Sunday, May 10, 2009

Animal Spirits, Vegetable Math

Benjamin Friedman, in the NYRB, on what's wrong with behavioral economics:
The more important question is whether what Akerlof and Shiller have offered in Animal Spirits amounts to "a theory" in the sense that it could stand in place of the current theories that they criticize for being based entirely on rational responses to economic motives. There is a difference between a series of ideas about different aspects of economic behavior and an integrated account of macroeconomic fluctuations. Akerlof and Shiller are surely on the right track in pointing to elements that are missing from today's conventional models, and in arguing that incorporating them into mainstream macroeconomic analysis would help. But they have neither done this nor shown others how to.

In fact, it's worse than that. Here's Friedman a few paragraphs earlier:
[F]or purposes of macroeconomics—the study of the economy as a whole—most of the standard models do not admit the possibility of unemployment. The reason is not that no one knows unemployment exists. Rather, no one has figured out how to allow for it within the confines of sufficiently simple mathematics; and faced with the choice between excluding unemployment and sacrificing analytical simplicity, most macroeconomists have opted for the former.

But "Opted" suggests they had more of a choice about it than I think they did (or do). Tractability is essential; without it, you have laundry lists of effects that you can't quantify, like Akerlof and Shiller. On the other hand, if the price of tractability is complete irrelevance, that's problematic too. What makes people think macroeconomics is solvable?


Grobstein said...

Macroeconomics has usually been the shortest route from economics to power.

Yeah I agree that these are too-rarely acknowledged shortcomings of the behavioral economics "revolution."

dmnorton said...

Arguing that certain behavioral facts need to be recognized, rather than about how to implement them, also creates a convenient straw man: the economist who doesn't know that behavioral features are missing. But it's a useless critique when the real problem is that nobody knows how to incorporate behavioral insight in a consistent and useful way.

Sarang said...

Yeah, I agree. On the other hand, it's reasonable to claim that standard theory is sufficiently far off in its predictions -- e.g. Newton's laws without friction are a terrible way to describe motion in viscous fluids -- that it's entirely irrelevant.

Grobstein said...

I'm reading that review now. It's really striking that such brilliant guys apparently don't realize how bad their arguments are when they step outside their narrow topical comfort zone.

It's actually genuinely weird to me.