Yesterday in NRO, Robert VerBruggen provided a nice history of the ongoing debate among economists over whether the legalization of abortion in the 1970s resulted in lower crime rates starting in the 1990s. John Donohue and Steven Levitt’s working paper on this topic started receiving media attention in 1999. Their study was published in The Quarterly Journal of Economics in 2001 and was popularized in their book Freakonomics. While the study itself has received a great deal of media coverage, most outlets have given subsequent critiques of the study considerably less attention.
One problem with the original study is that it failed to properly account for interstate travel. The authors argue that states that legalized abortion before the Supreme Court decision in Roe v. Wade legalizing abortion, experienced crime-rate reductions earlier than other states. However, it is not clear that the Centers for Disease Control data, which Donohue and Levitt used, was able to effectively track the residence of women who crossed state lines to obtain abortions. In particular, when New York legalized abortion throughout the first 24 weeks of pregnancy in 1970, there was no requirement that women be state residents. As a result, approximately 58 percent of women obtaining abortions in New York in 1971 and 1972 resided in other states. Additionally, it is not clear that the authors properly controlled for interstate migration; many people born in one state grow up elsewhere.
Economist Ted Joyce has argued that reductions in crack cocaine use were primarily responsible for the crime-rate decline in many cities. Economists John Lott and John Whitely, meanwhile, have argued that Donohue and Levitt’s hypothesis can be tested by analyzing the demographics of crime rates. If their theory is correct, there should be a decrease in the number of crimes committed by young people. However, between 1976 and 1998, murder rates for the oldest age groups steadily decreased, while the murder rates for younger age cohorts — which eventually began to include people born after Roe — increased. Lott and Whitely also found that murder-rate trends among young people did not vary much between the states that legalized abortion early and the states where abortion became legal only after Roe.
The “abortion cuts crime” hypothesis never did much to inform the ongoing public-policy debate over life issues. Abortion opponents believe in the intrinsic value of all human life and therefore were not persuaded by the argument that the legalization of abortion might have reduced the population of criminals. Additionally, supporters of legal abortion recoiled from the eugenic implications of this type of argument. Even so, it is interesting that Donohue and Levitt are continuing to publish research on this subject. Hopefully media outlets are willing to give scholarly critiques of this research the attention they deserve.