THE ABORTED CRIME WAVE?

A controversial article links the recent drop in crime to the legalization
of abortion two decades ago

Since the early 1990s crime has fallen annually in the U.S., last year by
about 7 percent.

Many explanations have been put forward for this drop: more police walk the
beat, more people are in prison, the economy has improved, crack use has
fallen, alarms and guards are now widespread. The emphasis given to any one
of these rationales varies, of course, according to philosophical bent or
political expediency. In New York City, for instance, plummeting crime has
been attributed to improved policing.

Yet the decline exists even in cities that have not altered their approach,
such as Los Angeles. The above explanations are unsatisfactory to many
researchers, among them two economists who have studied crime.

Steven D. Levitt of the University of Chicago and John J. Donohue III,
currently at Yale University, have proffered an alternative reason: the
legalization of abortion in 1973 reduced the number of unwanted
children--that is, children more likely to become criminals.

In 1992, the first year crime began to fall, the first set of children born
after 1973 turned 18. Because most crimes are committed by young adult
males between the ages of 18 and 24, Levitt and Donohue argue that the
absence of millions of unwanted children led to fewer crimes being done by
that age group.

In total, the researchers maintain, the advent of legal abortion may be
responsible for up to 50 percent of the drop in crime. Their hypothesis,
presented in the as yet unpublished paper "Legalized Abortion and Crime,"
has triggered everything from admiration for its innovative thinking to
outrage for its implications. Groups on both sides of the abortion divide
remain wary: some right-to-life representatives describe the findings as
strange, while pro-choice groups worry that the conclusions will make
people view abortion as a vehicle for social cleansing.

The response has shocked both academics.

The work "is not proscriptive, but descriptive," Levitt maintains. "Neither
of us has an agenda with regard to abortion." Some economists, for their
part, want questions answered about certain aspects of the methodology--and
they want more evidence. "Most interesting is that they put forth an
alternative explanation that is conceivably possible," says Phillip B.
Levine, an economist at Wellesley College. "In terms of the evidence, I
think it is somewhat suggestive. I wouldn't go so far as to say it is
conclusive." Levine also points out that although the paper surprised the
public, it actually follows logically from previous work in this area.

Indeed, Levitt and Donohue are not the first to connect crime and abortion.
As they note in their paper, a former Minneapolis police chief made the
same suggestion several years ago. But they are the first to examine data
to determine whether there could be a correlation. They looked at how crime
rates differed for states that legalized abortion before the U.S. Supreme
Court decision on Roe v. Wade: New York, Washington, Alaska and Hawaii. In
those states, crime began to drop a few years before it did in the rest of
the country, and states with higher abortion rates have had steeper drops
in crime. Fewer unwanted children, the two conclude, ultimately means fewer
crimes.

The idea that unwantedness could adversely affect children is also not new.
Levine and several colleagues explored the economic and social
ramifications for children of the legalization of abortion in a paper
published earlier this year in the Quarterly Journal of Economics. They
estimated that children who were aborted would have been from "40 to 60
percent more likely to live in a single-parent family, to live in poverty,
to receive welfare, and to die as an infant."

Real-world evidence also links unwantedness to some poor outcomes for
children. A 1995 Institute of Medicine report, The Best Intentions:
Unintended Pregnancy and the Well-Being of Children and Families, reviewed
studies on this topic, concluding that women who did not mean to get
pregnant were more likely to expose their fetus to harmful substances and
that these children were at higher risk for low birth weight and abuse. And
a few long-term studies have found an association between unwantedness and
criminality. Levitt and Donohue cite a handful of European studies that
have followed for several decades children born to women who were denied
abortions they had requested--repeatedly, in some cases.

These studies did find that unwanted children had somewhat higher rates of
criminality and psychiatric troubles. "It is correct that there is more
evidence of difficult behavior and criminal behavior," says Henry P. David,
co-author of an ongoing 38-year study of unwanted kids in Prague and an
editor of the 1988 review Born Unwanted: Developmental Effects of Denied
Abortion. "But the numbers are small; it would be difficult to say that
they became criminals because of unwantedness. Certainly that was a factor,
but we don't know how much."

The "how much" seems the crux of the matter for some economists. Theodore
J. Joyce of Baruch College argues that when Levitt and Donohue factor in
regional variability, the strength of their correlation vanishes.

In other words, one of their own charts seems to suggest that some
underlying--and unspecified--differences ("omitted variables," as they
write) between the regions explain the drop in crime, not the abortion
rate, he says. In addition, Joyce and other scholars note that relying on
abortion occurrence data is problematic. Levitt and Donohue use figures for
the number of abortions performed in a state--which do not specify whether
the woman came from out of state.

When Joyce recently reviewed estimates for abortions by state of origin
that were made in the early 1970s by the Alan Guttmacher Institute in New
York City, he says he found that 30 percent of New York's abortions were
performed on women from elsewhere.

Such dramatic interstate movement was not accounted for in Levitt and
Donohue's paper, Joyce states, and it suggests that their correlations
could be off-kilter. "To say that legalization has some kind of effect is
certainly plausible," he concludes. "But I think it should be questioned
because the magnitude of the finding is so large: 50 percent seems way too
large." Despite these concerns, scholars generally agree that Levitt and
Donohue are asking a reasonable question.

And if the two are right, the association should show up in other realms as
well: teenage pregnancy should be dropping, as should adolescent and young
adult suicide, unemployment, and high school dropout rates, and education
levels should be rising. Levitt says that the 2000 census will allow
researchers to investigate some of those other correlates but that for now
he and Donohue are focusing on teen pregnancy.

At first glance, at least, their expectation seems to be holding up. A 1998
article in Pediatrics notes that teen pregnancy has been declining steadily
this decade--a total of 13 percent between 1991 and 1995--and the extent of
the decline varies enormously by state and ethnicity. In addition, teenage
and young adult behavior is changing on many fronts.

In 1994 and 1995, notes Laura D. Lindberg of the Urban Institute in
Washington, D.C., drug use, sexual activity and suicidal ideation began to
decline in adolescents after what had seemed a never-ending increase. "But
how you connect very recent declines with [Levitt and Donohue's] idea of a
shock to the system is very unclear," Lindberg cautions. "Many things are
changing over time." So the jury remains out. Researchers are waiting to
see whether the paper withstands ongoing scrutiny and whether other
evidence emerges. "It is a fascinating theory," David declares. "I suspect
there is some kernel of truth, but how much is hard to say."