Does Abortion Lower The Crime Rate?

Crime in the U.S. has fallen dramatically since 1991. By 1998, the homicide rate was down by more than one-third, and the rates for all violent crime and property crime were each down by around one-quarter.

Many explanations have been offered for the decline, including increased expenditures on prisons and police, better policing strategies, the strong U.S. economy, and the diminished role of crack cocaine. Spending on prisons and police has been increasing since the 1970s, however, so this factor therefore does not explain why crime rose until 1991 and then fell sharply. Better policing methods, as promoted by Mayor Rudolph W. Giuliani in New York, may be making a difference. But crime rates also fell substantially in Los Angeles and the District of Columbia, which are not renowned for their policing skills. As to the strong economy, it is hard to show generally that growth in income and employment leads to less crime. For example, the economy performed well from 1983 to 1989, while national crime rates rose. Finally, the fall in crime rates after 1991 occurred in places where crack cocaine had never been much of a factor.

A recent study, "Legalized Abortion and Crime," by Professors John Donohue of the Stanford University Law School and Steven Levitt of the University of Chicago, proposes a new causal factor: the legalization of abortion in the early 1970s. The idea is that the children who were not born would have been disproportionately likely to grow up in poverty and on welfare with a young and poorly educated single parent. Because these factors are known to breed crime, the children not born would have been prime candidates to be criminals 15 to 25 years later. Hence, the absence of these children contributes to the drop in crime rates since 1991. My reactions to this idea: (1) This is a surprising hypothesis. (2) It may well be correct. (3) The implications are explosive politically and are likely to cause a major ruckus.

TIMING. Donohue and Levitt present three types of evidence to support their theory. First, the sharp rise in legal abortions--from fewer than 750,000 in 1973 to a plateau of around 1.5 million a year since the early 1980s--fits with the timing of the drop in national crime rates since 1991. The cumulative effect from abortions 15 to 25 years earlier can also explain why the national crime rate continued to fall through 1998 and is projected to keep falling for a further 10 to 20 years. In fact, the authors estimate that the number of abortions and the consequent reduction in crime would have been significantly greater had not the 1976 Hyde Amendment, which restricted federal funding of abortions through Medicaid, been in effect in various forms.

Second, a few states, including New York and California, legalized abortion by 1970, three years before the U.S. Supreme Court's Roe vs. Wade decision in 1973. As the theory implies, the early legalizers experienced falling crime rates sooner than the rest of the country. Finally, abortion rates responded to the legalization differently across states, and those with the highest rates of abortion in the 1970s experienced the sharpest drops in crime in the 1990s.

WEEDED OUT. The researchers estimate that, for every 1,000 extra abortions in 1973 to 1976, there were 380 fewer property crimes, 50 fewer violent crimes, and 0.6 fewer murders in 1997. Overall, the abortion effect accounted for one-half of the drop in crime from 1991 to 1997. The rest is explained by increases in prisons, police, and other factors, or goes unexplained.

About 20% of the abortion-related drop in crime arose because of the reduced population of 15-to-24-year-olds (the high-crime age group) in the 1990s. The main effect, however, is the reduced frequency of crime among those in the 15-to-24 age group who did manage to be born and grow up. Apparently, abortion particularly weeded out the children who would have been likely to follow criminal careers.

Donohue and Levitt argue reasonably that they are carrying out objective scientific research about the determinants of crime and that the policy implications can be left to others. The effect on crime, even if confirmed by further study, would probably not moderate the views of pro-lifers, who view abortion as murder. Similarly, the evidence would have little influence on pro-choice advocates, who already view a woman's right to an abortion as a fundamental liberty. But for people with less extreme views, including me, the policy implications could be important. If abortion rights turn out to be a strong crime fighter, then opinion is likely to move in favor of these rights.

Before it's here, it's on the Bloomberg Terminal.