Monday, July 24, 2006

Weakonomics


Freakonomics, A Rogue Economist Explores the Hidden Side of Everything, by Steven D. Levitt and Stephen J. Dubner, Read by Stephen J. Dubner, HarperCollinsPublishers, Inc., 2005


The original title of this blog came to me one day when I had been kicking around the idea for this space for about a week. I didn't love the title, but I went with it anyway, because it was better than nothing. The new name that occurred to me nearly a year later was PostHype because that's really the idea behind what I was aiming for. Maybe you read a book review when something just came out and it sounded really good and you saw it's name everywhere, but for one reason or another you just didn't get around to it and you forgot. Or maybe you thought maybe you might like to get around to reading something, but all the hype surrounding its initial release kind of put you off. (I know that shit repels me).

Well, the idea I had was that I would either remind you, hey, that book everyone was talking about in January? It's really good. Or I'd comfort you, hey, that book everyone loved in January? What a pile of shit.

Which is kind of the message I come bearing today. No, Freakonomics is not quite a pile of shit, but it is nowhere near matching its hype and its hype was huge and everywhere. Now, I'm not an economist nor even a scientist by training and my exposure to economics theory consists of taking micro and macro 101 classes over ten years ago (classes I did remarkably poor in), but what I do challenge are Levitt and Dubner's sociological assumptions (a class I did well in, if memory serves).

In a variety of little ways, the authors present interesting information, but some of their assumptions require them to assess all data as value neutral facts of reliable provenance when anyone could tell you that people lie; they lie all the time. How many books do you read per year? How much TV do you watch? On average how long does it take you to reach climax during sexual intercourse? How much do you drink? Everyone fudges these numbers if asked in a survey, or nearly everyone.

Thus the entire chapter on the negative correlation between parental behavior and a child’s school performance and the positive correlation between parental identity and a child’s school performance is an apples to oranges comparison. Your income level is a fairly easy number to quantify and verify, likewise the mother’s age at her first child’s birth and whether or not the parents speak English. Fuzzier numbers that don’t appear to Levitt to have much bearing on students’ academic performance include how often you spank, how often you take your kids to museums, how much television you let them watch, and how many minutes in the day you spend reading to them. All the kinds of things, sitting face to face with a governmental interviewer, many parents might be likely to fib about. In this sense, it is as if Levitt is comparing facts with opinions. I can't imagine a scenario in which not reading to your kids would be just as effective on influencing their academic careers as reading to them daily.

It is curious that Levitt calls himself an economist, as he clearly admits that he is bad at math, knows little economic theory, and is unable to answer any particular questions about whether the stock market will go up or down, whether deficit spending can be beneficial, or what the effect of taxes would be. That strikes me as somewhat the point behind being an economist. It's like saying I'm an English major, but I don't like to read, don't understand basic grammar, and couldn't explain anything about Shakespeare, Milton, Hemingway, or Faulkner.

As for the suggestion that he’s some kind of “rogue economist” I hardly know what to make of that peculiar appellation. Seeing as little of Levitt’s work actually deals with economics per se, I’d imagine “rogue statistician” is more likely a better choice. Why rogue too? Is there some kind of trend he’s bucking or bulkhead of economic theory he’s striving against? No, not really. Levitt just thinks it more interesting to apply statistical analysis towards topics that haven’t yet really been given the spreadsheet once over. The whole subtitle thing is a lot like much of the book’s writing and packaging, which is maximal intrigue with limited pay off. How are Sumo wrestlers like public school teachers? What’s so similar between the Ku Klux Klan and real estate agents? The book jacket and introduction suggest tantalizing possibilities with these wild pairings.

Well, it turns out both of the first parties have cheaters in their midsts. Good heavens, what an insight! I’ll bet garbage men and stock traders are surprisingly alike in that both procrastinate on unpleasant household chores. And how is the Klan like my real estate agent? They both have code words that sound innocuous but are really charged with subtext. What a groovy groovy insight. The story Dubner and Levitt tell about Stetson Kennedy who infiltrated the Klan and passed on all their secrets to the makers of the Superman radio show for inclusion into the script is amusing and well-told, the KKK going crazy when all the children near them start spouting their passwords in games, but the final summation of this essay “Knowledge is power” is obvious and hardly a bolt of lightning.

If there are any particularly enjoyable pieces both for their peculiar insights and their novelty it would be the look at the internal economics of a crack gang and their analysis of whether or not obviously ethnic names lead to punishment in the job marketplace. In the latter case, most of us have heard of the study indicating that identical resumes sent out for LaTisha Moore and Amy Moore ended in a disproportionate number of interviews and call backs for the more Caucasoid name. Levitt and Dubner use California Social Security rolls to slice and dice the data and provide not only some surprising results but also a funny and enlightening trip down baby naming trends of the decades past and present. The trickle down effect of upper class naming inclinations as they reach the lower classes ten years later is particularly comical. If what they’ve predicted comes true, we’re in for some even more fucked up names than are on display even now.

Crack gang economics is appealing and interesting in that the common misconception is that there is mad money to be made as a dealer, when in reality most in the game pull down actually a little less than what they’d get flipping burgers. Levitt’s source for all this is a former MBA turned dealer who happened to organize his gang on corporate principles as well as kept detailed accounts of all cash flow. At first it’s rather eye-opening to suddenly grasp that low-level street corner dealers make quite specific economic choices on wages versus gunshot wounds.

Then you realize that Levitt and Dubner are pushing a very specific though rather plain idea throughout all of these essays, an idea which is somewhat sound if absurdly reductive: humans seek to maximize the good for themselves and minimize the bad. The single greatest difficulty with such a one-size-fits-all argument like this is that aesthetically it allows for such a wide variety of interpretations of good and bad in human judgment that it’s no more than a vague epigram. Thusly, Levitt can write, “Whenever I try to answer a question, I put myself in the shoes of the actors and I ask myself ‘What would I do if I were in that situation?’” and never truly ask himself if a thirteen year old crack-addicted dealer might make the same calculated, rational choice. Much like my hypothetical uneducated English major, Levitt seems supremely unaware of absurd rationalizations of humans as well caricatured by Dostoevsky in Notes from the Underground. He has no interest in a supremely personal supreme good that might appear completely irrational to an outside observer. Such things fall outside the realm of statistical likelihood.

But what’s most galling in the book is that Levitt and Dubner repeatedly warn against confusing correlation and causation then go about doing precisely that in their most controversial and least convincing essay. The collection’s fourth chapter is a longer rehash of a paper Levitt presented entitled “The Impact of Legalized Abortion on Crime” in which the argument runs that the 70s and 80s saw a huge upswing in crime rates that completely dropped away starting in 1992 and this can be neatly tied back to one fact, the legalization of abortion nationwide with Roe v. Wade. Their hypothesis is that since criminals are most commonly males aged 18-24 from impoverished backgrounds, legal abortion allowed mostly poor women in bad situations (or with the kind of bad parenting skills that might lead to criminal offspring) to avoid raising future jail cell citizens. 1973 plus 18 equals 1991 and ta-da!

There are a number of obvious weaknesses with such an argument such as legalized abortion simply took away the punishments doled out for a rather large number of illegal abortions previously taking place; the legalization of the Pill a decade earlier and its widespread use should have lead to falling crime rates even earlier and on a much larger scale as more women would take oral contraception than have abortions; and Levitt’s cavalier dismissal of any other factor such as increased police forces in cities as a result of rising crime rates as well as the dramatic and fashionable appearance of crack cocaine in those years. Likewise Levitt’s attempt to bolster his shaky argument by pointing to how many of the young generation born after Romania’s dictator Nicolae Ceau┼čescu outlawed abortion were criminals not-so-neatly blurs what might be substantial differences between who is considered a criminal in a repressive autocracy and in an open democracy.

This kind of fuzzy methodology and shady comparisons fill the book from cover to cover, and Dubner’s dumbing down on Levitt’s more complicated econo-speak obscures the duo’s shaky methodology and even shakier conclusions. As the book provides little if any background on their work’s analysis (who needs charts in a beach book anyway?), there’s little to bolster their more energetic claims and almost nothing on which to attach specific criticisms. Dubner himself reads the book and his almost boyish voice takes on a tone of abject hero worship when he feeds us tidbits about Levitt’s personality and biography. In some ways, these peeks behind the curtains are some of the more fascinating parts of the book — human beings not quantified down to their buttonholes always remaining infinitely more interesting than any measurement of them might suppose.

No comments: