The case of Dutch social psychologist Diederik Stapel fraud, now in the news, which led to the retraction of several of his papers by academic journals suggests that this might be the right course of action for the American Economic Review (AER). Even if Reinhart and Rogoff's (RR) results do not necessarily amount to fraud, something that I'm sure could become a matter of dispute, it's still a fact that they are incorrect, as admitted by the authors. So the AER should clear the record and retract the paper that suggests that growth collapses when a country has a debt-to-GDP ratio of more than 90%.
PS: As the NYTimes notes there is a blog about scientific papers that are retracted here. The blog dealt with RR case here.
PS: As the NYTimes notes there is a blog about scientific papers that are retracted here. The blog dealt with RR case here.
Hi Matias,
ReplyDeleteAlthough I do like your blog and your thoughts very much, I think you are going a little too far in demmanding that the paper be retracted. After all, science is supposed to progress by error.
I think the problem with the paper is not the excel error, but the underlying neoclasscial assumptions (interest rate set by the market/crowding out/loanable funds) that make people expect a causal relationship between public debt and low growth.
If you wanted to purge serious economic journals or institutions of such thinking, there would be hardly any left.
Papers with errors (irrespective of the causes of the error) are regularly retract in sciences. Economists should do it too. There is no evidence to support the conclusion in the paper that above 90% debt-to-GDP ratio output growth collapses. I can see the point of not doing it. In fact, I would kind of respect them if they asked for retraction like these guys http://retractionwatch.wordpress.com/2013/04/26/a-model-retraction-in-the-journal-of-neurochemistry-for-unexpected-effect-of-a-filter/
Delete