Abortion and crime: one has an effect on the other, but it may not be the effect you think.

VIOLENT crime in the United States shot up like a rocket after 1960. From 1960 to 1991, reported violent crime increased by an incredible 372 percent. This disturbing trend was seen across the country, with robbery peaking in 1991 and rape and aggravated assault following in 1992. But then something unexpected happened: Between 1991 and 2000, rates of violent crime and property crime fell sharply, dropping by 33 percent and 30 percent, respectively. Murder rates were stable up to 1991, but then plunged by a steep 44 percent.

Several plausible explanations have been advanced for the drop during the 1990s. Some stress law-enforcement measures, such as higher arrest and conviction rates, longer prison sentences, "broken windows" police strategies, and the death penalty. Others emphasize right-to-carry laws for concealed handguns, a strong economy, or the waning of the crack-cocaine epidemic.

Of all the explanations, perhaps the most controversial is the one that attributes lower crime rates in the '90s to Roe v. Wade, the Supreme Court's 1973 decision to mandate legalized abortion. According to this argument, the large number of women who began having abortions shortly after Roe were most likely unmarried, in their teens, or poor, and their children would have been "unwanted." Children born in these circumstances would have had a higher-than-average likelihood of becoming criminals, and would have entered their teens--their "criminal prime"--in the early 1990s. But because they were aborted, they were not around to make trouble.

This is an attention-grabbing theory, to be sure. But a thorough analysis of abortion and crime statistics leads to the opposite conclusion: that abortion increases crime.

Debate about the relationship between abortion and crime was greatly influenced by a Swedish study published in 1966 by Hans Forssman and Inga Thuwe. They followed the children of 188 women who were denied abortions from 1939 to 1941 at the only hospital in Gothenburg, Sweden. Their study compared these "unwanted" children with another group, this one composed of the first child born at the hospital after each of the "unwanted" children. They found that the "unwanted" children were much more likely to grow up in adverse conditions--for example, with divorced parents, or in foster homes. These children were also more likely to become delinquents and have trouble in school. Unfortunately, the authors never investigated whether the children's "unwantedness" caused their problems, or were simply correlated with them.

Forssman and Thuwe's claim, notwithstanding the limits of the data supporting it, became axiomatic among supporters of legalized abortion. During the 1960s and '70s, before Roe, abortion-rights advocates attributed all sorts of social ills, including crime and mental illness, to "unwanted" children. Weeding these poor, crime-prone people out of the population through abortion was presented as a way to make society safer.

Recently, John Donohue and Steven Levitt--a law professor and an economist, respectively--revived the debate. They presented evidence that supposedly demonstrated abortion's staggeringly large effect on crime rates, and argued that up to "one-half of the overall crime reduction" between 1991 and 1997, and up to 81 percent of the drop in murder rates during that period, was attributable to the rise in abortions in the early to mid 1970s. If that claim was accurate, they had surely found the Holy Grail of crime reduction.

Most people who challenge the "abortion reduces crime" argument do so on ethical grounds, rather than trying to rebut the empirical evidence. But it is worth looking at the data, too--because they do not prove what they are supposed to.

To understand why abortion might not cut crime, one should first consider how dramatically it changed sexual relationships. Once abortion became widely available, people engaged in much more premarital sex, and also took less care in using contraceptives. Abortion, after all, offered a backup if a woman got pregnant, making premarital sex, and the nonuse of contraception, less risky. In practice, however, many women found that they couldn't go through with an abortion, and out-of-wedlock births soared. Few of these children born out of wedlock were put up for adoption; most women who were unwilling to have abortions were also unwilling to give up their children. Abortion also eliminated the social pressure on men to marry women who got pregnant. All of these outcomes--more out-of-wedlock births, fewer adoptions than expected, and less pressure on men "to do the right thing"--led to a sharp increase in single-parent families.

Multiple studies document this change. From the early 1970s through the late 1980s, as abortion became more and more frequent, there was a tremendous increase in the rate of out-of-wedlock births, from an average of 5 percent (1965-69) to over 16 percent twenty years later (1985-89). Among blacks, the number jumped from 35 percent to 62 percent. While not all of this rise can be attributed to liberalized abortion laws, they were certainly a key contributor.

What happened to all these children raised by single women? No matter how much they want their children, single parents tend to devote less attention to them than married couples do. Single parents are less likely than married parents to read to their children or take them on excursions, and more likely to feel angry at their children or to feel that they are burdensome. Children raised out of wedlock have more social and developmental problems than children of married couples by almost any measure--from grades to school expulsion to disease. Unsurprisingly, children from unmarried families are also more likely to become criminals.

So the opposing lines of argument in the "abortion reduces crime" debate are clear: One side stresses that abortion eliminates "unwanted" children, the other that it increases out-of-wedlock births. The question is: Which consequence of abortion has the bigger impact on crime?

Unfortunately for those who argue that abortion reduces crime, Donohue and Levitt's research suffered from methodological flaws. As The Economist noted, "Donohue and Levitt did not run the test they thought they had." Work by two economists at the Boston Federal Reserve, Christopher Foote and Christopher Goetz, found that, when the test was run correctly, it indicated that abortion actually increases violent crime. John Whitley and I had written an earlier study that found a similar connection between abortion and murder--namely, that legalizing abortion raised the murder rate, on average, by about 7 percent.

The "abortion decreases crime" theory runs into even more problems when the population is analyzed by age group. Suppose that liberalizing abortion in the early 1970s can indeed explain up to 80 percent of the drop in murder during the 1990s, as Donohue and Levitt claim. Deregulating abortion would then reduce criminality first among age groups born after the abortion laws changed, when the "unwanted," crime-prone elements began to be weeded out. Yet when we look at the declining murder rate during the 1990s, we find that this is not the case at all. Instead, murder rates began falling first among an older generation--those over 26--born before Roe. It was only later that criminality among those born after Roe began to decline.

Legalizing abortion increased crime. Those born in the four years after Roe were much more likely to commit murder than those born in the four years prior. This was especially true when they were in their "criminal prime," as shown in the nearby chart.

The "abortion decreases crime" argument gets even weaker when we look at data from Canada. While crime rates in both the United States and Canada began declining at the same time, Canada liberalized its abortion laws much later than the U.S. did. Although Quebec effectively legalized abortion in late 1976, it wasn't until 1988, in a case originating in Ontario, that the Canadian supreme court struck down limits on abortion nationwide. If the legalization of abortion in the U.S. caused crime to begin dropping 18 years later, why did the crime rate begin falling just three years after the comparable legal change in Canada?

Even if abortion did lower crime by culling out "unwanted" children, this effect would be greatly outweighed by the rise in crime associated with the greater incidence of single-parent families that also follows from abortion liberalization. In short, more abortions have brought more crime.

Mr. Lott is a senior research scientist at the University of Maryland. This piece is based on his new book, Freedomnomics, as well as his published academic research.

[GRAPHIC OMITTED]