O Financial Times pôs em causa os dados utilizados por Thomas Piketty no livro Capital in the Twenty-first century, numa crítica que fez imediatamente lembrar a polémica em torno do Excel de Rogoff. Agora, o próprio Piketty respondeu.
E, pelos vistos, o debunking do FT foi um tiro ao lado: os ‘erros’ dos dados Piketty terão sido, na verdade, ajustamentos aos dados de base devidamente fundamentados no apêndice técnico do livro – ajustamentos normais no tratamento de dados desta natureza, mas de que Chris Gilles, do FT, não terá dado conta. Entretanto, as duas partes parecem ter chegado a um quase-consenso.
Sobre esta polémica ler: A new critique of Piketty has its own shortcomings (New York Times), Did Piketty get his math wrong? (New York Times), Thomas doubting refuted (Paul Krugman), Piketty’s errors aren’t errors: they are questions, and he answered them (Matt O’brien), e Mistakes e What the Financial Times got very wrong, de Simon Wren-Lewis. O último tem uma meta-lição que vale a pena citar.
When an academic, or student, thinks they have found a mistake in an academic paper or book, what do they do? Check their calculations again and again, or course. Ask someone else to do the same, maybe. But then they will write to the authors of the original work, and ask them to comment. What they will notdo, in that letter or email, is to give the original author a deadline of one day to respond. That was how much time Chris Giles of the Financial Times gave Thomas Piketty to respond to his long list of alleged errors and unexplained adjustments.
I think it might have been very different if Chris Giles had written a piece about the difficulty of interpreting wealth inequality data, and had wanted to get clarification of what Piketty had done and why. I suspect in that case the paper would have given Piketty more time to respond (what was the urgency?), and the article would have benefited greatly from that dialog.
But that was not the article that Chris Giles chose to write and the Financial Times chose to publish. Instead they wrote an exposé, in much the same way as you would expose some wrongdoing by a politician. (Is an academic making a spreadsheet error the equivalent of a politician having an illicit affair?) The phrase they use in football is playing the man and not the ball.
Now, in the unlikely event that I ever warranted a headline story, I know I would not want to be treated in the way Giles treated Piketty. There were only two possible justifications for writing a story of that kind. One was if the paper had clear evidence that Piketty had fiddled the numbers to get the results he wanted, and it is obvious they did not have that evidence. The other is that they had found so many simple mistakes that this discredited Piketty as an academic. Again this was not the case.
I also get very cross with academics who suggest that, because his book had become a bestseller and he had accepted invitations to talk to White House staff, he somehow deserved this kind of treatment. This seems to me like hypocrisy at its worst. Given this treatment, both Thomas Piketty’s initial response and his more detailed response issued yesterday are remarkable and impressive in their restraint.