To be clear, however, this curve doesn't support the WSJ authors point either, because the y axis doesn't make any sense at all. Why is it corporate tax revenues as a % of GDP instead of just corporate tax revenues? Longer posts were written on this point at the time, but the synopsis is: the graph is meaningless.

The biggest problem with the whole editorial and paper was that the variables are not chosen properly for a Laffer Curve.

To get a Laffer Curve, the y-axis has to represent an amount of revenues rather than a percentage of revenues. The idea is that more taxes cause a decrease in growth, which leads to less taxes collected. However, if the y-axis is a percentage, then it is not impacted by economic growth one way or the other.

In other words, WSJ and AEI drew a fraudulent curve that, even if did exist, would have nothing at all to do with the Laffer Curve, and they then claimed that it proved the existence of the Laffer Curve and that all Liberals are idiots.

Of course, what this does is show again that correlation does not mean causation. There are many more important variables than the single, top tax rate, starting with amount of deductions.

But they just couldn't bring themselves to treat the data honestly, and the result was to undermine their own cause.

Not unlike the trolls here, I might add.

And what is the correlation coefficient? Looks to be pretty low. This stuff is all garbage.

So, if the curve is accurate, then instead of getting 2.2% of GDP as corporate tax revenue, we could be getting 3.2-3.3%?

So with stricter enforcement of existing law, and closing loopholes, we can gain 50% revenue, about 1% of GDP? With GDP at $13.7 trillion, we're talking increasing revenues up to $137 billion a year? Sign me up.

Liberal Chris is exactly right about the Y axis issue. My other problem is that, to my knowledge, the corporate tax rates shown on the X axis are nominal and not effective; I believe US corporations really only have about a 16% or so effective rate, after tax credits, subsidies, and other shelters are factored in. I'm guessing that plotting the effective rates would take some of the curve out of that line, and probably render any remaining curve statistically insignificant (is is even significant now?)

Kevin, Brendan's "analysis" and "regression" are full of fallacies. But I'll just explain two of them.
First, the regression must be wrong because both curves say there would be non-zero corporate tax revenue even at zero percent tax rate. This is plainly wrong and absurd.
Second, Brendan totally misunderstands the Laffer curve. The Laffer curver says cutting taxes causes greater economic growth and tax revenues. It says nothing about causing greater tax revenue as a percentage of GDP.
Seriously, liberals like you and Brendan should learn more about what you're talking about.

No one pays any attention to that shit anyway. You'll see.

Right, GDP is the wrong measure, it should be corporate earnings. Corporate Tax revenue = rate*earnings. So if was Corporate Tax revenue/earnings you'd have something. But thats is not whats there.

Al:

Take another look at Brendan's graphs. Most of them don't run to zero on the x-axis; he cuts them off at about 8% (presumably to expand the most relevant part of each graph). The only point where he suggests positive revenue from a zero tax is the linear-model projection in his first chart.

Maybe WSJ just forgot to carry the seven.

Hmm, back on 13 July I took some flack for writing:

"... 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."

You took your eyes off the ball, Kevin. I posted this in your comments the last time. Btw, the argument that "linear is better" is nonsense. For one, a 2nd order polynomial will always get a better fit, in the case where the linear is the best fit the polynomial is simply a straight line. For two, if you're trying to answer the question whether there's a Laffer curve somewhere you look for a curve and not impose a straight line. That doesn't incorporate that the Y-axis is total bullcrap though.

Al,
Neither of those things is a fallacy. Also, the second one (about the meaninglessness of the y axis) is not Brendan's error: he inherited it from the WSJ. (I note that you didn't attack the WSJ editorial board for this in the earlier thread.)

If you think the WSJ has a right-wing bias now, wait untill that impotent old pervert Murdoch takes over!

So essentially the WSJ's first graph was "fake but accurate"?

Sounds familiar.

Corp. tax collections as a % of GDP allows comparability for purposes of the graph. Luxembourg's corp. tax collections, for example, would be minuscule compared to the US or UK at any tax rate.
Making that the y axis makes perfect sense to me.

Al,

Do you even know what a linear regression is? Brendan didn't just go and make a chart and say, "This is my 'regression'." You know that right? Given that you think the regression line should start at zero and go up, I'm thinking no.

There's actual mathematics going on here. The line you see represents the best fit to the data points that you see, meaning that it minimizes the sums of the squares of the distances from each point to the line. That's what linear regressions of normally distributed data do. Unless you can articulate why you think the distribution of the datapoints is not normal, this is a perfectly acceptable analytical technique.

What all this tells us (and would tell you if you had a clue), is that, unsurprisingly, moderate increases in the corporate taxation rate will result in modest increases in corporate tax revenue as a percentage of GNP. As noted in other comments, the y-axis used is kind of pointless, but there it is anyway.

Perhaps you should take another stab at it Al. Or maybe leave the statistical analysis to Brendan, and reveals some of the other "fallacies" you've identified.

Actually reading Hassett's paper reveals some interesting details:

1. The X axis is indeed statutory corporate tax rates and not effective;
2. When he excluded 3 countries he identifies as idiosyncratic in their corporate tax/revenues - Ireland, Norway, and Switzerland - the revenue-maximizing rate is 37%, pretty close to where the US is now;
3. Nowhere is corporate tax revenue as a share of GDP defended or even discussed as the optimal, or even a legitimate, dependent variable;
4. The results that he likes the best - i.e. that show the lowest revenue-maximizing rate - are when he lags the tax rates by 5 years, as compared to tax revenues.

Now, a solid case can be made for lagging 1 year, maybe 2, but 5 years? Nowhere does he explain the economic theory that suggests that a change in tax rates in Year 1 will have their most significant impact to tax revenues in Year 5. All he says is, "There is some evidence, however, that the lag relationship is somewhat stronger, as evidenced by higher quality of the fit." That kind of analysis, my friends, will earn you an F in your average college quantitative methodology course.

"Kevin, Brendan's "analysis" and "regression" are full of fallacies. But I'll just explain two of them.
First, the regression must be wrong because both curves say there would be non-zero corporate tax revenue even at zero percent tax rate. This is plainly wrong and absurd.
Second, Brendan totally misunderstands the Laffer curve. The Laffer curver says cutting taxes causes greater economic growth and tax revenues. It says nothing about causing greater tax revenue as a percentage of GDP.
Seriously, liberals like you and Brendan should learn more about what you're talking about."

What a fucking moron. Regression is a perfectly standard technique, and there is NO reason why the intercept must go thru 0 - none whatsoever. You can force it thru 0 if necessary, but that always results in biased estimates (bias is a technical term, not a political one). And most importantly, Al, you compleat and total boob, even INCLUDING 0 in the model violates the most FUNDAMENTAL tenet of regression modeling - you do NOT go outside of the range of the predictor to evaluate the model. Anyone who is even semi-statistically literate knows that, who is not a total boob like you, boz.

Why is it that conservaboobs are all so CONGENITALLY stupid these days?

I never thought I would agree with Al, but he makes one point that I thought of when I saw the regression lines. Why wouldn't the regression line be forced to have a y-intercept of zero. I don't believe it makes any difference for Kevin's points, but I'm curious. If the corporate tax rate is zero, shouldn't the corporate tax revenue (as a percentage of GDP) be zero as well?

TJM,

Doesn't the whole "as a percent of GDP" kind of confuse the issue, though? If you look at the UK and Australia, they seem to have comparable corporate tax rates, but Australia's corporate tax revenue comprises a much higher percentage of its GDP. Why is that significant? What economic factors play into that? How does it relate to each country's choice of a corporate tax rate?

It seems like if you wanted to clearly show the relationship between corporate tax rates and corporate tax revenue, you'd show just that. The scale would admittedly be different (i.e., luxembourg would be way lower, and the US way higher), but I think the curve you'd get a much better sense of the relationship between reality and the Laffer Curve. Which, come to think of it, is probably why the WSJ didn't use corporate tax revenue as their y-axis in the first place...

Cheers!

Liberal Chris, et al, are 110 percent right about nominal vs. effective rates, due to the large amount of loopholes in the corporate tax code.

That's why, when somebody from a place like Heritage wants to talk about a flat tax, I'm ready to talk -- starting with a corporate income flat tax.

I'm sure Grover Norquist and other tax-wingers don't have that in mind.

Everett Volk wrote:

"What all this tells us (and would tell you if you had a clue), is that, unsurprisingly, moderate increases in the corporate taxation rate will result in modest increases in corporate tax revenue as a percentage of GNP."

But if anything, what the data shows us is that at a certain point a higher corporate taxation rate may actually result in a modest decrease in corporate tax revenue as a percentage of GNP. I agree that there are limitation to how far we should interpret the data, but it hardly says definitively, what you claim it said. There is at least room for disagreement.

Why is it that conservaboobs are all so CONGENITALLY stupid these days?

I once saw a study that found a correlation between low IQ in children and alcohol use by the father.

Just to be clear, I accept being less than semi-statistically literate, so that I don't have to cop to being a total boob.

Hacksaw,

Point taken. I may have been caught up in my own rhetorical excesses.

Everett

"Now, a solid case can be made for lagging 1 year, maybe 2, but 5 years? Nowhere does he explain the economic theory that suggests that a change in tax rates in Year 1 will have their most significant impact to tax revenues in Year 5. "

Well, I just betcha - and I would betcha 45 cents here - that this is some kind of stepwise technique. When you start with lagging at all possible terms, and let the best win out in some kind of backward elimination stepwise, you get this kind of moronic, idiotic and nonsensical result, in which the BEST FITTING but NON-SENSIBLE result comes thru.

Brendan Nyhan, who clearly has too much time on his hands, took the raw data provided by Kevin Hassett and replotted it using standard tools. The result is on the right: if you do a linear regression you get the red line, while if you apply a quadratic model you get the blue curve.

Uh, yeah. While there was no graphic (hard to do that in your comments), someone made that exact point in the comment thread here on the original WSJ joke:

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.

CTN: "All he says is, "There is some evidence, however, that the lag relationship is somewhat stronger, as evidenced by higher quality of the fit." That kind of analysis, my friends, will earn you an F in your average college quantitative methodology course."

Finally a place where the term "begging the question" can be used in its original sense!

"If the corporate tax rate is zero, shouldn't the corporate tax revenue (as a percentage of GDP) be zero as well?"

Simple reason. 0 was not in the original range of the X values. The X values run from 10-35. Thus, it is a violation of the rules of regression prediction to use 0 as a value in generating a predicted value. The reason here is simple - the values within the range are assumed to follow a rule (defined by the least squares line of best fit called the regression line) but we cannot say if values outside follow that same rule, and it is very easy indeed to find many violations of rules outside of the range of the X values.

Why wouldn't the regression line be forced to have a y-intercept of zero.

For the same reason that the regression line is not also forced to be zero at 100% taxation -- it would give us a trivial and wholly unhelpful solution.

We are interested in modeling the behavior in the middle of the curve.

But if anything, what the data shows us is that at a certain point a higher corporate taxation rate may actually result in a modest decrease in corporate tax revenue as a percentage of GNP.

Sure, because lifting the burden on workers (and thus, most consumers) compared to the public benefits received (increasing the corporate income tax rate allows the government to do more, drag the economy down with public debt less, or cut other taxes) accelerates the economy so much that the resulting GDP growth is greater than the increase in corporate tax collections.

So, higher corporate taxes are good idea on the right side of the curve.

I never thought I would agree with Al, but he makes one point that I thought of when I saw the regression lines. Why wouldn't the regression line be forced to have a y-intercept of zero.

No, regressions are empirical.

If the corporate tax rate is zero, shouldn't the corporate tax revenue (as a percentage of GDP) be zero as well?

Probably, but its quite likely a special point case, in which case that won't be reflected in any regression which provides a useful generation of the behavior anywhere else on the curve. Its possible that driving corporate tax rates toward zero redirects what economic activity there is into corporations so much as to increase the share of the overall economy, which may also shrink as government can do less to support it, represented by corporate taxes until you actually hit 0%, at which point nothing is collected.

cmdicely,

Now that's the kind of analysis only a liberal could love!!!

On a serious note, though, I think it is interesting that so many of the countries liberals think we should model ourselves after have significantly lower corporate tax rates. Part of the answer may in fact be that their actual rates are closer to their nominal rates but the reality is their governments are able to "do more" chiefly through higher personal income tax rates. In other words, they have declined to shift the tax burden away from workers for some reason.

the WSJ is now officially irrelevant--you can expect much worse than this from now on.

It's curious that the modern neo-supply siders avoid the original Laffer-Wanniski argument (see Wanniski's book The Way the World Works) and adopt pretty much a modern post-Keynesian argument over tax rates; they now argue that tax cuts stimulate growth, and that this is the reason for their fanciful curve. Au contraire: Wanniski spent the first part of his book looking at banana republics in which there was a very high nominal rate of taxation (ie: 90 percent for everything above twenty-five thousand dollars yearly income) and in which the rich paid very little. The point they made was that in such economies, people go to great lengths to avoid taxes, including camoflaging their wealth and misstating their income. The book then went on to talk about the black market economies that resulted.

What's curious about this is that it said essentially nothing about the stimulatory effects of deficit spending, presumably because Wanniski and Laffer were into going back to the year 1900 in terms of tax, wage, and monetary policies.

One problem with the original Laffer curve is that it postulates the tax rate as the x axis; Since a modern progressive tax system has several tax rates which scale up from zero, the Laffer curve could only be applied to the situation for marginal income at the highest rates, and even then the argument is illogical: Under the pre-JFK rates at 90 percent, people still paid taxes on their top dollars and continued to make money in the highest brackets; whether or not this is good for the economy is another question, but the Laffer curve certainly never described anything remotely like what the U.S. has had for a tax system. A more likely real world prediction is that a nominal 100 percent tax rate will fail to collect 100 percent of top bracket income, but it will collect something. The Swedes used to have something like this; the result was that people who could shield income by moving to Monaco did so, but others couldn't and didn't.

On a serious note, though, I think it is interesting that so many of the countries liberals think we should model ourselves after have significantly lower corporate tax rates.

(1) Liberals don't generally think we should model ourselves after other countries. They occasionally think that we ought to emulate particular, successful policies of other countries, but that's not the same thing.

(2) Most of the countries you are looking at also have tighter regulations of corporate behavior.

Part of the answer may in fact be that their actual rates are closer to their nominal rates but the reality is their governments are able to "do more" chiefly through higher personal income tax rates.

Perhaps on average, but largely, where true, through higher top end rates, and greater progressivity.

Chris: "Why is it corporate tax revenues as a % of GDP instead of just corporate tax revenues?"

I tell you, Chris, this "percentage of the GDP" sure comes in handy. Why, when measured against that standard, our national debt magically becomes insignificant. We can thus overspend ad infinitum and our economy will still expand indefinitely.

Also highly recommended for use in conjunction with that little-understood but important-sounding standard is the judicious and well-placed Latin phrase, like the one in the previous sentence. This grants almost any proponent of "unfettered free-markets" that veneer of intellectual expertise so often necessary nowadays to convince an otherwise-hesitant "John & Joan Q. Public" that they needn't worry because you obviously know whatever it is that you're talking about.

Therefore, it is imperative that you provide the general public that misplaced confidence by which citizens can freely ignore their initial gut instinct that this is really all a crock of shit, and they can continue to go on obliviously about their business.

By the time they do finally catch on, the proverbial horses are already long-escaped from the equally-pproverbial barn, and you will have made your millions and moved on to your next mark.

See how easy it is? For additional effect, try drawing it out on a cocktail napkin for your favorite GOP presidential candidate, and who knows? You, too, may be able to profoundly influence the public debate on our national economic policy for the next quarter-century.

As a supposed proof of the Laffer Curve, however, having the y-axis be tax revenues as a % of GDP is singularly stupid. The idea of the Laffer Curve would be that as tax rates go too high, they stimulate lowered economic activity. Thus, the Laffer Curve's insight is that tax increases, at some tax rate, won't actually increase tax revenues because of the disincentive effect.

That might be visible if the y-axis were tax revenues. Then you might see that at high tax rates, tax revenues actually declined.

But these bozos use tax revenues as a % of GDP as their y-axis. What the hell will that prove? According to Laffer, as the tax rate increases, tax revenues go down, but that's because GDP itself has gone down. Indeed, the revenue decrease predicted by Laffer is proportionate to the decline in GDP. Thus, these idiots' y-axis is necessarily eliminating all even arguable Laffer Curve effects. And whatever accounts for the data, it is necessarily not the effect predicted by Laffer.

"There is some evidence, however, that the lag relationship is somewhat stronger, as evidenced by higher quality of the fit."

He might as well write:
"There is some evidence, however, that the lag relationship is somewhat stronger, as evidenced by higher quality of the fit!!1!1!"

Does his employer know that he wrote this garbage? This type of statement in a manuscript for a real journal would get a note back requesting that the writer never submit another manuscript.

cmdicely,

You wrote "Perhaps on average [governments that are able to "do more" do so chiefly through individual taxation], but largely, where true, through higher top end rates, and greater progressivity."

This study sheds some interesting light on that proposition:

http://elsa.berkeley.edu/~saez/piketty-saezJEP07taxprog.pdf

The discussion of US tax progressiveness versus France and the UK starts on page 17. Granted it is only one study, but it found among other things that "In France, as of 2005, the regressivity of the payroll tax system undoes the progressivity of the individual income tax system, so that the resulting tax system is basically flat."

Anyway, I'm not suggesting this invalidates your point, I just thought the study makes for an interesting read.

it found among other things that "In France, as of 2005, the regressivity of the payroll tax system undoes the progressivity of the individual income tax system, so that the resulting tax system is basically flat."

It's basically the same way in the US. That is why those who call for a flat income tax are actually calling for drastic regressivity in the over-all tax structure.

many of the countries liberals think we should model ourselves after have significantly lower corporate tax rates

They probably have significantly different corporate charters too.

I would have no problem with zero percent taxation of corporations, if corporate law were changed to require equal voting power per stockholder; not one-share-one-vote, but one-shareholder-one-vote.

Also, let's require corporations to disperse all profits every year, and require all board members to work 40 hours a week. And CEOs must have the same size cubicle as the support geek.

Since CEOs are so damn brilliant, they surely don't need a huge office, a jet and all the other perks that just get in the way of their important brainstorming.

See, a corporation is nothing more then a mask for aristocrats. Time to take away the damn mask, so we can all get a good look at the clowns who think they're so much better then the rest of us.

Chris, your point would make sense if the graph was a time-series chart of, say, US corporate tax revenues at various top rates. This particular graph is meant to replicate a Laffer effect using various countries tax rates as a proxy.
Relax, have a drink.

Has NEWS CORruPt and Rupert Murdoch already taken over the WSJ?

I would have no problem with zero percent taxation of corporations, if corporate law were changed to require equal voting power per stockholder

As long as stockholders and corporations are not both totally contained in one country, there needs to be taxes on each. Each type of entity both demands and causes government services within each country they operate or reside.

TJM--
While we're relaxing, please explain to us what the Laffer Curve predicts about Corporate Tax Revenue as % of GDP.

BobG: I've always thought the Reagan-era supply-sideism was just Keynesian economics in drag.

Joey Girard: I'll take your idea one better and hope for the day SCOTUS overturns the idea that a corporation is a "person" before the law.

TJM - I'm pretty relaxed. I'll explain it again though so you can explain how I'm wrong.

Laffer argues that in some situations, when you raise tax rates, revenue actually goes down, because economic growth stagnates as a result of the tax hikes. GDP goes down so much that we actually lose revenues from raising tax rates.

Here's a very simple example. Imagine we have a 30% tax rate and GDP is 100. So tax revenues are 30.

Now let's imagine we increase tax rates to 50%. If GDP stayed at 100, we'd get tax revenues of 50. But Laffer points out that the tax increase might slow the economy and thereby cut both GDP and tax revenues. So let's imagine the economy takes a huge hit and goes down to 50.

Now even though the tax rate has been increased from 30% to 50%, the hit to the economy is so great that GDP has declined from 100 to 50. 50% of 50 is 25. So we actually get less tax revenue despite our tax rate increase - we now get 25 in revenues (at a 50% tax rate) rather than 30 in revenues (at a 30% tax rate). This would be an example of the Laffer effect.

But the WSJ guys are graphing not tax revenues, but rather revenues as a % of GDP. In my example, at the initial 30% tax rate, with tax revenues of 30, tax revenues as a % of GDP were 30% (30 in revenues divided by 100 GDP). At the 50% tax rate, with tax revenues of 25, tax revenues as a % of GDP would be 50% (25 in tax revenues divided by 50 GDP).

Thus, my example - which was specifically designed to show that the Laffer effect occurred - would actually not show any such effect when graphed by these idiots with their stupid y-axis. Indeed, the y-axis of their graph would just end up showing the tax rate itself, by virtue of the fact that the loss in GDP is affecting both tax revenues and GDP and is therefore cancelled out when you calculate tax revenues as a % of GDP.

Now by all means TJM, explain why the graph makes sense and why I am wrong.

So the WSJ was right and Kevin was wrong.

And Kevin presents us with what has to be the least graceful concession we have ever seen.

Since you already know that you know the answer, it won't make a difference, but I'll try. Your example above uses a theoretical outcome measuring the same economy over a period in time. The graph shows different economies at different tax rates of collection at the same point in time. Since the design is to show that at higher tax rates, tax revenue declines, whether the measure is of gross tax revenue (in the currency of your choice)or as a % of GDP, the outcome is the same.
The graph replotted here, doesn't show that there is a Laffer effect, which is a dramatic decline in both taxes and economic activity, only that at some certain high rates of tax, a decline in tax revenue and/or a decline in economic activity occurs.
It shouldn't matter whether the y axis is gross revenue or as a % of GDP.
What isn't known about these economies is at what point(s) in time the countries graphed raised or lowered their tax rates. At this single point in time (lagging the effect for all the economies doesn't matter either) using the same data for each (tax revenue as a % of GDP) yields the effect shown. which effect is not at all what the WSJ showed.
There's nothing wrong with using GDP %ages since the point isn't whether the y axis is valid or invalid, the point is that the WSJ made up a graph to fit the theory when a real graph of the same data showed no such thing.

We can now look forward to that same commitment to accuracy on the news side now that Rupert's bought the paper - no?

How about throwing in the British Virgin Isalnds, Cayman Islands, Dubai, Hong Kong, Dubai, Gibraltar, and a few other tax havens. Luxembourg looks lonely.

I'll help out.

BVI has no corporate taxes (just small licensing fees averaging 400 dollars/company). They would fall at 18%.

Scanning the original and comments, I do not see mention of the correlation coefficient, i.e. how much of the data is actually accounted for by the regression. It is obvious looking at the graph that the cc is very small, meaning that there is probably no significance at all to either curve.

TJM states that "Since the design is to show that at higher tax rates, tax revenue declines, whether the measure is of gross tax revenue (in the currency of your choice)or as a % of GDP, the outcome is the same."

This is incorrect. The only reason tax revenues arguably decline due to the Laffer effect is because the economy is hurt by the tax rate increase. That is to say, GDP goes down when tax rates go up, and tax revenues go down because even though rates are higher, the economy being taxed is smaller.

You would arguably see this Laffer effect if the y-axis were tax revenues. But this y-axis is double-counting the changes in GDP caused by tax rate increases. The GDP effect lowers both the denominator of the y-axis (GDP) and the numerator (tax revenues), therefore cancelling itself out.

The fact that the graph is comparing different nations to each other is irrelevant. The point is, the points on the graph are in different places for reasons other than the Laffer effect. The researchers themselves have made it impossible to see the Laffer effect by choosing such a stupid y-axis.

First off, there is no Laffer effect except in very narrow circumstances.
It seems to me, though, that you are making your point and missing your point at the same time. The data used shows what you argue should occur. The countries in the lower right are showing lower tax revenue as a % of GDP due to lower GDP because the higher tax rates cause GDP to decline. Of course those in the upper right how just the opposite, hence the curve fit.
Not sure why you keep going around his circle but I suppose you have your reasons.

TJM, i really think you don't understand. Why would tax revenues decline if tax rates INCREASE? The only reason is because the tax rate increase hurts the economy (GDP).

How can this graph possibly get at whether this is occuring or not? The y-axis includes GDP twice. Tax revenues, if lower, are lower because GDP was hurt. But in that case, GDP is lower too. So who knows what will happen to "tax revenues as a % of GDP". That value will say nothing about whether the Laffer effect is occuring.

Thus, you are wrong when you say that "The countries in the lower right [of the graph] are showing lower tax revenue as a % of GDP due to lower GDP because the higher tax rates cause GDP to decline."

If the only effect of the tax rate increase were to decrease GDP, "tax revenues as a % of GDP" would actually go up. If we then also remember that the tax revenues will decline as a result of the GDP decline, the GDP decline has been eliminated from the y-axis.

The pattern of data in this graph says literally nothing about the Laffer curve as a matter of necessity because the GDP effect of tax rate increases is not visible on the y-axis.

whatever.

I love playing with excell.

So according to the wall street journal the share of GDP that is in the form of corporate income for various nations is:

Norway......34%
France.......8%
US...........6%
Australia...19%
UK..........12%
Germany......6%
Luxembourg..26%
Canada......16%
Ireland.....28%
Iceland......7%

This was easily by mutiplying the inverse of thier "corporate tax rate" (which is really the top corporate profit tax rate) by the Revenue generated as percent GDP.

I don't think the massive differences evident in this data are likely to be real. The more plausible explanation (as others have suggested) is that these "tax rates" the WSJ uses are not at all realistic for this purpose.

I'm having trouble googling up any international corporate profit as %gdp data from outside the US, but using the number for the US (10%) and corporate taxes as %gdp yeilds a real corporate profit tax of 21% for the US. I'm not hugely confident because one way to reduce tax paying is to reduce "income" in a technical way that might affect both taxes and the profit as %gdp number.

Interesting.

If you plot the corp tax revenue per capita vs real tax rate assuming that in each nation corporate profit as a share of GDP is approximately 10% (which is surely off to some extent) you get a very nice straight upward trend that points right toward the 0% effective tax rate of UAE, and just a bit below the 100% effective tax rate you end up with for Norway (which is obviously nonsense, but that 10% assumption is pretty ludicrously simplistic, my guess is that norway has a higher than normal corporate share of GDP because of its massive oil revenues).

If only I could find that corporate share of gdp data... the internet is just too cluttered with articles about corporate taxes as percent of gdp, and I haven't found a search string to differentiate the two.

Liberal Chris wins.

am and TJM lose.