Friday, January 16, 2009

Brooks vs. Quantum Mechanics

Brooks channels what-the-bleep in favor of behavioral economics:
Once there was just Newtonian physics and the world seemed neat and mechanical. Then quantum physics came along and revealed that deep down things are much weirder than they seem. Something similar is now happening with public policy.

Quantum mechanics is just as "neat and mechanical" as classical mechanics, which is why it caught on. It was clear from the beginning how to do quantum-mechanical calculations, and in what regimes the answers would limit to Newtonian physics, which clearly described the everyday world. This is why quantum mechanics is a terrible analogy for behavioral economics. As far as I know, all the behavioral work so far consists of a bunch of trivial psychology experiments that have no coherent message; there are no quantitative or even qualitative predictions; and it is not clear that behavioral "theory," such as it is, limits to classical theory when it should (and when is classical economics true, exactly?).

Oh, and I don't know what this means:
Mechanistic thinkers on the right and left pose as rigorous empiricists. But empiricism built on an inaccurate view of human nature is just a prison.

But facts are facts, no matter what human nature is. An empirical finding is, say, that when you give a pretentious idiot a column at the Times, he tends to spew pretentious garbage. Said pretentious idiot might be driven to this by difficult and ill-understood forces, but the fact that it happens is just a regularity of nature. You don't, after all, need to get the chemical structure of the air right to realize that clouds portend rain.


Grobstein said...

Durf durf no one tell this guy about Godel's Theorem.

Grobstein said...

And I rather agree about your assessment of "behavioral economics," although it seems conceivable that we will eventually get new models that incorporate the insights of those experiments.

Sarang said...

Oh I agree it might go somewhere. (Or not. I'm a little skeptical that it'll ever yield much more than common sense.) At this point though it's just an excuse for pundits to replace half-baked economic analysis with half-baked theories about human nature.

icety said...

Never really understood your "common sense" point. A lot of psychology / behavior economics wasn't common sense until they did the "trivial" studies. I mean, a lot of psychology is really wishy-washy, but so is a lot of neuroscience. You have to start somewhere.

icety said...

As for the "half-baked" economic analysis—you need some foundation to make policy decisions (unless you're Ron Paul, in which case all economic problems stem from the existence of the Fed, which should be abolished so that Americans can go back to the gold standard and barter sheep and chickens with each other like God intended). I share your frustration with all the hand-waving in a lot of economic models that drive policy, but it's the best we have. Putting that together with what we know from behavior economics (such as how to structure tax rebates/stimuli so they're more like to be spent instead of saved) doesn't seem terrible to me.

Sarang said...

Re common sense: the sunk cost fallacy, loss aversion, virtually everything in Brooks's article, etc. were familiar long before the 20th cent. The experiments would have added to our knowledge if they gave reliable quantitative predictions but they don't seem to.

Re half-baked econ: I was talking about David Brooks. I wasn't calling the entire discipline half-baked (though I'm not saying it isn't, either). Obviously policy must sometimes be made in the dark, but that doesn't mean David Brooks isn't a fuckwit.

I disagree that "you have to start somewhere" means you have to be wishy-washy. Mechanics as of 1690wasn't wishy-washy, just limited in the range of things it could talk about. It always had _some_ surprising, testable predictions. The downside to looking for new phenomena is that they might not have policy relevance, but scientific progress depends on finding the right problems and answering them correctly, not doing a half-assed job on whatever others consider important -- rolling balls down slopes rather than predicting the weather.