(January 2000) Abortion is a divisive and emotional issue. Add racial overtones and the implication that public officials do not deserve all the credit for reductions in crime during the 1990s, and combustion is inevitable. Informed debate is not.

Stanford Law School professor John J. Donohue III and University of Chicago economist Steven D. Levitt ignited a debate last August when they released a study on the relationship between abortion and crime. Their findings suggest that legal abortions have prevented the births of many would-be criminals. The absence of these people, according to their research, is behind at least half of the dramatic drop in crime rates seen between 1991 and 1997.

Were the researchers just trying to stir up controversy with their results? Not at all, according to Donohue. "I was asked to write a paper on why crime was falling. In the course of looking at . . . various social programs and their impact on crime, I sort of stumbled across the data on abortion."

The "sheer magnitude" of abortions prompted Donohue to dig deeper. When he and Levitt did, they found three factors especially compelling:

The timing. The abrupt drop in crime coincided with the coming of age (roughly 20 years later) of young people born after the 1973 Roe v. Wade Supreme Court decision legalizing abortion. Also, in states where abortion became legal before 1973, the drop in crime started sooner.

The magnitude. States with high abortion rates have experienced more dramatic declines — the authors estimate 15 percent higher declines — in crime.

The pervasiveness. The reduction in crime has occurred both in cities that have waged war on an acknowledged cause of violent crime, crack cocaine, and in cities relatively untouched by crack. Similarly, crime has fallen both in cities that have expanded their police forces and changed policing techniques and in cities that have not changed their policing techniques or expanded their police forces.

The researchers also observed a decrease in the rate of crime among young people. Young people ages 18 to 24 engage in more criminal activity than any other age group. If the size of the 18-to-24 population decreases, whether because of lower birth rates or because of any other event, the number of crimes also drops. But Levitt and Donohue identified a decline in the rate of crime in this age group, signaling that there was more behind the drop than a shrinking population.

The Demographics of Abortion

The researchers then looked at abortion demographics and noted that women who have abortions tend to be young, unmarried, and poor. Children born to these mothers, like children of mothers who have little education or who do not want to be mothers, are at higher risk for committing crime, according to the authors. If today's 18-to-24-year-olds did not include as many of these high-risk children, that might explain the group's lower per capita crime rate.

The authors stress that the availability of legalized abortion may not simply have reduced fertility but may also have delayed it and improved the environment of subsequent children. Wanted children who are born to married, supportive parents are less inclined toward criminal behavior. In many instances, the authors reason, abortion may have given young, single, poor, and often minority women time to mature, get jobs, marry, and become more economically stable and then have children who would have a lower risk for committing crime.

Strong Reactions

Interest groups have questioned where this research leads. Abortion supporters, wary of a study that could be misconstrued as promoting eugenics through abortion for the poor and for blacks, have interpreted the findings carefully: "Healthy women, healthy children, and healthy families are beneficial for society as a whole, and the right to choose legal abortion is instrumental in promoting all three of those," said William Lutz, a spokesman for the National Abortion and Reproductive Rights Action League.

The National Right to Life Committee issued a press release acknowledging that the staff of the organization had not read the study but that they rejected "the notion that the appropriate way to solve any of society's problems is to kill unborn children." Americans United for Life general counsel Nikolas T. Nikas concluded: "We've been presented with a new twist to the old adage: 'Stop the criminal before he commits the crime.' Does this mean executing a death penalty before any crime is committed?"

Criminologists have also weighed in, focusing on how the research was done rather than on the controversy it has stirred. Charles Wellford, acting chair of the Criminology and Criminal Justice department at the University of Maryland, heard Levitt present the research recently and was impressed with the care that went into it. "Levitt's a serious guy. He's not after inflammatory results." Nonetheless, Wellford noted problems with the research methods. For crime, he said, "aggregation at the state level masks trends within states" and county-level analysis would be more useful, but abortion data are unavailable at the county level. Also, the research relies on crimes known to the police as an indicator, whereas a crime victimization survey "would be a better measure of what's going on out there," said Wellford.

Alfred Blumstein, J. Erik Jonsson University Professor of Urban Systems and Operations Research and director of the National Consortium on Violence Research at Carnegie Mellon University's Heinz School, shared the view that the research was carefully done. But he disagreed with its conclusion that abortion accounts for as much as half of the downturn in crime. "They're probably onto something. Abortion may be relevant, but it's a much more complicated story than they were able to deal with," said Blumstein. He noted that the researchers had not accounted for factors such as the changing prevalence of handguns and changing drug markets. He also noted that the researchers should have disaggregated crime by single year of age: "One would expect that, if the abortion argument were the case, there would be a sequence of peaks in successive [age] cohorts. But in 1994, everyone started down, at all ages."

Revisions and More Debate

Some of these criticisms may be addressed in a revision that is underway. Donohue believes the final version of the study may be out in the first quarter of 2000.

When Donohue and Levitt's research is finally released, there will no doubt be even more debate. But as for greater scrutiny, that depends on people's ability to do what James Finckenauer, a professor with Rutgers University's Institute for Criminological Research, said he did: "Step back and say, 'If I look at this objectively, putting aside the emotional sort of implications, what does this tell me?'

"It's a funny thing about criminology," said Finckenauer. "All of a sudden in this decade we have witnessed these dramatic declines in crime rates. . . . But now we're hard pressed to explain why this is occurring. It's almost as if it were easier before this occurred to explain why we have the kind of crime that we have. But to explain the converse has not been so easy."

And that, according to Donohue, is the point. "People are always pushing that it's because of the way [New York City] Mayor Giuliani polices or because of the vast increase in incarceration that we're getting the drop in crime. To the extent that this paper is correct, you would attribute less of the drop in crime to those other factors. And that might suggest less reliance on those other factors as a method of crime control. If this paper gives us insight into that, that's very valuable. It can help us from going down the wrong path."


Allison Tarmann is editor of Population Today.