Convincing empirical ... and theoretical work

We are currently in the middle of the junior job market, and it is refreshing to read from Paul Krugman:

solid empirical evidence, even of the complicated econometric sort, changes plenty of minds.

I have a few quibbles. Surely some — perhaps all too many — economists are indeed locked into ideologically motivated beliefs. Consider the response of fresh-water macroeconomists to the utter failure of their predictions about inflation; who other than Narayana Kocherlakota has made the slightest concession to the people who got it right? I’m also skeptical about the persuasive power of complicated econometrics; my sense is that mind-changing empirical work almost always involves not much more than simple correlations, usually from natural experiments — that is, even multiple regression turns out, in practice, to be too complicated to persuade.

On the other hand, I would argue that empirical work isn’t the only thing that can change minds: really clear analytical arguments can do it too, by letting economists see things that were in front of their noses but overlooked because they didn’t have a framework.

 OK, it is time to go back to a seminar room, hopefully for a convincing empirical &/or theoretical work ;-)