“Up until Giles's shock revelations, Capital in the Twenty-
First Century was the liberal-left's philosopher's stone:
the magical device capable of transforming base socialist
instinct (greed, envy, hatred, control-freakery, love of the
state, obsession with equality) into intellectual gold. This
week, it's just another worthless leftist screed in the dustbin of history.”
- James Delingpole, “Piketty -- Author of Das Kapital in the 21st Century -- Rumbled”
“Piketty said the discrepancies were simply
adjustments to raw data to smooth
them over time and across countries.”
Kate, from Small Dead Animals blog,
from her post “Everything He Knows, He Learned
From Temperature Records”
Thomas Piketty’s best-selling book Capital in the 21st Century has sailed forth arguing that capitalism holds a fatal flaw. Having been virtually sainted in the four short weeks since the book’s release, the SS Piketty appears to have hit a reef now that reviewers have had time to delve into his working.
Numerous reviewers have already pointed out problems with Piketty’s theory that mature capitalism inexorably produces a fatal wealth inequality only remedied by a global wealth tax – many of which you can read here at NOT PC -- problems to which when pointed out he and his opponents have simply responded “but look at the data.”
So the Financial Times’s senior economics editor Chris Giles did look at the data, comparing it to the sources Piketty claims for it. They gave it a really close look, trying to emulate his charts and tables from the data sources. Their conclusion, after trying, is that Piketty’s results are too often overcooked, cherrypicked, mistakenly transcribed, incorrectly fitted to each other, and (in the case for example of U.S. records for over a century) simply “made up out of thin air.”
What Giles and his colleagues found appears devastating.
They found that the data, as presented, contained (to say the least) substantial inaccuracies. More bluntly, if the correct figures from the sources he cites are used, and the calculations are performed correctly, the effects he claims to describe vanish entirely.
You can click the FT link above to read the big-picture (ungated) story, and this (gated) link to see the specific allegations about Piketty’s data errors.
Giles explains what led him to start questioning Piketty’s historical figures on wealth and income concentration, which most other reviewers (including me!) had simply assumed were in the ballpark:
[W]hen writing an article on the distribution of wealth in the UK, I noticed a serious discrepancy between the contemporary concentration of wealth described in Capital in the 21st Century and that reported in the official UK statistics. Professor Piketty cited a figure showing the top 10 per cent of British people held 71 per cent of total national wealth. The Office for National Statistics latest Wealth and Assets Survey put the figure at only 44 per cent.
Whoa! Piketty’s figure for wealth concentration in the present–so we’re not quibbling over 1810 here–was off by 27 percentage points. Like Will Ferrell, that’s kind of a big deal.
It gets worse. This chart showing US data is an example:
Remember, the whole thrust of Piketty’s book is that there is inexorable income and destructive income inequality under capitalism, and his book has the data and theory to prove it. But, well, you see the blue line above, which is Piketty’s line purporting to show the “wealth share” of the “top 10% over the last two century’s? Turns out that for 90 of those years, between 1870 and 1960, the data doesn’t even exist. (That’s the gap in the red lines, above.)
So how did he produce his continuous blue line?
According to Giles, Piketty literally just made that blue line up (presumably guided by the trends in the other countries, and in the US 1%, for which there were better data).
Giles describes several categories of issues that he found with Piketty’s data:
a) Fat fingers
Prof. Piketty helpfully provides sources for the data he uses in his work. Frequently, however, the source material is not the same as the numbers he publishes. …
On a number of occasions, Prof. Piketty modifies the figures in his sources. This might not be a problem if these changes were explained in the technical appendix. But, with a few exceptions, they are not, raising questions about the validity of these tweaks. …
Prof. Piketty constructs time-series of wealth inequality relative for three European countries: France, Sweden and the UK. He then combines them to obtain a single European estimate. To do so, he uses a simple average. This decision (shown in the screen grab below) is questionable, as it gives every Swedish person roughly seven times the weight of every French or British person. …
d) Constructed data
Because the sources are sketchy, Prof. Piketty often constructs his own data. One example is the data for the top 10 per cent wealth share in the US between 1910 and 1950. None of the sources Prof. Piketty uses contain these numbers, hence he assumes the top 10 per cent wealth share is his estimate for the top 1 per cent share plus 36 percentage points. However, there is no explanation for this number, nor why it should stay constant over time.
There are more such examples. …
e) Picking the wrong year for comparison
There is no doubt that Prof Piketty’s source data is sketchy. It is difficult to find data that relates to the start of each decade as his graphs demand. So it is only natural that he might say 1908 is a reasonable data point for 1910 on the graph.
It becomes less reasonable when, for example, Prof. Piketty uses data from 1935 Sweden for his 1930 datapoint, when 1930 data exists in his original source material. …
f) Problems with definitions
There are different ways to compute wealth data ranging from estimates based on records at death to surveys of the living. These methods are not always comparable.
In the source notes to his spreadsheets, Prof. Piketty says that the wealth data for the countries included in his study are all obtained using the same method. …
But this does not seem to be true.
g) Cherry-picking data sources
There is little consistency in the way that Prof. Piketty combines different data sources.
Sometimes, as in the US, he appears to favour cross-sectional surveys of living households rather than estate tax records. For the UK, he tends to avoid cross sectional surveys of living people.
Prof Piketty’s choices are not always the best possible ones. A glaring example is his decision relative to the UK in 2010. The estate tax data Prof. Piketty favours comes with the following health warning.
“[The data] is not a suitable data source for estimating total wealth in the UK, or wealth inequality across the whole of the wealth population; the Wealth and Asset survey is more suitable for those purposes”.
These choices matter: in both the UK and US cases, his decision of which type of data to use has the effect of showing wealth inequality rising, rather than staying constant (US) or falling (UK).
This appears to be the key chart, says John Hinderaker at Powerline, “particularly the two graphs on the bottom. It is apparent that the superficial plausibility of Piketty’s account derives from his own ‘tweaks’ and misrepresentations, not from the underlying data”:
Unbelievably, there is still worse, leaving Piketty’s defenders to face
the awkward fact that Giles has arguably just shown that when you correct Piketty’s factual mistakes, then the trend in both the UK and the US is the exact opposite of what everyone took away from the book.
And while these countries represent two of the major sources for his argument, his data for the rest of Europe looks equally dodgy.
Nick Sorrentino at Against Crony Capitalism reckons that the fudging of data is actually unnecessary.
He didn’t have to fudge the data however, Piketty is fundamentally wrong about how capital snowballs and why the wealthiest have increased their share of wealth in the USA since about 1970. In a nutshell it is because of the financialization of the economy enabled by a truly fiat currency… Easy money benefits the bankers and the politically connected first, the 1%.
But Piketty has failed to even notice this valid point, leaving him, as Bob Murphy argues, wrong in data and wrong in theory.
The book contains foundational theoretical problems, a misreading of the empirical literature that blows up his whole case, sloppy and absurd factual errors concerning tax rates and minimum wage hikes, and shocking quotations that reveal he has no desire to actually raise government revenue with his massive soak-the-super-rich schemes, but instead merely wants to prevent the formation of fortunes in the first place.
Because even on its face, Piketty’s theoryas stated doesn’t hold water.
You have to be ignorant as the day is long not to know that capitalism has made us wealthy beyond all possible expectation, even going back thirty years never mind three hundred. We now have a vast number of people who do not work because we produce at such a prodigious rate that it just doesn’t require more than about a quarter of the working population to produce enough for us to maintain a 1950s and better lifestyle for those who choose not to bother actually earning an income. In our society who hasn’t got a phone, a car, a colour TV, enough to eat, clothes to wear and a place to live. There are always people on the fringe who circumstances have dealt a bad card, but really we are beyond any issue of deprivation that had existed for the entire course of human history up until say around that same 1950s mark.
So Piketty lied. The people who line up behind the book will care about that as much as they did about Climategate. It is about power and wealth, with the facts of the case as close to a non-issue as it is possible to be. The only interesting question about wealth distribution to these people is that they would like more of our wealth distributed to themselves.
As Steve Horvitz observes, “Piketty has a potential problem here and we need to see what his response is. His first round reply in the Financial Times was a whole bunch of nothing.” And it’s not enough to respond on his behalf, as many have already done, saying “but look, he’s so honest he’s put all his data on line. Because as Tyler Cowen points out, the question is not whether or not Monsieuer Piketty is a nice guy (he may be, but it’s irrelevant), nor whether he’s more honest than Michael “Hide the Decline”Mann (which wouldn’t be hard): the real question at stake is whether or not he’s right. Because if he is, or can be made to appear that way, then politicians the world over are ready and willing to swing a global wealth tax your way.
So this matters.
LINKS REFERRED TO:
- Data problems with Capital in the 21st Century – Chris Giles, FINANCIAL TIMES
- Piketty findings undercut by errors – Chris Giles, FINANCIAL TIMES
- Piketty response to FT data concerns – Thomas Piketty, FINANCIAL TIMES
- Is Thomas Piketty a Fraud? – John Hinderaker, POWERLINE
- Piketty Tricketty & Historical US Wealth Data – PHILIP W. MAGNESS
- Phil Magness on Piketty's U.S. Wealth Data – David Henderson, ECON LOG
- Are You Kidding Me? Now Allegations That Piketty Fudged His Wealth Data – Bob Murphy, MISES CANADA
- Piketty and scientific fraud – Steven Kates, CATALLAXY FILES
- What do the Piketty data problems really mean? – Tyler Cowen, MARGINAL REVOLUTION
- Piketty’s Data, Krugman’s Shame – Perry Metzger, SAMIZDATA
- The Piketty Controversy – Steve Horwitz, COORDINATION PROBLEM
- More Picking on Piketty – POWERLINE
- Piketty and the Shoe Event Horizon – Perry Metzger, SAMIZDATA
- Criticisms of Piketty – James Hamilton, ECONBROWSER
[Hat tip for cartoon to Steven Kates]