The failure of modern macro

Thomas Sargent complains that the arguments for the fiscal stimulus “ignore what we have learned in the last 60 years of macroeconomic research.” This is his fault. He and the other big names of modern macro have failed to produce a simple model (or a series of simple models) to replace the Keynesian one in undergrad textbooks.

Policy makers, and even economist that don’t specialize in macro, were taught to think about the economy like Keynes did 75 years ago. They shouldn’t be expected to absorb cutting edge research only taught in grad schools. The problem is that this stuff is no longer cutting edge. The rational expectations revolution is nearly 35 years old.

In macro, there remains a gearhead culture. To understand 30 year old models you need to learn overly complex jargon and graduate level mathematics. There still isn’t a standard tool box for solving models… you’re expected to program everything from scratch. Serious macro job candidates are expected to know how to prove fixed point theorems and its not enough in your job market paper to have an interesting economic problem (if that’s even necessary at all), but you have to extend some already complex model. Those extensions have to be “tractable” but not necessarily “understandable” (its not clear to me the point of the former without the latter). This means wildly unrealistic assumptions that happen to make the math work, even if they contradict each other in spirit, are ok.

I actually don’t mind all that stuff. That’s why I signed up to do it. Someone, though, needs to take stock of the economics — rather than neat mathematical theorems or cool econometric techniques — modern macro has taught us and write it down in an undergrad textbook.

3 Responses to “The failure of modern macro”

  • Gabriel says:

    The major achievement of modern macro is to show us how little we know (about the trend, the fluctuations and the big depressions).

    Neither Bernanke nor Cole&Ohanian can provide you with a model that can quantitatively account for the length of the great depression.

    Endogenous growth and its flirting with IO and so on have not elucidated trend growth rates and account for the 25x international differences.

    You can sort of approach fluctuations with your favorite mix of shocks and rigidities or whatnot, but to solve the model you have to have policies are rules but real policy making is not rule-based. — The policy advice that can come out of these models is what rule is best, but policy makers don’t need/want that.

  • pushmedia1 says:

    I guess, then, THE great achievement of the enlightenment was show us how little we know? It is true that the Church thought we knew everything (“god did it”) and science showed otherwise, but I wouldn’t say it was THE great achievement.

    Yeah, we used to think because some dude (Keynes) had a theory and wrote a book (General Theory), we knew how the economy worked. Modern macro gave us an empirical tradition and we started realizing lots of those theories weren’t true. That’s hardly the absence of knowledge.

    The existence of questions doesn’t necessarily imply the absence of knowledge. We know what we don’t know and we have the tools to begin to find answers. Why can’t we teach that?

  • swong says:

    The existence of good questions, in fact, implies a good foundation of knowledge. Teachers rarely recognize this.