Ever since John Kenneth Galbraith coined the term conventional wisdom, economists and other big thinkers have been trying to fashion their reputations by overturning it.
The latest aspirant for fame as a tilter against conventional wisdom is University of Chicago economist Steven D. Levitt. Acclaimed by many in his field as a brilliant thinker, Levitt has used the analytic tools of an economist in unusual ways: hes constructed algorithms to determine that some Chicago teachers were helping their students cheat on standardized tests; hes used contestant voting patterns on a TV game show to test theories of racial discrimination, and perhaps most famously, hes attempted to correlate the legalization of abortion in America with the beginning of a nationwide decline in crime that began 18 years later.
Though Levitt has published most of this research in obscure, specialized journals, hes lately distilled many of his theories in the best-selling book, Freakonomics, whose subtitleA Rogue Economist Explores the Hidden Side of Everythingtells you all about what Levitt (or his publisher) wants you to think about him.
In a world apparently hungry for unconventional wisdom, sales of Levitts book, co-authored with journalist Stephen J. Dubner, have not only soared, but the book has won lavish, and often uncritical, praise from the critics. A Wall Street Journal review compared Levitt to Indiana Jones and opined that [c]riticizing Freakonomics would be like criticizing a hot fudge sundae. Reviewers have been especially impressed with Levitts work on abortion and the decline in crime: The Economist notes that the book moves methodically and persuasively through the statistical evidence linking the two. And in the New York Times, business journalist Roger Lowenstein has written that Freakonomics is a splendid book, full of unlikely but arresting historical details that distinguish the authors from the run of pop social scientists. Readers may even find themselves developing a dose of skepticism about the worldno bad thing.
Whats too bad, however, is that those rushing to praise the book havent developed the same dose of skepticism, especially about Levitts work on crime. Not only does the work of other economists who have tested Levitts theories cast serious doubt on his conclusions, but some of the latest crime stats also undermine Levitts notion that policing innovations undertaken in places like New York City had little to do with the drop in crime.
Levitt proposed his correlation between abortion and crime rates several years ago. To test the theory he looked at the decline in crime rates in states where abortion was legalized firstincluding New York and Californiaand found that crime started falling there first. He posited that many of the women having abortions were precisely the kind who might raise children likely to be criminalsthat is, poor, single womenso that legalization essentially reduced the pool of potential wrongdoers. He concluded that abortion accounted for more than half of the decline in crime in the 1990s.
But in economics, new theories based on innovative research are almost immediately tested to see if they can be replicated, and in Levitts case what quickly emerged were counterstudies that questioned his methods and conclusions. Professor Ted Joyce of City University of New York found that incidents of homicide by perpetrators in age groups too old to have been affected by legalized abortion declined faster than murders by younger perps. John Lott of the American Enterprise Institute and John Whitley of Adelaide University in Australia noted that research on the legalization of abortion suggested that it actually increased illegitimate births and single-parent families. They concluded that rather than decrease crime, legalized abortion probably contributed slightly to its increase.
Levitt has addressed some of these criticisms in academic journals, though not convincingly, to my mind. But in Freakonomics, he completely ignores this counterevidence and presents his own work as if its Scripture. Only a few reviewers, most notably the eminent sociologist James Q. Wilson in Commentary, have noted the way the book merely disregards the work of others.
But the flaws in Levitts notion of what caused crime to fall go beyond his theory of abortion. In Freakonomics he also dismisses the idea that innovative policing methods of the type that New York instituted during the Giuliani years had any effect on crime. This idea has been particularly attractive to reviewers of the book, prompting Lowenstein in the Times to crow that [w]hile all the world was congratulating Rudolph W. Giuliani for reducing violent crime . . . the authors demonstrate that Hizzoner probably had little to do with it.
Levitt bases this conclusion not only on his work on abortion but on another study hes done, which attempts to correlate declines in crime with the expansion of a citys police force. Looking at data from 1970 through 1992, Levitt concludes that adding to the size of your police force probably accounted for about 10 percent of the 1990s drop in crime. With that in mind, he rejects the idea that New Yorks policing methods helped rein in crime because, once one adjusts for the growth in Gothams force, which Levitt says increased by 45 percent, the citys reduction in crime is no greater than that of Los Angeles, a city that never incorporated the broken windows policing that Giuliani championed.
But there are serious questions about Levitts work here. For one thing, once you adjust the growth in New Yorks police force for the expansion of its population during the 1990s and then compare it with per capita police rates in other cities (as even Levitt does in his other work), the numbers tell a different story. New Yorks police force grew by 18 percent per capita during the 1990s, according to an FBI study, Police in Large Cities. This put the city behind police force increases in Newark, and in line with or slightly ahead of gains in cities like Baltimore, where the size of the force relative to the population rose by 20 percent, in Philadelphia by 13 percent, St. Louis by 10 percent, and Chicago by 9 percent.
Whats startling about this list is how little these increases in staffing paid off for most of these cities during the 1990s. While New Yorks violent crime rate declined by 65 percent in the 1990s, most of these other cities saw only small decreases in crime, and in a few cases violent crime actually rose. Chicago, for one, ultimately passed New York as the place with the highest total of murders per year, even though Chicagos population is only 38 percent of New Yorks.
Moreover, Levitt never explains why its merely an increase in police that matters so much to reducing crime, while the absolute number of police in a city seems not to matter. The two cities in the country with the highest rate of cops to population, Washington D.C. and Newark, both have much higher rates of crime than Gothamthough Newarks did decline significantly in the 1990s.
Even more startling, however, is whats happened in New York since 2000. In the last four years, New York Citys violent crime rate has continued to decline, down by nearly another 25 percent, even though the citys police force, caught in a post-9/11 budget crunch, has shrunk by several thousand officers. Not surprisingly, the citys current police commissioner, Ray Kelly, attributes the continuing declines to innovative policing methods, which have allowed the NYPD to target crime where it appears and to use its shrinking resources more effectively to combat it. Kellys Giuliani-era predecessors, William Bratton and Howard Safir, offered the same analysis.
In Freakonomics, Levitt does nothing to acknowledge, much less explain, all of these statistical exceptions to his theories. Its one thing to challenge the conventional wisdom with your own fresh data. Its quite another simply to ignore all the facts that dont fit neatly into your new way of thinking.