ever show if at some point higher tax rates lead to lower revenues.

However, if you want to play the game out, if you took a linear regression of the data for the tax rates higher than 20 or 25% (even excluding Norway), I bet your line would in fact end up headed down the higher the tax rate goes, in effect "demonstrating" that based on this data, there is an inverse relationship between higher tax rates and higher revenues, once you get above a a rate of 20 to 25%.

Mind you I'm not suggesting that is the sweet spot (I think it's around 30% looking at this data). But it is what your linear regression would probably look like.

IT'S NOT A CURVE. It's an AREA below a hypothetical curve based on a 10,000 variable equation and the data clearly shows that.

Revenues are not fixed to that curve like the Seattle Monorail. They can move all over the place. Thus you can cut taxes AND lose revenue.

Wait a sec. I actually read the WSJ article. Germany, who has nearly the same % of corporate tax rate as Norway (according to the graph) just lowered their corporate tax rate. And they celebrate that as a good thing. But Germany should have been on the peak of that curve, given that the y-axis is supposed to be the dependent variable. So I think by the reasoning of this curve, Germany just hurt itself, although it's already one of the lowest tax revenue countries on the graph itself.

Martin Gardner illustrated how the Laffer Curve really looks in Scientific American in 1981. You can see his graph here:

http://ase.tufts.edu/cogstud/graphics/tempfig4.gif

Egbert confused the y-axis with the x-axis [he's no Donald Duck in Mathamagic Land !]

Also, as Mr. Drum pointed out, the US effective corporate tax rate in America isn't 35%, it's about 26%. That would put the US to the "left" of Norway, so, by the "Laffer curve logic", the US would substantially increase tax revenue by increasing the rate to Norway's 29% !

oops, must make links...

Martin Gardner's Neo-Laffer Curve

Without North Sea oil Norway would be some pathetic replica of a Warsaw Pact country.

What makes this Laffer Curve hypothesis even stupider is that you can't possibly make any statement about the relationship between tax rates and revenue/GDP across countries without taking into account other variables that affect the business climate in each country.
If Hassett can call himself an economist, then I am a brain surgeon, an astrophysicist, and the King of France.

Why don't we ask Mr. Vitter about "trickle down" economics? Unless he swallows, that is.

Egbert for President!

If I were the Republican candidate, I would start off every speech my just holding this graph up to the crowd for about a minute and just let it sink in. Powerful.

Run, egbert. Please run. We've got your back.

(I get so rapturous at the thought of egbert running for Pres that my knees go all weak and rubbery. And why do I need to tell my fingers not to type "norbert" instead of "egbert"? What's up with that?)

"Without North Sea oil Norway would be some pathetic replica of a Warsaw Pact country.

Posted by: DaGall on July 13, 2007 at 4:39 PM"

And without water, Lake Superior would be a big, dry divot. Your point?

Hack,

Try switching to Christianity or Islam. If you are going to believe in nonsense, you might as well get the benefit of thinking you'll go to heaven for it.

It occurs to me that if you took the graph seriously, it would be the conservative argument for high corporate tax rates.

As I noted above, even if the trend curve it shows were justifiable, it doesn't show a "Laffer Curve", since it doesn't show revenues going down with higher tax rates. What it shows is tax burden—the overall percentage of the economy extracted by taxes—declining with higher corporate tax rates (past the Norway peak). So, if you want small government compared to the overall size of the economy, you can acheive that by raising corporate taxes.

The chart doesn't suggest whether this reduced tax burden is a result of lower revenues or just economic expansion, only if that, if we take it seriously as a cause-and-effect relationship, increasing corporate taxes from where the US is now should decrease the overall US tax burden.

And all good conservatives want a lower overall tax burden, economic expansion, and small government, don't they? So they should love higher corporate tax rates!

The Hassert Curve—its like the Laffer Curve, but with more laughs.

" if you took a linear regression of the data for the tax rates higher than 20 or 25% (even excluding Norway), I bet your line would in fact end up headed down the higher the tax rate goes"

I don't think so. There's necessarily a ceiling (at 100% certainly), but based on that scatterplot no one would think you had a curve there if they weren't already convinced that there should be one. The natural interpretation is that you do in fact have a linear relationship at least within the range of data depicted on that graph. There's no sign of a downturn anywhere in that neighborhood. As rates go up, so does revenue.

We can hypothesize that it curves downward at higher rates but from the looks of that graph it might be at a very high tax rate in the 70's, 80's, wherever. Simply put, these empirical data show that within the range of actual tax rates, revenue goes up when the rate goes up.

I imagine these are totally valid comparisons because every tax office across the world is run just like the IRS. Because how the money is collected and how efficient the agency is could never have an effect on revenue right? It's all about the rate itself.

Feel free to answer my sarcasm with actual information I am kind of curious...which regulatory powers, rates and enforcement mechanisms make for the most revenue at the most reasonable rates?

This chart doesn't make any sense. It appears to indicate that the US government collects only 2% of GDP in taxes. The real figure is more like 20%. The rest of the government's tax revenue must be coming primarily from the individual income tax.

All the chart actually proves is that corporations are good at tax avoidance.

Suppose that 2/3 of the GDP is individual income and the rest is corporate income. Then individuals (with income equal to 66% of GDP) pay 18% of GDP in tax. Corporations (with income of 34% of GDP) pay only 2% of GDP in tax. How is that fair?

Waddya expect from a Murdoch-owned paper?

You and I are saying roughly the same things with regard to what a reasonable analysis of the data would show

No, I don't think we are.

For instance, I've said above that, even before you get to the numbers involved, just considering the dependent and independent variables involved, its obvious that the data being analyzed are entirely irrelevant to the idea of the Laffer Curve.

I was, like you, confused about the use of revenues as a % of GDP. What I came up with is that, since GDP measures economic activity within a country (as opposed to GNP measuring the country's income as it were), you might have a case where at higher tax rates you still see economic activity within the country (i.e. part of GDP) but the firms are taking tax-avoidance steps which could explain the lower revenue percentages.

You might, if both the relationship Hassert suggest and the Laffer Curve, which are completely independent of each other, we correct.

OTOH, the relationship Hassert suggests does not demonstrate the existence of a Laffer Curve. To wit, it does not demonstrate (even if the relationship is taken as true) that increasing taxes should be expected to reduce total revenues. First, because it addresses only corporate taxes rather than overall taxes, and thus takes the wrong independent variable to show the Laffer Curve, and second because it measures tax burden (revenues/GDP), rather than overall revenue.

Now, you might say, but you can't compare overall revenue across different countries and do this, because there are too many intervening variables to get useful results. And that's possibly true. But measuring the wrong dependent variable doesn't solve that problem: it may make a more viable comparison, but it doesn't make one relevant to the point trying to be made.

At best, this relationship might suggest, for US policy, that if you want to increase the share of the economy consumed by government, the best way to fund that in the US isn't to increase the nominal corporate income tax rate. OTOH, it doesn't say that you shouldn't raise taxes, or even raise corporate taxes by closing loopholes and exposing more income to them, or to raise taxes on investors by taxing capital gains as income.

Kevin's basic point is that this is an example of bad statistics, and there is no way to argue against him intelligently. Even if you agree with every Conservative Principle in the book, the idea that you could draw a conclusion from the data as presented is laughable. Even if Thoma's graph is statistically weak, there is no way that it is weaker than WSJ's.

If you think about the two quantities being graphed, it makes no sense to graph them. The x-axis represents the tax rate, and the y-axis also represents the tax rate. If you think that lowering taxes causes revenue to go up because it increases growth, then you can't show that without taking growth or revenue amount into account. You can argue that the best way to collect more money is to lower the tax rate, but you can't argue that the best way to collect a higher percentage of incomes is to lower the tax rate.

In addition to being logically incoherent, the graph is based off of bad data. If you want to draw conclusions on what happens with certain tax rates, then use real tax rates.

Thoma's curve is by no means plausible. It is perfectly obvious that the line should start descending at around the 30% mark.

So even after (sensibly) eliminating Norway, the WSJ's point is in fact correct, and it is Kevin, Thoma and his ilk who demonstrate "obvious contempt for their readers' intelligence".

"It is perfectly obvious that the line should start descending at around the 30% mark."

As long as you wish real, real hard, tap your shoes together, and say "there's no place like home!"

Donald Duck in Mathmagic Land! Haven't thought of that one for a long time.

Goodness of fit, folks. Looks pretty low here. For example, if we cut taxes to ~26%, would we end up lower like Germany or higher like whoever is up there near UK? There are other explanatory variables not taken into consideration in this simplistic analysis. Until they are this is useless propaganda from the WSJ.

Without North Sea oil Norway would be some pathetic replica of a Warsaw Pact country.

Similarly, without all his money Bill Gates would be a poor nerd with a bad haircut. It is indeed true that if you take away something it will no longer be there. Powerful, powerful insight....

I recall a social experiment a few years back where researchers went to various countires and "lost" a bunch of wallets loaded with, like, $50, and waited to see if the wallets and money were returned. Norway, if I recall, was one of the few places in the world where every "lost" wallet was returned, cash intact.

So, maybe Norway is a place with unusually honest corporate accounting practices, even at tax rates of 29%.

Norweigans, unlike Republicans, still believe in the rule of law, ie they pay their taxes.

The "Wall Street Urinal." Does it really matter if Rupert Murdercoch purchases it? How much worse could it get?

And if Iraq didn't have oil, it would just be another sandbox in an uncomfortably hot part of the world....and we wouldn't be fighting a war there.

Does that qualify for a QED?

AEI is treated with much respect by the rest of the neo-con press when prognosticating about terrorism and the Muslim Menace.

They're worse than hacks. When the topic is the Middle East, ever notice how many of the editorials in mainstream newspapers, or "experts" on political news shows, are these AEI schmucks?

If you stare at the data long enough, you can see the face of Jesus.

Come on folks.

To argue about about shape of the curve that best fits the data points on the chart is an exercise in futility. That´s because that chart itself is as idiotic as the the " Laffer curve ". The only thing a country´s position on the chart can tell us is it´s size of corporate profits relative to GDP

Because:
corporate taxes as percentage of GDP = (effictve) corporate tax rate times corporate profits as percentage of GDP ; or

y = x times corporate profits as percentage of GDP

For example France and USA have similar x-values while France has a higher y-value. Which only shows only that corporate profits most be higher as percentage of GDP in France than in USA. Which, incidentally, makes one wonder about the quality of the underlying data.

Similarly Luxemburg and Australia share an y-value, but Australia has a higher x-value. Which simply shows that corporate profits have higher share of GDP in Luxemburg than in Australia.

Norways position similarly only shows exceptionally high share of corporate profits in GDP . My trusty calculator tells me that it clocks in at 35% of GDP.
The calculus behind that is simple.
Norwey is an oil exporting nation
Oil price are very high.
Hence vastly profitable oil companies.
Hence exceptionally high rate of corporate profits as % of GDP.
And speaking of Norway: Without oil Norway would be a Nordic country without oil as are Sweden, Finland, Denmark and Iceland. None of which are anything like a pathetic replica of a Warsaw Pact country.

If you follow the x axis you will see that in general the percentage of revenue increases until you reach Norway and then drop sharply. But I think that are too many other variables to say its a direct result of the tax rate.

The real question is whether there should be any corporate income tax. Any tax on a corporation is just passed on to the individual. Just have this thing about paying hidden taxes.

Probably the only fair taxes are the personal income tax and usage taxes.

No doubt about it: my high school math teacher would give me an F if I told him that U-shaped line best fit the given data points. What a joke.

I've noticed for years that logical fallacies are such a staple of right-wing bluster. For example, a rich Democratic Congressman would say, raise taxes on the wealthy, and my dittohead acquaintances (following their talk-radio leaders) would say, "Oh, what a hypocrite." That's as idiotic as saying that a black person opposed to affirmative action is a hypocrite, or a heterosexual opposed to anti-gay rules etc. Being a member of the group that you think doesn't deserve the privileged is not being a hypocrite, but not practicing whatever you preach is the definition. They can't get simple things like that, so are also prey to cheap manipulation of facts or factoids.

PS: The news on NBC tonight, included Iraqi police firing on US soldiers to defend some of their death-squad buddies. To consider that mess possible and worthy of defense is madness at this point.

tyrannogenius

"If I were the Republican candidate, I would start off every speech my just holding this graph up to the crowd for about a minute and just let it sink in. Powerful."

Sure you would Egbert-- and then all those dullard Republicans who go to the Creation Museum to see Noah's arc will clap and drool. "What a pretty piture! Duh"

The odd thing is that this is a completely unnecessary conceit. If you take the data and fit a second-order polynomial over it, it still has its mode at 27% and doesn't change much if you remove the Norway outlier. It's just that the magnitude of the effect is much smaller. Of course the implication is still sketchy because it controls for nothing, but it would at least take a bit of effort to refute the point.

What it shows is tax burden—the overall percentage of the economy extracted by taxes—declining with higher corporate tax rates (past the Norway peak).

Cmnicely is exactly right. What this graphs shows is tax burden on the economy as a function of corporate tax rates. You don't want to be Norway!

Btw, Thoma's graph is equally bunk though. You can't just go and refute an inverse U-shape claim by assuming linearity.

Ooh, oooh, ooooh!

I was real good in kindygarten at connect-the-dots -- let's see if I still got it...

Yes! YES! I found a ducky, an' a horsie, an', an', an'... an' a ANGEL (if you lean to the right a bit and squint.)

(Make that lean to the right a LOT, actually.)

LOL! Thanks, Kevin. This shit is way too rich!

So, maybe Norway is a place with unusually honest corporate accounting practices, even at tax rates of 29%.

I think there's something to this...

In another poll, IIRC, Norway had the highest trust level of any country in the world, followed by the other Scandinavian countries, Canada and Japan. U.K., France, Germany and even China were just a notch below this. I believe the U.S. was closer to Mexico than the other first world nations.

Some good news on the tax cutting front: Last week lawmakers approved an 8.9 percentage point reduction in the corporate income tax rate.

What the WSJ, and some posters, seem to miss is that the figures represent an aggregate, and include personal, corporate, social security, and other taxes. To present that data as validation of supply side economics, or as a basis for determining an optimal corpporate tax rate is patently absurd (see here):

Based on these figures and figures for 2004, OECD analysts say, a trend towards lower tax burdens witnessed from 2000 to 2003 appears to be going into reverse. Between 2000 and 2003, the tax ratio in the OECD area as a whole fell from 36.6% of GDP to 35.8% of GDP, but in 2004 it moved back up slightly to 35.9%.

Higher revenues from taxes on incomes, including both company profits and personal income, were the main factor behind the higher 2005 tax-to-GDP ratios in Iceland, the United States and the United Kingdom (see Table B). In Iceland, an additional factor was increased revenue from taxes on goods and services. By contrast, the decline in Hungary was mainly due to lower revenue from taxes on goods and services.

Tax on personal income and corporate profits is one of the main sources of tax revenues in many OECD countries. But social security contributions and taxes on goods and services also play a major role, and the relative importance of these various taxes varies across countries (see Chart 1). In New Zealand, for example, taxes on income and profits are the largest single source of revenues, while in the Czech Republic social security contributions provide the main single source of revenues and Mexico taxes on goods and services are the main source of revenues.

"You can't just go and refute an inverse U-shape claim by assuming linearity."

You don't have to assume it, the linear relationship is fairly clear right there in the graph. When it comes to social science you rarely get all your points exactly on a line, but it looks like you could estimate a pretty good regression model on those data with an r-squared somewhere around .4 (eduated guess) which is not bad for social science.

I agree with several criticisms upthread about choice of variables, and, of course, a simple 1 variable linear regression model is a vast oversimplification. But that said, for what it's worth, there is a fairly decent linear relationship exhibited in that graph.

The simple linear model may be simplistic and may only account for about half the variance but its not preposterous. However, to argue for a curve based on those data is simply absurd.

Sorry, link in previous post to OECD data should be here.

However, to argue for a curve based on those data is simply absurd.

Really?

"And all good conservatives want a lower overall tax burden, economic expansion, and small government, don't they? So they should love higher corporate tax rates!"
Posted by: cmdicely on July 13, 2007 at 4:45 PM

Yep! And that's where we can get the funding for single-payer health care. And the administrative savings, etc. just might be enough economic boost to overcome all the outsourcing, who knows?

It is Kevin who doesnt understand the Laffer curve. It represents the maximum tax revenue that can be obtained versus the corporate tax rate. So it is not meant to fit the scatter data, rather it is the upper boundary that one can expect.

The editorial is correct. The lack of understanding is here

The head-case leftardos on this thread just can't fathom the record economy figures. Paul Krugboy keeps on predicting recessions that just never come!

So because the figures don't lie, they have to ridicule the economic theory that makes the economy work so well.

Typical liberal logic!

Two things this graph and other analyses by right-wingers always ignores is the notion of marginal rates and when these rates kick in. It isn't 33% on the first dollar of income. You have to make a certain amount of income before you pay anything. This is conveniently ignored by the wingers. This leads to the second point, which Kevin mentioned, and that is the effective tax rate, which you can derive by dividing the actual tax paid (or tax liability) by the corporation's total income. Most U.S. corporation's effective tax rates are far lower than 33%! In some cases, it is zero or less!!

As Mark Twain said, "There are three kinds of lies - lies, damn lies and statistics."

Let me repeat myself: The graph cannot be a Laffer Curve because it does not show the correct variables. If there was an ideal tax rate, you would never figure it out based on this data.

If you want to find an ideal tax rate, you need data that takes economic growth into account. This data does not. Therefore, it does not matter what curve or line you draw through it. It doesn't measure the right quantities to draw the conclusions that the WSJ tried to draw.

If I bought the Wall Street Journal, I would start the first editorial meeting by holding this graph up for a minute. After letting it sink in, I would announce that everybody in the room was fired.

Thoma and Drum did not criticize this graph enough. It is much worse than they described. The physics equivalent would be tracing a projectile using horizontal motion on the x-axis and vertical motion on the y-axis and using the data to prove that the acceleration due to gravity is not 9.8 m/s^2. The acceleration due to gravity is only 9.8 m/s^2 if you plot time on the x-axis.

This comes via Mark Thoma, who draws a more plausible straight line through the data over at his site and finds that as tax rates go up, so does tax revenue.

And of course liberals love increased tax revenue to fund their social engineering schemes.

It's also predictable that liberals, with their linear thinking, would find a straight line more plausible than a more subtle curve.

j lewd: The "Wall Street Urinal." Does it really matter if Rupert Murdercoch purchases it?

first rule of business:

know your customer

if they are easy to fool...

that's just icing on the cake...

As a statistician, I am appalled by the misuse of statistics and graphical methods inherent in this ridiculous plot. There is no way in hell that ANY reputable econometrician or statistician would EVER under ANY circumstances end up drawing this curve.

It's also predictable that liberals, with their linear thinking, would find a straight line more plausible than a more subtle curve.

Statistically speaking, Al, you're a fucking moron.

Don't be so hard on the AEI. Most whores have high heels and miniskirts to work with--all the AEI has is bad curve-fitting software.

A couple of problems with using this graph as "proof" of the Laffer Curve:

-The graph is just one moment in time, you would need to review historic trends to really come to any conclusions. Using this graph of proof is kind of like when the Republicans were claiming that the Democrats would crash the stock market because the Dow took a dip the day after the elections.

-There are a lot more industrialized\high-income nations than are represented on the graph. Why are some left off ? What criteria were used to select
the countries that appear on the graph ?

- There are a lot of factors that influence tax revenue besides tax rates.

Aside from what taxation percentages are optimal as such, how about trying a graduated corporate income tax? I mean, the percent levied on the profit would rise with profit margin (after a deduction, to avoid punishing little companies that have to markup higher to make enough money at all.) Then we reward the industries that serve the public by selling by volume at low profit margin, and penalize the “gougers� without having to directly regulate prices.

You're leaving out the possibility that, since the graph presents tax revenue as a percentage of GDP, when the curve goes to zero, GDP goes to infinity. Yea, we're all rich!

Take all oil related revenues out of GDP and all oil related corporate taxes out of tax receipts.
Your curve would go flattish and the U.S. would probably rise above the U.K. and France as would Germany.

I still advocate that the major oils should do a sale and leaseback of, say, three trillion dollars worth of Pentagon/military assets.

The oil's could take the depreciation to cover all their profits and the U.S. debt's gross annual interest payments would plumment.
Voila, we have balanced the budget.

This leaseback arrangement was originally proposed by Ken Lay in Cheney's secret meeting. He, George, Dick and Rummy et al had an offshore subsidiary all teed up to take a wee commission when EnRon's demise made it too hot for the scabrous looters.

Did I say that? Scabrous looters? Ouch!

--at cognitorex--

If the Corporate Tax Rate is 35% or 29% why do 99% of all Corporations only pay 5% or less while the average American Citizen pay 23% or more?

Where's Edwin Tufte when we need him? All media graph-drawers should be required to take his one-day course on the visual display of quantitative information.

ESPECIALLY those who work for "business" publications or those that tout economic news.

I've suffered with the stigma of paying the wingut-welfare recipients at the Journal's editorial page for years (I want the great news coverage) but this is the worst piece of garbage they ever published.

ANY attempt to fit a line, curve, whatever, would have that entity flow through the center of the data cluster(s). That curve has almost all of the data points (far) below it! (Maybe it was drawn in Lake Wobegon, but in this case, all of the children are BELOW average!)

Second, following Okham's Razor, we assume a line unless we can prove a higher-order explanation (i.e. a curve). It seems easy to fit a line to that data, a line which would slice through the right-hand part of the curve at a near-right angle. (For a plot with N data points, one can exactly fit an n-dimensional polynomial to it, but justifying a function of that high an order would be nearly impossible.)

Finally, the Laffer Curve would be a curve on a graph with the y-axis having tax revenues, not revenues as a percentage of GDP!

Only a matter of time until someone more artistically gifted than I produces an actual connect-the-dots portrait of the Virgin Mary from this scatterplot.

Please post prominently when it happens.

Since Norway is at the top of the curve, here are two factoids:
1. Norway has more millionaires in U.S. dollars per captia than any other country in the world even if you exclude the value of the primary residence. See: http://www.aftenposten.no/english/local/article1881847.ece
2. In some years the dollar value of fish exports from Norway is higher than the value of oil exports. Of course, this is unlikely to happen at 70 USD per barrel unless the oil price stablizes at this level and there is a multi-year period of compensatory price adjustments in other sectors. (Exersice for the reader: at what point did at the pump gas prices, adjusted for inflation, go below 1975 prices in the U.S.?)

Oh my God, how pathetic..........................

1. Norway has more millionaires in U.S. dollars per captia than any other country in the world even if you exclude the value of the primary residence. See: http://www.aftenposten.no/english/local/article1881847.ece

The top personal income tax rate in Norway is 44%.

The oil industry in Norway is nationalized and the profits are controlled by the government.

Statoil (the national oil company in Norway) is 70.9% owned by the Norwegian government. Shares are traded on the MYSE and Oslo stock exchanges. For a list of the top shareholders see: http://www.statoil.com/ir

Statoil stock seems to be popular with U.S. banks.

Norwegian oil prospects are auctioned to oil companies in much the same way as the U.S. government auctions prospects in the Gulf of Mexico.

Most major oil companies are active in the Norwegian sector.

Statoil (the national oil company in Norway) is 70.9% owned by the Norwegian government.

I believe that 70.9% ownership counts as "controlled".

Another factiod:

Norwegian budget surpluses are saved to the "Government Petroleum Fund" for a time when oil and gas stop producing such high revenues.

A top rate of 44% and budget surpluses going into a special government controlled fund. Sound like "supply side economics" to you ?

In your earlier post you referred the oil industry in Norway is nationalized. It is not. Does Norway control the oil industry in Norway. Certainly.

Nationalized means that only the national oil company can exploit oil reserves. Many oil companies bid on leases and produce oil in Norway. Statoil is one of many companies producing oil in Norway. The Norwegian oil industry is not nationalized.

Nationalization and control are different.

Saudi Arabia nationlized their oil industry and Venezuala had started to nationalize their oil industry.

Governments control the oil industry is done by regulation and taxation. For example, the Texas Railroad Commission controlled oil prices for years until the U.S. became a net importer of oil. The U.S. and Norway control the oil industry through regulation and selling leases. And taxes.

I should really get my money back...