Laffer Curve Revisited

You may recall The Wall Street Journal's July 13 editorial which proved that the Laffer Curve is real if you restrict your attention to corporate tax rates, mis-code Norway, and draw your line wrong. Thanks to Kevin Hassett's efforts to spell this argument out in more detail, Brendan Nyhan was able to look at the exact same data and draw the line correctly:

If you exclude Norway, who's oil tax revenue shouldn't really be lumped in with corporate income taxes in general (it's more like a royalty), things look even less like the WSJ version of reality.

Comments

However you draw the line, the graph is irrelevant to the Laffer curve - it has the wrong axes. Laffer didn't claim lowering the tax rate would increase the percentage of GDP which gets collected as taxes. That's just silly, and the above graph is more about how many corporate tax loopholes each country has.

The point of the Laffer curve is that lowering tax rates will increase economic growth, thereby (if you are at the right region of the curve) increase total revenues in terms of dollars, even though (by definition) taxes will be a lower % of GDP.

Posted by Mark | August 2, 2007 10:52 AM

Not to mention: what the heck is the UAE doing on that chart? If we had taxable migrant residents making up 95 percent of our private-sector workforce, plus huge direct income to the state from lots of state-owned oil, and a tiny citizen population, we could have zero corporate tax too!!!!
Remove all the oil states, not just Norway.

Complete tendentious fraud, typical of the WSJ ed page.

Posted by Jim M | August 2, 2007 10:52 AM

Mark is absolutely right. I would guess (if we adopt the mindset of pseudo-analysis typical of WSJ) that the reason the y-axis is % of GDP is because gross dollars of revenue will obviously depend on the size of the economy over all, so the % of GDP number is an attempt to make different countries comparable. Yet Laffer's hypothesis is not that decreasing tax rates increases the proportion of tax revenue to GDP.

For exactly that reason, it simply isn't possible to draw a Laffer curve across countries because (inasmuch as the curve exists at all) it will have a different shape in each country. This whole exercise is bullshit on so many levels.

Posted by Marshall | August 2, 2007 11:21 AM

Norway probably doesn't belong, but it's pretty clear that a UAE point makes a lot of sense: at a zero corporate tax rate you know that you'd collect zero taxes.

Most arguments about the Laffer curve turn not about whether you collect zero revenue at the two extremes, but on what the shape of the curve is between those two extremes: assuming a quadratic, and assuming you hit zero at both ends, just about guarantees that the highest revenue will be at 50%. If you don't assume a quadratic, though, there are lots of shapes that could connect the end points. Given the degree of scatter in the graph, and what we know about all the special circumstances that could be cited for each country, I suspect that this approach is unlikely to shed any light on the underlying relationships.

Posted by Barry | August 2, 2007 11:24 AM

Why pay any attention to this pointless analysis by Nyhan? The curve obviously doesn't fit the data very well anyway, but more important is that the whole exercise would make no sense even if it did. The axes are not the ones used for the Laffer curve, and comparisons across countries have too many variables to be meaningful. With garbage in, you get garbage out, regardless of whether you draw the curves "correctly" or not.

The proper response to the WSJ graph isn't to give it legitimacy by "fixing" one of the vast numbers of errors that produced it (though one might expect a member of the cult of false balance like Nyhan to try such a thing). The proper response is to denounce it as a shocking example of innumeracy from a major newspaper.

Posted by KCinDC | August 2, 2007 12:14 PM

I think that it would be more informative to have one of the axes measuring the percentage of redheads in the population, rather than the rate of corporate income tax.

The yield of that tax depends on too many factors, making the tax rate no more of a predictor than the distribution of hair color.

First, small states with low rates may benefit from multinational corporations manipulating internal transactions so the profits are accumulated in low-tax jurisdictions. This explains very high yield for Ireland. Luxemburg is also a good place to declare profits.

Second, the details of tax law of a country may classify a large portion of the cash flow of a corporation as non-income (sometimes even a loss). In this way USA currently has very high tax rate, very high portion of GNP in corporate profits and low yield. I guess that this phenomenon explains why many financial corporations have very profitable branches in Luxemburg, in spite of not very small corporate tax rate.

So the chart is in a way interesting, but more as illustration for the phenomena of tax loopholes and tax heavens than anything else. One can consider if some vast simplification of corporate tax laws, closing most of the loopholes and decreasing the rates, would be beneficial.

I think that "Laffer curve" phenomenon is supposed to be that if the taxes are too high, the taxable economic activity is depressed and eventually it leads to the loss of revenue. Of course, there are other phenomena at work too. You keep a high tax rate and you allow to register profits in a tax heaven. Surprise! the revenue goes down. Predicting that does not require any curve fitting and what-not.

Posted by piotr | August 2, 2007 12:35 PM

Yeah, this has been beaten to death. One more point though: none of the curves is going to be that predictive, because there's huge scatter. The real point is there's not much of a relationship here!

Posted by chris | August 2, 2007 12:50 PM

If every country on the right side of the deflection point suddenly came to their senses and said, "We could lower corporate tax rates and still collect as much revenue (% GDP) as Canada!" and adjusted their tax policies accordingly... how would that change the curve?

Posted by Grumpy | August 2, 2007 12:55 PM

As Brad DeLong has pointed out, you don't need to remove Norway: you just need to plot it correctly. Norway's effective tax rate here is more like 52%.

Posted by phil | August 2, 2007 1:34 PM

The whole idea of the Laffer curve is that raising taxes reduces GDP, so it makes no sense for the Y axis to be adjusted by GDP, as if that were an independent characteristic. (Of course it would also make no sense to use absolute amounts of taxes raised across a collection of countries with widely varying populations, socioeconomic levels, government policies, and other characteristics.) Suggesting changes to this "curve" at all, rather than ridiculing it as a meaningless collection of points, is granting it a level of acceptance that undermines your credibility.

Posted by KCinDC | August 2, 2007 1:39 PM

I realize that at some point each highlighting the differences between countries outside of their tax rates can get out of hand, but Luxembourg also seems like a bit of an odd duck.

Posted by Warren Terra | August 2, 2007 2:00 PM

Actually, that graph shows strong evidence for the existence of a Laffer effect -- you'll notice that none of the countries have a corporate tax rate over 36%. Presumably, 64% of the possible corporate tax rates aren't used by major countries these days because they are self-defeating.

The old fogeys out there can remember days in the 1950-1970s when there were self-destructive tax rates clearly on the downward side of the Laffer Curver. There aren't many of them anymore.

Posted by Steve Sailer | August 2, 2007 3:00 PM

Ireland has done extremely well for itself by lowering its corporate tax rate to 12% -- setting the vertical axis in % of GDP understates how well Ireland has done because its GDP has been skyrocketing. Whether that can be replicated in a big country is a different question -- lots of multinationals contrive to take profits in Ireland to pay the low tax. But, still, the impact on Ireland is real -- Ireland's standard of living is much higher than it was before it cut taxes.

Posted by Steve Sailer | August 2, 2007 3:39 PM

caywh yvhzlgd cxkrjv psbjwvyc hkxjviwl iakdumpzg qnmt

Posted by jhqrgbm kluf | August 8, 2007 5:17 AM