Thomas Piketty and His Critics

     In my last post I refereed to Chris Giles' attack on Piketty as largely nitpicking and quibbling. I also quoted Krugman who said that even if he had found genuine errors in Piketty's work he proves too much. 

     "Sorry but in reading Giles this is the first thought that comes to mind. What first jumps out at you is the triviality of his concerns, not to say quibbles. It's not clear that even if Piketty has failed to dot an i or cross a t why this matters so much materially for his work. Of course, you know what the conservative man in the street, a la Scott Sumner answer will be for this: well R-R's Excel error was at least as trivial if not more."

      "In saying this, though Sumner and friends give away the game-they're drawing equivalences with R-R because in their mind this is payback for what was done to R-R. Another bad sign is that Giles immediately goes from pointing out various 'errors' real or imagined and then makes the categorical declaration that inequality has not increased since 1980 and so the problem is greatly overstated by the Piketys of the world. As Krugman notes, in saying this, Giles says too much. Of course Sumner's favorite debunking of inequality in America and Europe is to talk about inequality in the world and suggest declare the answer is more free trade and low corporate taxes-and more fiscal austerity. "

      http://diaryofarepublicanhater.blogspot.com/2014/05/chris-giles-attack-on-piketty-as.html

      Now Krugman:

      "It’s just not plausible that this increase in the concentration of income from capital doesn’t reflect a more or less comparable increase in the concentration of capital itself.

     "The point is that Giles is proving too much; if his attempted reworking of Piketty leads to the conclusion that nothing has happened to wealth inequality, what that really shows is that he’s doing something wrong."

       However, some might read this and want to see some examples of what I mean by quibbling and why Giles is proving too much. 

        Well here is Giles allegedly skewering Piketty. 

       "There appear to be few problems with the choices made by Prof. Piketty for Sweden. These are mostly data omissions, transcription errors and odd choice of data to represent the years in the graph.

For 2010, I use the latest data from 2006, which shows a small decline. Prof. Piketty uses an average of 2005 and 2006, but does not explain why."

       "He also chose to use 2004 for 2000, when the data point for 2000 was available in the sources he cited. I prefer to stay with the 2000 data."

         The main problems relating to the French numbers used by Piketty seem to relate to the arbitrary tweaks he uses for 1910 which raises the wealth share at the top around the turn of the 20th century (see 1-b).

       "The other main difference is that I have taken data for the year in question rather than an average of the data for the rest of the decade. This makes the series more compatible with other countries."

      http://blogs.ft.com/money-supply/2014/05/23/data-problems-with-capital-in-the-21st-century/

       Yes, that's what I call quibbles. It's different to see why using 2005 and 2006 rather than 2005 in the Sweden data should immediately prove that inequality has not been on the rise. This is what Krugman means by proving too much-even if this quibble is correct how it demolish the idea that inequality has risen since 1980? Giles is expecting too much from such small potatoes. Giles 1980 to 2010 numbers show that inequality has risen during this period whereas Giles' counter numbers which are supposed to be a corrective show it basically flat. In theory then even he admits that inequality stops falling in 1980 after having fallen for the previous 6 decades.  If he's right about using 2005 rather than 2005 and 2006 does this show that inequality has actually decreased? Hardly. 

       For this reason he seems to me to also be quibbling over France in 1910 to 1920 by claiming that Piketty somehow overstates its rise. Even Giles own FT chart shows that inequality peaked in France in 1920 then sunk for 6 decades so how can much be at stake here other than showing an alleged 'pattern' of Piketty somehow doing shoddy work. He also goes way too far in accusing Piketty of 'cherry picking data' based on what he's criticized him on. What actually happened is actually Giles mixing apples and oranges as Krugman noted in the above quote. 

       So sorry to disappoint the Sumners and Tyler Cowens of the world but this attempt to abort Piketty has proven itself abortive. No doubt there will be other tries-we'll see if/when Sumner tries to rescue Piketty. My guess is that if he does he'll use that usual gadget in his toolbox, the claim that liberals by worrying about inequality in America are actually ethnocentric monsters themselves for only caring about inequality within their own countries but not world inequality-which Sumner would tell you, may make the life of the average American or Briton worse but improves the lives of peasants in Asia or Latin America.