24 May 2014

Piketty vs. Giles

Capital & Interest, Shameless Self-Promotion, Tyler Cowen 6 Comments

Here is my initial reaction to the affair. Let me emphasize my main position: I am not an expert on the various estimates of wealth concentration. So I don’t know how to referee on the specific allegations. However, let us not forget that Piketty clearly put in bogus information on both the minimum wage and tax rates that served his narrative. Those weren’t transcription errors. In fact, they’re so bad that I don’t even know what to make of them.

However, what is really frustrating in the reaction to the FT bombshell is people saying, “Well, in the grand scheme Giles isn’t really challenging Piketty’s main thesis.” Huh? From my post, talking first about the UK:

As the figure shows, there is an enormous discrepancy in Piketty’s figures and those that the FT computed. In addition to the figure for the top 10% being off by up to 27 percentage points–the discrepancy that originally caught Giles’ attention–we also see that, depending on which data point one uses for 2010, Piketty’s figure for the wealth held by the top 1% is off by around 9 to 18 percentage points, in the worst case with Piketty reporting a figure that is almost triple the actual value (about 29% instead of 11%, just eyeballing the bottom lines).

Yet beyond the discrepancy in values for given years, step back and look at the overall historical trend if the red lines are correct (as opposed to Piketty’s blue lines): We see that wealth inequality continued its downward trend even throughout the Thatcher years, and after a spike upward during the 1990s, is now (if we use the lower data points for 2010) at the lowest level in recorded UK history.

And regarding the US:

[I]f you look at either of the longer data sets (the bottom red lines) upon which Piketty presumably based his blue composition, you’ll see that neither shows an upward trend in recent decades. In particular, if you look at the longest red series, it shows that the U.S. is currently hovering near the lowest level of wealth concentration in the hands of the 1% in recorded history.

To repeat, I am not saying Giles of the FT has necessarily produced better estimates of wealth inequality than Piketty. But what I am saying is that Giles’ accusations are quite serious. This is yet another issue that completely overturns Piketty’s whole thesis if it turns out in Giles’ favor.

One last thing from me: The reason the different charts don’t look all that different when you eyeball them, is that Piketty hasn’t actually demonstrated that wealth inequality is anywhere near the level it was in 1910 (even though you might have thought that, from hearing people discuss his book). But Piketty was showing a (modest) upward trend since the drop in inequality bottomed out in the 1970s or 1980s (depends on the country), and now Giles is saying that that modest upward trend itself is dubious.

* * *

Here are a few great MR posts on this:

==> Alex Tabarrok explains why it’s important to understand that measured financial capital can increase because interest rates fall. Critics kept asking me, “Who cares Murphy? So what if we can’t just add up tractors and farmland and car factories to get a single number?” Well, Tabarrok shows you the potential pitfalls.

==> You rarely see Tyler Cowen this feisty, here and here.

6 Responses to “Piketty vs. Giles”

  1. Beard Face says:
    • Beard Face says:

      Didn’t realize you just posted it, oops.

  2. W. Peden says:

    On the price of capital issue, Tabarrok says-

    “You can’t, therefore, multiply P by some fixed r and conclude that rents have increased. Indeed, in this case rents, the dividend, haven’t increased at all. For the most part, Piketty simply ignores this issue (at least in the book) by arguing that changes in the price of capital wash out over long time periods but that does not appear to be the case in his data.”

    I am very confused as to why this isn’t a problem even IF the changes “wash out over long time periods”, because Piketty is making use of these statistics to make causal judgements about short time periods. For example, if I have an inflation index that does a good job in the long-run but isn’t good for judging things like intra-cyclical stability (like the pre-Romer stats for pre-20th C. inflation rates) and I know they are flawed, then I can’t reasonably use them to make judgements about things like the stability of RGDP within the cycle, the timing of peaks and troughs etc.

  3. Kevin Donoghue says:

    “Piketty’s theory is based on rents from physical capital” says Tabarrok.

    No it isn’t.

    I do agree that the FT’s accusations are quite serious. Like Ryan Avent and Mark Thoma, I think the FT overdid the rhetoric. But I’m mostly just curious as to how Piketty got some of his numbers. I find his spreadsheets hard to figure out. Granted they aren’t meant for the likes of me, but he needs to post more information so that people like Branko Milanovic can say whether his number-crunching is reasonable or not.

  4. Benjamin Cole says:

    The art market says…Piketty don’t know nothing. A derivative Roy Lichtenstein 1960s comic painting sells for $57 million…people yawn…other “works” by a Klimt or Picasso much higher…seriously the elites have wealth in unimaginable oceans…

  5. guest says:

    consultingbyrpm blog tag piketty

Leave a Reply