Man of records: Freakonomics author's statistical microscope
reveals a life less ordinary
by Robert Fulford
(The
National Post,
3 May 2005)
One of the many fascinated admirers of Steven D. Levitt has called him "a kind of intellectual detective trying to figure stuff out." Levitt is a 37-year-old economics professor at the University of Chicago who spends much of his time dealing with issues that rarely interest others in his profession. He uses economics as a way to strip a layer off ordinary life and reveal what's happening underneath.
Relatively trivial questions occupy him as much as grand themes. For instance: Why, if great fortunes flow through the cocaine business, do so many dealers live with their moms? It happens that Levitt can now answer this question because (through the closest thing to a miracle that a social scientist ever experiences) he got his hands on the account books covering four years in the business life of a certain J.T., the leader of a crack-cocaine gang in Chicago.
Having gone to college and majored in business, J.T. knew how to keep records -- and used them to impress his bosses higher up in the chain. His numbers showed Levitt and a colleague that a few gang leaders (they call themselves "the board of directors") make big money while about 98% of dealers receive tiny incomes, under $1,000 a month. They live with their moms because they can't afford to live anywhere else.
This glimpse into the drug life appears in Freakonomics: A Rogue Economist Explores the Hidden Side of Everything (Morrow), by Levitt and Stephen J. Dubner. Levitt first stimulated controversy when he and a colleague, John J. Donohue of Stanford, explained in Harvard's Quarterly Journal of Economics that legalized abortion was the reason the United States began experiencing less crime after the early 1990s.
About 14 years ago, America's horrendous crime rate suddenly began declining. Why? Levitt and Donohue noted that this was 18 years after the U.S. Supreme Court, in Roe v. Wade, made abortion legal. In three states where abortion was legalized earlier, the crime-rate decline also began earlier. Millions of unwanted babies had been aborted -- and, as many have noted, unwanted babies are at high risk of becoming criminals. Levitt says it's obvious: "Increased abortion reduced unwantedness and therefore lowered criminal activity." That controversial thesis was attacked as simplistic, but no one has disproved it.
Under Levitt's statistical microscope, the world becomes more interesting and also more appalling. Much of his work focuses on incentives and human nature. As an illustration, he mentions the seven million American children who disappeared one spring midnight in 1987. It wasn't a mass kidnapping. That was the year the American income tax service first required that parents provide the Social Security numbers of children they were claiming as dependants. The seven million were all, of course, fictional. The regulations till then had provided an incentive for parents to cheat by taking deductions for imaginary children.
Levitt believes that where incentives to cheat exist, people will find them and use them. "Who cheats? Well, just about anyone if the stakes are right ... Cheating is a primordial economic act: getting more for less."
Many schoolteachers, for instance, will cheat if the incentive is in place. If a school jurisdiction creates standardized tests and compares the results of different schools, some teachers will falsify the grades of their students, thus improving their own professional status.
To identify that kind of cheating, Levitt says, you mentally move to the side of the possibly deceitful teacher and ask how such a person would approach the task of cheating. Then, if detailed records have been kept, you can invent an algorithm that will detect suspicious anomalies -- such as a sudden improvement in a student's scores under one teacher and an equally sudden deterioration in the following grade. When the Chicago school system analyzed 700,000 sets of test answers from 1993-2000, many teachers were found to be correcting their students' answers on computerized tests. A dozen were fired; others, it was assumed, were frightened.
The 10-page description of this process in Freakonomics makes clear sense even to the mathematically incompetent. And it has a pleasant side: "The algorithm could also identify the best teachers in the school system. A good teacher's impact was nearly as distinctive as a cheater's."
Levitt and another scholar applied a similar system to uncover corruption in, of all things, sumo wrestling. It would be just about impossible for anyone watching a bout to know that one sumo wrestler had agreed to lose, since the decisive push-and-pull appears to end almost before it starts. But the managers of sumo tournaments in Japan also keep thorough records, which made it possible for Levitt and Mark Duggan to figure out which wrestlers were throwing their matches; they had the results of 32,000 bouts fought by 281 wrestlers.
As Freakonomics says, "the data tell the story." In a tournament, each wrestler must win a certain number of bouts to remain in the elite rank of the profession. As the tournament ends, a wrestler who has already won enough bouts will sometimes allow an opponent to win if the opponent needs just one more victory; the favour will be returned later. Levitt and Duggan apparently proved that this happens fairly often, but their results were rejected by the sumo authorities. They couldn't face the shame of corruption in their national sport.
In one sense, Levitt is a creation of journalism. As a personality he came to national attention in August, 2003, when The New York Times Sunday magazine published Dubner's article about him. As Levitt told an interviewer last week, "He created a fictional account of me." Levitt said the character in the Times article was smarter, more likeable and more interesting than the actual Levitt. But Levitt didn't mind. When publishers began asking him to write a book, he said he would do it if Dubner collaborated.
That turns out to have been a good idea. Freakonomics, published last week, has already won a good ranking on Amazon.com and on Sunday made its first appearance on The New York Times best-seller list, in fifth position. Levitt is working the TV-interview circuit, doing his best to imitate the charming young fellow Dubner described, maybe blushing occasionally when asked why he's written under such a silly title. If pressed, he will no doubt explain that it's all a matter of incentives.
