Which Economic Models Are Más Macho?

A few days ago Alex Rosenberg and Tyler Curtain wrote an op-ed titled “What Is Economics Good For?” In a nutshell, their answer was “not much.” Paul Krugman begs to disagree:

Rosenberg and Curtain completely misunderstand what’s been going on at the Fed. They also misunderstand the nature of economists’ predictive failures. It’s true that few economists predicted the onset of crisis. Once crisis struck, however, basic macroeconomic models did a very good job in key respects — in particular, they did much better than people who relied on their intuitive feelings….Wonks who relied on suitably interpreted IS-LM confidently declared that all this intuition, based on experiences in a different environment, would prove wrong — and they were right. From my point of view, these past 5 years have been a triumph for and vindication of economic modeling.

Something about this passage has been niggling at me since I read it yesterday, and I just now figured out what it is. Krugman has been banging this drum for quite a while, and regular readers know that I’m basically on his side. Basic Keynesian macro has done a pretty good predictive job in the aftermath of the financial crisis, and it’s fair to wonder why skeptics continue to be skeptics even after years of solid results from textbook macro.

But here’s the thing: I’m on Krugman’s side in hindsight. A better question is whether it was obvious in 2008 that “suitably interpreted IS-LM” was likely to be the best model for dealing with the post-crisis recovery. Maybe it was. Krugman makes the case, for example, that RBC models should have been abandoned decades ago for not fitting the data. But conservative economists would argue that Keynesian macro was quite justifiably thrown out even earlier for failing during the 70s. That’s obviously a matter of contention, but it’s certainly the case that the Keynesianism of the 70s has since been retooled into the New Keynesianism of the 90s and beyond. But that makes it a fairly new theory. So again: how obvious was it before the fact that Krugman’s preferred models were likely to be the best ones for 2008-13?

This is light years above my pay grade, so I’m throwing it out mostly in the hopes that some real economists will essay an answer. I’m not even sure I’m framing the question entirely properly. But the basic problem is that economists change their models the way most of us change our television viewing habits, and the best models often seem to be very dependent on a particular place and time. Wait a couple of decades, or examine a different kind of economy, and suddenly the old models don’t work so well anymore. So how do we know in advance? Can Krugman legitimately say that his models have had a long track record of success in different environments, and therefore should have been the obvious incumbents when the economy went kablooey in 2008?