Wednesday, July 05, 2006

Freakonomics Cultists Must Die

What in the world is the deal with all this "Freakonomics" worship?

I am, as Cheech and Chong famously put it, sick up and fed with all of this Freak-o crap.

It's a perfectly good book. I admire Levitt. He is a fine man. Very smart, and may bring Economics as a discipline back toward an empirical focus. (I'm sure Stephen Dubner is also wonderful).

But Freakonomics is simply arm-chair, superficial, first-thought-that-comes-into-your-mind social science. (Yes, it is TESTED empirically, but not all alternatives occur to you in the first thought that comes into your mind).

At the IHS conference, I talked a bit about several results in political science and econ. Two different students demanded that I immediately explain why all the journal article results were different from the Freakonomics results.

(For example, the Freak-o claim was that campaign spending doesn't matter, at all. That's nonsense. People have shown that both challenger and incumbent, but especially challenger, spending matters a lot. SL is flat wrong about that. Plus, he mismeasures things, and ignores the fact that he would need an equilibrium model to solve the endogeneity of spending).

Now, I love Freakonomics. The emphasis on testing, the idea that we can learn things from empirical work...all great.

But one of these kids at the Stanford conference demanded that I come with him and go over the Freakonomics claims. I said I wasn't interested. He said, "Why are you so afraid of the truth?"

Gosh, fella, I've studied this question myself, and the literature on this question, for 20 years. There are dozens of different papers that show that Levitt has this wrong. And the reason is simple: he has misspecified the model, in ways we have understood for decades. I have no interest in explaining in detail why Levitt is wrong in this instance, at 11 pm when we are having cocktails. I am particularly uninterested in explaining it to a junior grad student who is proud (and rightly so!) of the great depth of his ignorance of political science. That makes me AFRAID?

(SIDENOTE: Does that argument actually work for you sometimes, kid? Is that how other people got you to drink your first beer in high school? "Do it or we will call you 'big chicken-face'?" You need to get out more, pumpkin.)

More and more people have this belief that Levitt is some kind of prophet, a god-like figure. Levitt does not present himself that way in person, and makes NO such claims in his book. He just wants to make the read think. So it is not HIS fault.

He got a lot of profit (I'm punning on "prophet", you see) out of Freakonomics, sure. But much of the work in that book is simply provocative, and intentionally so. It is designed to make the point that it is fun and interesting to do your own tests. I agree with all that.

But it is not surprising that he got a lot of things wrong. That is one of the great things about empirical work: I *know* he got it wrong, and I can prove it. Challenger spending definitely matters, even controlling for challenger charisma. And the data are not IID, because in equilibrium it is true that charismatic challengers get more money. But that doesn't mean the spending doesn't matter.

Anyway, bottom line: Steve Levitt--way to go.

Young goofballs who think that Levitt's empirical work is perfect, and doesn't itself need to be tested--you are missing Levitt's own point.

This is statistical inference, not religion class.

UPDATE: I should be more careful. On rereading, my interpretation of the book is unduly harsh, and inaccurate. SL does not really say campaign spending doesn't matter at all. In particular, he never makes the equilibrium claims I attribute to him above. Those claim were made by his self-appointed hit men in my audience. Mea culpa. The hit men can still bite me, however.

I would edit the post, but that would violate basic blogger ethics, since someone may have seen, and linked to, the earlier (admittedly incorrect) version. So this update will have to do.

6 comments:

Anonymous said...

I think your point in your correction is the point. Science, like politics, is about discussing and analyzing an issue. Not necessarily to reach agreement, but to understand the strengths and weaknesses of an idea (and perhaps to recognize that you may change your view upon reconsideration). The problem of your encounter can be seen in the larger society. Al Gore is profitatyzing (I'm riffing on your pun, sorry) on global warming, and the title of his movie says it all: "An Inconvenient Truth". Why is his theory the "truth"? Perhaps an attempt to shift the burden of supporting a theory from those who propose it to those who may disagree with it? Similar to the mike munger experience, the student wanted to shift the burden of supporting his (incorrectly) borrowed theory to someone who disagrees with it. And like Al Gore, if you have the burden and desire to disprove my theory, then you must be clearly morally reprehensible.

I think the larger question is who created these students and how are they getting into graduate school where the whole endevour can be summed up in "cogito, ergo sum". In this case the student wanted you to do the thinking for them, and clearly had been rewarded for it in the past. Your failure to do this was clearly a victory in his book. I just hope he's not the one who will be teaching my kids one day.

Anonymous said...

Sounds like you and that grad student may have something in common - you don't read carefully?!?!

mungowits said...

it's worse: I don't even WRITE carefully. How embarrassing....

Anonymous said...

Careless as you may be, at least you write a lot from what I can tell.

mungowits said...

"At least you write a lot..." AAAARGH! Worse and worse!

Reminds me of the Groucho Marx line. "The food here is terrible. Fortunately, the portions are small...."

Anonymous said...

Mike:

You forgot about the part where he was also demonstrably wrong on the abortion - crime arguement. See:

http://online.wsj.com/public/article/SB113314261192407815-HLjarwtM95Erz45QPP0pDWul8rc_20061127.html?mod=tff_main_tff_top

--Scott