June 28, 2005

James Q. Wilson reviews Levitt's Freakonomics abortion-crime theory in Commentary

From the new July issue of Commentary, not yet online:

During my many years of lecturing on crime, invariably the first two questions I would be asked were: "What do you think of the death penalty?" and "What do you think of gun control?"

No more. Now the first question is whether I believe that legalized abortion has cut the crime rate. For this I can thank Freakonomics, the weirdly named book by Steven D. Levitt and Stephen J. Dubner that has been high on the New York Times best-seller list for weeks now. My answer, by the way, is no: I do not believe the evidence shows a causal link between legalized abortion and our reduced crime rate.

Levitt, an acquaintance of mine, is an immensely talented economist whose restless mind has inquired into all sorts of fascinating topics....

Back to abortion and crime. Levitt's argument is that, with the legalization of abortion by the Supreme Court's 1973 decision in Roe v. Wade, many fetuses were killed in America that would otherwise have led to the birth of unwanted children. Such unwanted children, receiving little affection and guidance, would have been more likely to commit crimes when grown. Ergo, their removal from the population had something to do with our lowered crime rates.

Why should we think such children would have been unwanted? Because, Levitt contends, they would have been born to thousands of poor, single, teenage mothers. Levitt conspicuously refrains from saying so, but a very large fraction of these poor, single, teenage mothers would have been African American: over 60 percent of all black children are born out of wedlock, and the abortion rate is roughly three times greater among black than among white women.

To prove that abortion reduced crime, Levitt and his coauthor on the original paper, John Donohue, examined crime rates 15 to 18 years after the Roe decision, and found a drop. Moreover, they pointed out that five states had already legalized abortion three to four years before the decision: in these early-legalizing states, crime rates fell sooner than in states that did not permit abortion until Roe.

You would never know it from this book, but not only have these claims been criticized, but several scholars have offered rival theories. On the issue of abortion rates alone, the economists John Lott and John Whitley have written that, even before Roe, many anti-abortion states allowed abortion if the life or health of the mother was at risk; in these states, there were at least as many abortions per 1,000 live births pre-Roe as in states that had made abortion legal. Why, then, attribute falling crime rates to legalized abortion?

Levitt and Donohue have rejoined that, in those states where abortions were still nominally illegal, it was well-to-do white women who mainly availed themselves of the loopholes in the system. But there is no evidence of this; to the contrary, black women were over-represented among those having abortions in such states.

Now look at homicide rates by the age of suspected offenders. In the late 1990s, roughly a quarter century after Roe, the murder rate was falling for offenders aged twenty-six and older -- a class of offenders much too old to have been affected by Roe one way or the other. As for the youngest offenders, those between sixteen and twenty, their murder rates had jumped up in the early 1990s, probably because of involvement in the crack cocaine trade. Again, no Roe effect.

George Akerlof, Janet Yellen, and Michael Katz have argued that legalized abortion actually increased the number of out-of-wedlock first births -- because the availability of abortion, along with the advent of new contraceptive devices, rendered sex "cost-free" for men but not necessarily for the women they impregnated. Were the children who were increasingly likely to be born to unmarried women "unwanted"? Perhaps they were, but we do not know; Akerlof and his colleagues have not given us sufficient evidence.

As of now, no one is entitled to decide who is correct in this matter, whether Levitt or any of his critics. But it is certainly premature to say that Levitt is right, and positively disconcerting to take the work of an enamored journalist that Levitt must be right.

On another controversial matter, however, Levitt is clearly right, and I am his victim. I once wrote that the proportion of juveniles in the population was going up and that therefore the crime rate would go up. Levitt correctly takes me to task for this unwarranted assertion, which was later proved wrong. His criticism reminds me of something my Ph.D. adviser once said, no doubt quoting someone whose name I have forgotten: social scientists should never try to predict the future; they have enough trouble predicting the past.

Touché ... Levitt, of course, being a classic example of a social scientist who has failed to predict the past. When Levitt concocted his theory in 1999, he only looked at crime rates in 1985 and in 1997, and he forgot to look at crime rates by narrowly defined age groups. So, he completely overlooked the fact that, in direct contradiction of his theory, the first cohort born after legalization went on an enormous teen violence spree in-between 1985 and 1997. Ever since then, with his name and reputation linked to his half-baked theory, he has been hustling like P.T. Barnum to turn his slapdash hypothesis into conventional wisdom to preserve his marketability.

My published articles are archived at iSteve.com -- Steve Sailer

No comments: