Financial Times: Expert Review Finds That Thomas Piketty's "Capital in the 21st Century" Used Erroneous Numbers in Making Its Predictions; When Correct Numbers Are Used, His Central Thesis Fails to Eventuate

It gets better.

His defense gets better, I mean.

Some issues concern sourcing and definitional problems. Some numbers appear simply to be constructed out of thin air.

But, according to a Financial Times investigation, the rock-star French economist appears to have got his sums wrong.

Piketty's main "contribution," and I use that term advisedly, was attempting to prove that as the rate of return on investment in a country exceeded the growth in GDP, it tended towards a positive feedback loop (similar to that postulated in Global Warming) and caused even further gains in passive investment wealth growth while general wealth stagnating.

Basically, it's The Rich Get Richer proved (supposedly) by rigorous statistical analysis. And this then supports Piketty's conclusion that Something Must Be Done.

But when the correct numbers are used, his formulas demonstrate no rise in income inequality in, for example, Europe -- with the right numbers, Europe shows no tendency towards growing inequality since 1970.

Okay, here's where it gets very Global Warming-esque indeed:

Contacted by the FT, Prof Piketty, said he had used “a very diverse and heterogeneous set of data sources ... [on which] one needs to make a number of adjustments to the raw data sources."

The whole article is worth reading, but FT keeps putting up warnings about cutting and pasting, so I don't want to include more than that.

FT is pretty serious with its firewalls, so a lot of their previous articles on this can't be viewed, but you can watch the author of the article cited here discussing the problems with Piketty's math in this video.

Skip to 0:52-1:30 to see Giles discussing Piketty just adding a "2" to a formula for no apparent reason, because, Giles speculates, the formula "didn't seem high enough" without adding a little somethin'-somethin'.

Now maybe there is some justifiable reason for adding the two, but apparently Piketty does not alert people to the addition, nor explain his rationale for adding it; his claims are, therefore, not transparently arrived at, and of dubious value.

You have to be able to replicate a scientific experiment or, in this case, an ostensibly scientific analysis. An outside observer needs to be able to evaluate each estimation, assumption, and "adjustment to the numbers."

Beyond just that, though, the whole video is pretty devastating. Giles shows that, contra Piketty, there is no rising income inequality in the UK, and there simply isn't enough data to say anything much about the US.

Piketty's predictions -- or, I guess, his scenarios for helping you think about the future -- seem, at the moment, to be based on cherry-picked numbers and assumptions pulled from a dreamjournal.

Thanks to LoneStarHeeb.

A Very Heterogenous and Diverse Set of Data Sources

...proves that I cannot be blamed for what eventuated in Your Mouth.

posted by Ace at 03:03 PM

| Access Comments