Saturday, May 7, 2005

Inside The Mind Of Steven D. Levitt : A Review of Freakonomics

Steven D. Levitt (with Stephen J. Dubner) has a hot new bestseller Freakonomics: A rogue economist explores the hidden side of everything.

In the introduction, Levitt makes abundantly clear that his book has no central theme. I can almost agree with this assessment. The book is indeed much more a blog type compendium of different topics, rather than an exposition on one theme. But I did find one theme that runs through out the book. Levitt poses interesting questions, reports interesting facts and occasionally makes clever arguments, but these questions, facts and arguments are surrounded by misleading statements, hazy statements, inaccuracies, poor logic, sloppiness and outright errors.

These flaws run from the minor to the grand scale. Indeed,one must begin by considering that Levitt clearly believes that through out the book he is "doing economics." In fact, although he does tend to include some type of cost benefit analysis in most chapters, his analysis tends to be much more that of a sociologist than that of an economist. Consider the titles of some of his chapters: "What Makes a Perfect Parent?", "Where Have All the Criminals Gone?" and "What Do School Teachers and Sumo Wrestlers have in Common?"

Further although there is an implication by Levitt that he is writing theory, he is in fact more of a historian reporting on past data.(For the important distinctions between theory and history, see Ludwig von Mises' Theory and History.)So instead of a book of economic theory, we have a flawed book on sociological history.

On a minor scale, Levitt tends to use misleading chapter titles. His chapter "What do School Teachers and Sumo Wrestlers have in Common?" comes up with the answer: some in both groups cheat to get ahead. Since there are some in almost any group that will cheat to get ahead, there is nothing distinctive about this supposed "link"

It is the same as saying "What does Steven Levitt and the members of the offensive front line of the New York Giants football team have in common?" Answer: They all use cell phones. The facts in both cases are true but they result in no new insight, but the questions themselves tend to mislead one into thinking that there is some type of distinctive link in the answers when there is not. At most the chapter title is a sloppy effort at being cute.

Levitt goes from bad to worse in the title of his next chapter: "How is the Ku Klux Klan like a Group of Real Estate Agents?" His answer: they both use privileged information to their advantage. Again, nothing remarkable about this chapter, since everyone uses privileged information to their advantage (Indeed that is pretty close to the definition of an entrepreneur!)

Levitt's presentation is so sloppy that it almost fails to get across the point that privileged information is used by the Ku Klux Klan and real estate agents. In short, it is a pretty bad example used to get in a little, quite interesting, history about the Ku Klux Klan.

As for sloppy and hazy arguments, in one chapter, Levitt relates the story of how a pre-school attempted to solve a problem of children being left late after school. The pre-school instituted a fine for parents who left their children late. With the new stated policy (It was only a $3.00 fine), more parents left their children late.

Levitt in the next chapter on page 45 calls this cheating: "So parents...cheat are we to assume that mankind is innately and universally corrupt?"

Given all the cheating going on in the world, it is quite odd that Levitt uses this as one of the examples, which only by the wildest stretch could be called "cheating." In fact, it really is a story about the limits of knowledge, and how people will change their actions when more knowledge becomes available.

In Levitt's next chapter sloppiness is coupled with an implied wrong conclusion. Levitt does the math and shows that on a per hour basis "The per hour death rate of driving versus about equal." He then concludes "The two contraptions are equally likely... to lead to death." He ends his analysis of flying versus driving with this conclusion, which tends to imply that it doesn't matter whether you fly or drive. But, in fact, using Levitt's own data the clear conclusion is the exact opposite of what Levitt implies. The clear conclusion is to fly whenever possible. Why? Because you get there faster, which means you are traveling a shorter amount of time at the per hour death rate. If it takes five hours to fly from New York to Los Angeles and three days or 72 hours to drive, then if the death rate per hour is the same, the risk by driving is close to 15 times greater. If you do nothing else after reading his book other than take Levitt's implied conclusion on driving versus flying and drive instead of fly, Levitt has increased your chances of dying when traveling by nearly 15 times!

Levitt's chapter on names continues the trend of sloppiness, haziness, illogic and poor conclusions.

He tells us that " isn't famous people who drive the name game." He uses the fact that no parents are naming their daughters Madonna as part of his argument that this is proof that parents don't name their children after famous people, but this is just sloppy logic. Just because parents don't name their children after Madonna doesn't mean many parents aren't naming their children after famous people. In fact only a page away from where Levitt tells us that famous people don't drive the name game, Levitt lists the most popular black names in California in the year 2000. Number 4 was Michael and Number 2 was Jordan. Hmmm, it seems to coincide with a period when there was a pretty famous basketball player on the court, named Michael Jordan.

In short, I could literally write a book (maybe many books) detailing the errors, sloppiness, inaccuracies, haziness and poor theory going on in this book. And I haven't even touched on the errors in his chapter on Roe v.Wade. (Levitt is probably most well known for his theory that abortions reduce crime, see Steve Sailer for a critique of this theory.) Nor have I discussed his love affair with regression analysis, which is a questionable method to prove theory in the social sciences (See Leoni and Frola)

In a sense though, this book is a great mystery book. The great mystery being just what marketing plan was implemented to drive this disaster onto the best seller list?

1 comment:

  1. Robert

    Liked your review, although I never read the whole Freakonomics book. I was put off by the cute title and the even cuter chapter headings. I skimmed it at the book store and it seemed at best to be a compelation of well-known commen sense ideas in a clever and expensive package. Maybe thats the marketing plan...