The Washington Post

Is Piketty’s ‘Capital’ full of mistakes?

Thomas Piketty (LEON NEAL/AFP/Getty Images)

The nine most terrifying words in the English language for a researcher are: "You made spreadsheet errors like Reinhart and Rogoff did." But that's what Chris Giles of the Financial Times thinks rockstar economist Thomas Piketty did, among other mistakes, in his groundbreaking new book, "Capital in the Twenty-First Century."

How serious are these problems? And how much, if at all, should they change what we think about inequality?

Well, the answer seems to be not too serious, except for the recent British numbers. Let's go through them. Giles identifies three basic types of issues. The first are simple transcription errors. He finds, for example, that Piketty accidentally entered Sweden's wealth data from 1908 instead of 1920. But despite this mistake, Giles concludes that "there appear to be few problems with the choices made by Prof. Piketty for Sweden"—which you can see in Giles's chart below, comparing his corrected calculations (red) with Piketty's (blue).


In other words, there are data-entry errors, and they're embarrassing, but they don't change the big picture.

The next concern is methodological. Giles thinks Piketty should average European data by population, not by country. He doesn't like that Piketty draws trends between large gaps in the numbers. Or that Piketty labels data from, say, 2004 as "2000" on some of his charts. And Giles isn't sure why Piketty has put together some of his wealth data—which is sparse, and needs to be adjusted, if not constructed—the way that he has. But these aren't errors. They're questions. Ones that Piketty should answer, but still just questions.

The last problem is the most significant. Giles points out that Piketty seems to have mixed up different sources on British wealth the last few decades, and overestimated their inequality. You can see Piketty's numbers (blue) versus the raw data (red) below. This does need to be explained.


Bu this doesn't seem to be a Reinhart and Rogoff situation. Their Excel errors really did change their conclusions. Piketty's don't. Unless, like Giles, you average inequality by population instead of by country—which is debatable, at best, since Piketty is only concerned about inequality within countries.

Still, it's another reminder that economics is the least exact science. That's why it's critical for researchers to make their data public, so everyone else can double and triple check it. To his credit, Piketty has done just that. Now he should tell us a little more about his methods. (And don't worry, that won't bore us. We were already willing to read 700 pages about the history of inequality.)

Matt O'Brien is a reporter for Wonkblog covering economic affairs. He was previously a senior associate editor at The Atlantic.



Success! Check your inbox for details. You might also like:

Please enter a valid email address

See all newsletters

Show Comments
Most Read



Success! Check your inbox for details.

See all newsletters

Your Three. Videos curated for you.
Play Videos
Sleep advice you won't find in baby books
In defense of dads
Scenes from Brazil's Carajás Railway
Play Videos
For good coffee, sniff, slurp and spit
How to keep your child safe in the water
How your online data can get hijacked
Play Videos
How to avoid harmful chemicals in school supplies
Full disclosure: 3 bedrooms, 2 baths, 1 ghoul
How much can one woman eat?
Play Videos
What you need to know about Legionnaires' disease
How to get organized for back to school
Pandas, from birth to milk to mom
Next Story
Jason Millman · May 23, 2014

To keep reading, please enter your email address.

You’ll also receive from The Washington Post:
  • A free 6-week digital subscription
  • Our daily newsletter in your inbox

Please enter a valid email address

I have read and agree to the Terms of Service and Privacy Policy.

Please indicate agreement.

Thank you.

Check your inbox. We’ve sent an email explaining how to set up an account and activate your free digital subscription.