The Agony and the Ecstasy of Paul Krugman

In a series of recent posts, Paul Krugman argues--correctly, in my view--that raising the top marginal income tax rates is unlikely to have a substantial impact on work incentives. Unfortunately, he relies on some terrible data to make his point.

Krugman writes:

[We] have pretty good evidence on the (small) actual incentive effects of changes in top tax rates, enough to suggest that the optimal rate is in the 70-80 percent range — which is where it was in the 1960s, a decade of very good economic performance.

Here's the relevant graph:

Why is this so unbelievably misleading? 
Let’s start with the optimal tax rate.

Krugman doesn’t cite the 70-80 percent figure, but he seems to be referencing this recent paper by Christina and David Romer, which estimated “an optimal top marginal [income tax] rate of around 76 percent.”

It's important to note that the Romers’ study relies on data from the 1920s and 1930s, so the external validity already seems questionable. Other studies have produced dramatically different results using different data and methodologies.   

Either way, the Romers are careful to qualify their findings. They point out that their analysis only looks at ordinary income, and not capital gains:

One important complication in these calculations involves capital gains, whose tax treatment varied greatly over the interwar period. To address this issue, we exclude capital gains from our definition of income, and focus on the relationship between taxable income exclusive of capital gains and marginal rates on non-capital-gains income . . . . Excluding capital gains is standard in studies of tax responsiveness, both because they often reflect the timing of realizations rather than current income and because they are often taxed differently than other types of income (Saez, Slemrod, and Giertz, 2012).

All of this is pretty damning to Krugman’s analysis, whether or not he pulled his numbers from the Romers' paper. If it's true that excluding capital gains is “standard in studies of tax responsiveness,” then anyone claiming to have knowledge of the "evidence" on the top tax rates should be aware of this fact.

Of course, it seems clear that Krugman is invoking the Romers' work. Not only does the 76 percent figure fit his claim perfectly, but Krugman’s follow-up post specifically links to Christina Romer’s NYT op-ed about her findings.
In other words, the "evidence" that Krugman is citing seems to be based around a study that ignores capital gains, a common practice when estimating optimal marginal income tax rates. Yet Krugman goes on to compare the 70-80 percent figure against another study conducted by Piketty and Saez, which looked at changes in total income--including capital gains--from 1960 to 2004.

How can Krugman justify making this comparison? He doesn't say. 

To back up his claim, Krugman would have to show that tax rates on ordinary income  (excluding capital gains) have decreased dramatically between 1960 and 2004, since this is the kind of income on which the Romers and others have focused. But as Piketty and Saez point out, that isn't really the case.  

From Piketty and Saez:

[T]he larger progressivity in 1960 is not mainly due to the individual income tax. The average individual income tax rate in 1960 reached an average rate of 31 percent at the very top, only slightly above the 25 percent average rate at the very top in 2004. Within the 1960 version of the individual income tax, lower rates on realized capital gains, as well as deductions for interest payments and charitable contributions, reduced dramatically what otherwise looked like an extremely progressive tax schedule, with a top marginal tax rate on individual income of 91 percent. The greater progressivity of federal taxes in 1960, in contrast to 2004, stems from the corporate income tax and the estate tax.

This isn’t the only problem with Krugman’s analysis. In fact, his entire comparison is apples-to-oranges. The Romers' study--and almost every other study of optimal income tax rates--looks at marginal tax rates (the rates that people pay on their last dollar of income), while the Piketty and Saez paper shows effective tax rates (the rates that people pay on their total income). This seems like a pretty elementary error for a Nobel laureate.

We’re left with this: Krugman almost certainly referenced a paper on the incentive effects of raising marginal tax rates on ordinary income. He then compared it to a paper on the changes in effective tax rates on total income, including capital gains, between 1960 and 2004. This is not a consistent comparison, no matter how you spin it. 

None of this means that Krugman is wrong about raising marginal income tax rates on the wealthiest Americans or reforming the corporate income tax structure.  

But let’s not forget that Krugman is a pundit who routinely bemoans the intellectual dishonesty of his opponents. If nothing else, he should be able to make his points without fudging the numbers.