C rime is down across America. The nation's crime rate has been dropping for the best part of a decade now, and everyone is keen to take the credit. New York's Mayor Rudy Giuliani claims that zero-tolerance policing is responsible; former California Governor Pete Wilson credits three-strikes-and-you're-out laws; President Bill Clinton says gun control and federal funding for prison construction and new police officers have done their part.
Likely as not, they're all partly right. But what if it turns out that other factors are actually having far more influence on the crime rate than these get-tough policies? Why lock up two million people--more than half of them nonviolent offenders--at a cost of tens of billions of dollars a year and the disruption of untold millions of lives, if the real explanations for the drop in crime lie elsewhere?
Last summer word began circulating, first in the academic community and then in the media, that two professors, John Donohue and Steven Levitt, had found solid evidence of exactly that: They had discovered a link much stronger, more statistically demonstrable, than the link between anticrime policies and crime rates. More shocking still, the link they found was between abortion and crime. Or to be more precise, between the 1973 Roe v. Wade decision legalizing abortion and the much heralded fall in crime rates starting about 18 years later, in the early 1990s.
When laid out in its crudest form, the notion that crime was dropping because potential criminals were being aborted provoked a firestorm of objections from left, right, and center. What terrible policy conclusions, people demanded to know, were they supposed to draw from this information? "Racist, genocidal stupidity," the conservative monthly American Spectator labeled the new study. On the opposite political flank, Jackie Cissell, director of the Indiana Family Institute, wrote that "African Americans [are] in shock because it could threaten the very survival of the race."
But if the smoke ever clears, progressive critics will find that the policies this new study actually points to are not shocking at all. In fact, they're quite traditionally liberal.
John Donohue, a professor of law at Stanford University who once ran an unsuccessful campaign as a Democratic nominee for the Connecticut state senate, and Steven Levitt, a University of Chicago economist, have long sought explanations for the falling crime rate. In a 1998 paper published by the Journal of Criminal Law and Criminology, Donohue pointed out that the explanations people usually give just do not account for much statistically. According to his calculations, the large decline in crime during much of the 1990s was either a short-term anomaly or the result of factors that had not yet been identified. It was a hint of a fascinating new theory that he and his colleague were developing.
During the course of their research, Donohue and Levitt had almost accidentally stumbled upon the number of abortions performed in America and the fact that--in Donohue's words--"poor, unmarried, young, low-education women tend to have more abortions. And their kids tend to have higher rates of crime." The two researchers began crunching the numbers, and after several years they concluded, to their own surprise as much as anyone else's, that fully half of the decrease in crime that has occurred over the past decade can be directly attributed to the fact that women in the 1970s and 1980s had ready access to abortion.
After exhaustive peer review lasting the better part of a year, their study, titled "Legalized Abortion and Crime," is scheduled for publication in the May 2001 issue of Harvard University's prestigious Quarterly Journal of Economics. The authors hope that a longer version, specifically addressing many of the criticisms leveled against their work, will follow in a top law review.
Looking at state-by-state and year-by-year figures, the two professors found a remarkable correlation between abortion rates and crime rates 15 to 18 years later. And that's not all. They also determined that in the states that legalized abortion prior to the Roe v. Wade ruling, crime rates began falling earlier than in other parts of the country (see box). Moreover, while the rate of arrests did drop in other age groups, among young people (those whose mothers had the option of a legal abortion) it dropped far more. The authors factored in a host of other possible explanations, and the correlation between crime rates and abortion remained powerful. "According to our estimates," they boldly asserted, "legalized abortion is a primary explanation, accounting for at least one-half of the overall crime reduction... . The social benefit to reduced crime as a result of abortion may be on the order of $30 billion annually."
That such an idea, put like this, would raise the hackles of an extraordinary range of people should have surprised no one. Levitt and Donohue had stepped into the vicious ethical and political minefield of the American abortion debate--as well as the treacherous terrain of race politics, since African Americans have abortions at a higher rate than whites. But the authors, economists rather than ethicists, were unprepared for the response and initially seemed almost too stunned to prepare counterarguments. "What's odd about our study," Levitt now reflects as he prepares for publication of the work and, presumably, renewed assaults on its authors, "is it manages to offend just about everybody. [But] our worldview is an economic worldview--that people respond to incentives. I view it as being apolitical."
That lack of political savvy may explain much. "I don't think it's controversial to say crime is higher among African Americans," Levitt still insists. "We're not saying there's anything intrinsic about this. There's also higher poverty among African Americans. The causality is not important in our argument, in the sense of why it is that African Americans are disproportionately represented in the crime statistics."
Given recent political history, however, and given the ways in which the poor--especially the black and Latino poor--have all too often been blamed and punished for the circumstances of their poverty, causality was certainly on many other people's minds. Were Donohue and Levitt arguing that poor people were intrinsically, perhaps even biologically, predisposed to criminal behavior, or did they believe that the environmental conditions of poverty pushed people into crime? Were they conservatives or liberals, and why had they avoided showing their political hand in their academic presentation?
Add to this confusion the fact that conspiracy theories run rife in today's inner city--theories that AIDS is a form of biological warfare against black populations, that the war on crime and drugs is a none too subtle attempt to destroy the social fabric of the inner city through mass incarceration, and that the easy availability of abortion is a means of racist population control--and this study could hardly be anything but explosive.
At the same time, anti-abortion groups accused the study of advocating the murder of unborn children as a crime control stratagem. "Naturally, if you kill off a million and a half people a year," the executive director of the Pro-Life Action League declared sarcastically in a news release, "a few criminals will be in that number." Meanwhile, many on the left of the political spectrum, including some in the pro-choice movement, denounced the report as little more than a call to arms against the poor. As if it had advocated for forced abortion, sterilization, or euthanasia against select elements of the American population, the study was seen as a bloody addendum to such bibles of the New Right as Richard J. Herrnstein and Charles Murray's 1996 book The Bell Curve.
But this is not, in fact, what the two authors were arguing. In the years after Roe v. Wade, they found, the number of abortions performed each year in America grew rapidly. By 1980 the annual total had reached 1.6 million, a statistic that has remained fairly constant ever since. Before this, the authors argue, more unwanted children were being born, often after unhealthy pregnancies during which the mother failed to look after herself adequately, and often into difficult, non-nurturing, impoverished environments. Such children, Donohue and Levitt assumed, would be more likely than others to grow up to commit crimes as troubled, angry, gang-affiliated teenagers and young adults. The authors cite evidence from studies in eastern Europe and Scandinavia that "unwanted children are likely to be disproportionately involved in criminal activities." And it's certainly plausible that the same would be true in the United States.
Donohue and Levitt's data bear this out. Legalized abortion, they found, didn't just lower the absolute number of people in a given age group; it disproportionately lowered the number of children born to mothers in impoverished circumstances who hadn't intended to become pregnant and gave birth to babies they didn't want. The effect on subsequent teenage arrest rates suggests that these were indeed the conditions that put children "most at risk of engaging in criminal behavior."
Joseph McNamara, a Hoover Institute fellow, sees the connection as simple common sense. "Many years ago, when I was police chief of San Jose," McNamara remembers, "I cooperated with Planned Parenthood, and I said: 'Your organization prevents more crime than mine does.' Children need love and nurturing. If there's no one there to provide that [because a child is unwanted], many are going to commit crimes and violent behavior. You don't have to be a criminologist to see that the children growing up under these conditions are at high risk. It's an enormous problem for society to have children born that no one wants and no one's able to take care of."
Levitt and Donohue think of themselves as researchers, not advocates for any position--not even McNamara's--and they are annoyed that others see their study as promoting forced abortion. "It seems such a puerile logical step," Donohue says. "A completely incorrect reading of the study. One wonders how presumably intelligent people could make such an incorrect logical inference." Far from having any ax to grind about abortion and crime, Donohue, raised as a Catholic, says he resisted making this connection until the evidence he and Levitt compiled proved overwhelming.
But by presenting only their study's results and not the social meaning they saw in it, the two researchers left both the study and themselves vulnerable. "John and I are good at taking data and understanding [it]. We have no expertise in making moral or ethical judgments," Levitt says. "We bring this to the table to add to the debate, but certainly not to solve the debate." They might have diffused much of the criticism leveled against them, however, if they had joined forthrightly in the policy debate from the start, presenting their study as Donohue privately acknowledges they see it--as evidence that get-tough anticrime policies have less effect on crime than most people think and that allowing women to choose when to have children has more.
Still, the more important questions would have remained. Is their startling conclusion about abortion and crime actually true? And if so, are these its implications?
Other academic experts have mostly praised the work, while also noting its limits. The study does not, for instance, explain why crime rose so dramatically in the first place. Obviously, other social changes, unrelated to the availability of abortion in pre-Roe v. Wade days, were at work. Donohue and Levitt themselves talk of the large growth in the drug trade, the movement against prisons, the transformative but chaotic forces released by the civil rights, antiwar, and student movements of the 1960s, and the decline of respect for traditional institutions of authority. But given that these paradigm shifts served as a backdrop to a dramatically rising level of crime in the United States, the authors wanted explanations for why it subsequently fell. And whatever the other crime-reducing changes in the culture, they found that the change in abortion policy had a major magnifying effect on them.
Many scholars feel that Donohue and Levitt's study may exaggerate that impact. Ted Joyce, for example, a professor of economics at New York's Baruch College, says, "I think it's plausible that there's an association between fertility controls and better outcomes for women and children. The question is how much. A 50 percent drop in crime? I'm suspicious."
Joyce believes that the authors overestimated the effects of abortion in traditionally high-crime states like New York and California. In the years leading up to Roe v. Wade, when New York and California were among only five states in the nation that had already legalized abortion, many women traveled to these areas from elsewhere in order to have abortions. Because Levitt and Donohue based their calculations on where abortions were performed rather than where the mothers actually lived, Joyce argues, they ended up comparing apples and oranges. Abortion rates in New York during the early 1970s were as likely to affect subsequent crime rates in New Jersey or Connecticut.
Moreover, Joyce suspects that once abortion became routine and widely available it began to serve as an alternative, albeit radical, form of birth control. Many of those aborted fetuses, would not, Joyce believes, have been born into lives of crime had abortion been less available. Rather, the parents would simply have been more careful to use other methods of birth control to prevent pregnancy in the first place.
But the same reviewers who believe the study has probably overstated the effect of abortion liberalization agree that the authors most likely have uncovered an important mechanism contributing to the lower crime rates. Alfred Blumstein, a professor of public policy at Carnegie Mellon University and one of the country's leading criminal justice experts, views the study as "sophisticated research." It provides hard evidence, he says, in a fresh, innovative format, that poverty, neglect, and crime are linked, that the much maligned "root cause" arguments touted by liberals in the 1960s and 1970s and derided by the lock-'em-up crowd in the 1980s and 1990s can no longer be dismissed. Bring together a child and a certain set of impoverished material circumstances, the authors demonstrate, and there's a fair chance that the mix will, at some stage, produce crime. Abortion takes the child out of this equation. But society could equally well invest in social policies that would try to remove the poverty instead.
Blumstein, too, has quibbles with the research; he says that Levitt and Donohue's regression analysis did not adequately account for such changes as the booming national economy and the shrinking drug markets in California and New York as possible explanations for the decline in crime rates. But the study convinced him that there is some linkage between access to abortion and falling crime. And he says it is a connection that the public needs to contemplate. "Conservatives say, 'People are rotten, lock 'em away,'" Blumstein argues. "And liberals say, 'We have to get at root causes.' The dilemma is how to allocate resources to do both. The message has got to be [that] prevention is key regarding young people and criminality."
A t the moment, conservatives pretty much control the policy conversation about crime. But as Robert Weisberg, a Stanford University Law School colleague of Donohue's, says, get-tough policies from "Giuliani's policing techniques to three-strikes laws were open to the criticism that they were getting undeserved credit. Crime in three-strikes states hasn't fallen by more than crime in other states. The lack of that evidence is even more embarrassing in light of the analytical study Donohue and Levitt have done."
Indeed, their work has the potential to change the parameters of the debate. "The deterrent effect of the regular criminal justice [system] is not trivial," Weisberg asserts, "but it has already been accomplished." At a certain point, Weisberg says, incarceration policy runs into a wall of diminishing returns: With two million people already behind bars, increases in the size of the prison population mainly tend to put more small fry into harsh prison environments from which many emerge more dangerous than they were when they entered. If imprisoning lawbreakers cannot reduce crime beyond a certain point, then attention must turn to studies such as Levitt and Donohue's. Its strong evidence of a link between economic deprivation and crime suggests that the most sensible crime-prevention policy may well be the introduction of better antipoverty programs--perhaps the very social programs that have been dismantled over the past 20 years in favor of prison building. ¤
