When Thomas Piketty called for marginal tax rates that go as high as 80% on top of an annual global wealth tax in Capital in the 21st Century, he undoubtedly knew that it would draw heavy criticism. But as much as people have attacked his proposed solutions, most accept his empirical work to be sound.

“It is increasingly evident that a sizable fissure exists between the bold data claims that are made at the outset of Piketty’s Capital in the Twenty-First Century and his actual empirics under closer scrutiny,” write Phillip W. Magness and Robert P. Murphy, rejecting that consensus in their paper Challenging the Empirical Contribution of Thomas Piketty’s Capital in the 21st Century. “An abundance of questionable and problematic data claims may well mean that empirics are the book’s weakest point.”

Piketty makes some embarrassing factual errors

Some of the mistakes they bring up really are embarrassing, since a fact checker should have caught them somewhere along the way. Piketty gets the dates wrong for both tax increases after the Great Depression and minimum wage hikes in the 80s and 90s, and in both cases the mistakes better fit Piketty’s narrative. It’s not surprising for a few mistakes to slip into a book during writing, but it’s surprising that no one caught them before publication.

Piketty also extrapolates tax revenue in the US back to 1870 when he doesn’t have any data before 1902, basically averaging his 1902 – 1909 data and drawing a line straight back. This makes his US series lineup with the UK, France and Sweden.

us tax data piketty 0115

Piketty makes another strange decision when he looks at UK inheritance flows. His data goes back to 1896, but he starts his decennial averages at 1900. To include all of his data he staggers the first few decades so that 1900s are expressed as the average of 1896 – 1902, and the 1910s are the average of 1904 – 1910, but later decades are what you would expect (eg the data point for the 1950s is the average for 1950-1959). The result is a bump at 1910 that disappears if you average out the flows consistently.

UK inheritance flows 0115 Piketty

The FT/Piketty debate continues

Finally, Magness and Murphy rehash the debate originally brought up by Financial Times journalist Chris Giles, arguing that Piketty chose a specific method of measuring wealth inequality (based on capitalized income) that reinforced his main point about growing inequality while ignoring different results (based on estate taxes) showing that wealth inequality has been roughly flat since the 1980s, even though income inequality has grown.

estate tax measure of inequality 0115 Piketty

Saez, who is responsible for both the red and blue data series in the above chart says that the capitalized income data gives a truer picture of what’s going on, but Magness and Murphy argue that Piketty should have explained that different methods of measuring wealth find different levels of inequality. As the researchers themselves acknowledge, the consensus is that Piketty’s data is solid, and nine months’ of debate following Giles’ critique haven’t changed that. While Piketty should revise the mistakes in the next edition of Capital, it’s unlikely that this new research will undermine his empirical work either.