Chicago economist links abortion to falling crime rates
By Amy Rust News Office
Presented at seminars at the University, Stanford and Harvard but not yet published,
“Legalized Abortion and Crime,” Chicago economist Steven Levitt’s recent study that links the legalization
of abortion to the country’s falling crime rate in the 1990s, already is receiving national attention.
The study, co-authored by Levitt, Professor in Economics at Chicago, and StanfordUniversity’s John Donohue III, suggests legalized abortion may be responsible for approximately half of the crime rate’s recent
According to the researchers, the decline of the U.S. crime rate may be the result of two mechanisms
related to legalized abortion. First, following the Roe vs. Wade decision in 1973, more women at risk of having children who
could later engage in criminal activity—teen-agers, those living in poverty or those with unwanted pregnancies, for
example—opted for abortion. And second, improved maternal, familial or fetal circumstances may have led to better environments
for raising children.
Levitt and Donohue stress that their findings do not carry an endorsement of
abortion. “We do not take a position on abortion, and the study was not undertaken as a study of abortion, but crime,”
said Levitt. “Neither is the study about race or class. Many studies have shown that children who are born unwanted
have unsatisfactory outcomes, including involvement in crime.”
As evidence for their findings, the researchers point to data regarding the timing
of the crime drop: the first generation of pregnancies terminated under legalized abortion would have otherwise resulted in
children who reached the peak ages for criminal activity, 18 to 24, in the early 1990s. Increases in 1970s abortions by high-risk
mothers may have lowered the number of potential criminals coming of age in the 1990s.
The study also reports that states such as California and New York, which legalized abortion before 1973, experienced a drop in their crime
rates before the rest of the nation. Furthermore, empirical evidence suggests states with higher abortion rates have seen
more dramatic decreases in crime since 1985, and those drops in crime have been concentrated among individuals under age 25
in 1997––precisely the group possibly affected by abortion legalization in 1973.
While many explanations have been given for the dramatic decline of crime during
this past decade, the authors maintain in a study abstract that “each of them has difficulty explaining the timing,
large magnitude, persistence and widespread nature of the drop.” The researchers also predict crime rates will continue
to fall slowly for 15 to 20 more years as the full effects of legalized abortion continue.
“A better understanding of the reasons for declines in crime helps policymakers
as they formulate programs to reduce crime. For instance, with lower future crime rates, there may be less need to build prisons,”
Solve a Mystery Meet the economist who figured out that legal abortion was behind
dropping crime rates.
BY STEVEN E. LANDSBURG Wednesday,
April 13, 2005 12:01 a.m. EDT
If Indiana Jones were an economist, he'd be Steven Levitt. The most recent winner of the John Bates Clark
award for the best economist under the age of 40, Mr. Levitt is famous not as a master of dry technical arcana but as a maverick
treasure hunter who relies for success on his wit, pluck and disregard for conventional wisdom. Mr. Levitt's typical quarry
is hidden not in some exotic locale but in a pile of data. His genius is to take a seemingly meaningless set of numbers, ferret
out the telltale pattern and recognize what it means.
It was Mr. Levitt who nailed a bunch of Chicago public-school teachers for artificially inflating their students'
standardized test scores. I'm dying to tell you exactly how he did it, but I don't want to spoil any surprises. His account
of the affair in "Freakonomics" reads like a detective novel.
The evidence is right there in front of you: Mr. Levitt actually reproduces all the answer sheets from two
Chicago classrooms and challenges you to spot the cheater. Then he shows you how it's done. He points to suspicious patterns
that you almost surely overlooked. Suspicious, yes, but not conclusive--maybe there is some legitimate explanation. Except
that Mr. Levitt slowly piles pattern on pattern, ruling out one explanation after another until only the most insidious one
remains. The resulting tour de force is so convincing that it eventually cost 12 Chicago schoolteachers their jobs.
The Case of the Cheating Teachers would make a fascinating book, but in Mr. Levitt's hands it is compressed
into 12 breathtaking pages. Then he is on to his next adventure--the Case of the Cheating Sumo Wrestlers. Here an entirely
different kind of data (the win-loss records from tournaments) gets the Levitt treatment: the identification of a suspicious
pattern, a labyrinth of reasoning to rule out the innocent explanations and a compelling indictment.
Then it's on to another question, and another and another. Were lynchings, as their malevolent perpetrators
hoped, an effective way to keep Southern blacks "in their place"? Do real-estate agents really represent their clients' interests?
Why do so many drug dealers live with their mothers? Which parenting strategies work and which don't? Does a good first name
contribute to success in life?
Mr. Levitt is hardly the first to attack these questions; there is no end of books on parenting strategy,
for example. The difference is that Mr. Levitt knows what he is talking about. Where other parenting books rely on either
puerile psychological theorizing or leaps of logic from haphazard numerical correlations, Mr. Levitt relies on his instinct
for analyzing data. As a result, there is more valuable parenting advice in Mr. Levitt's single chapter than in all the rest
of Barnes & Noble. And some of it is going to shock you. One example: It turns out that reading to your children has no
appreciable effect on their academic success.
Back in 1999, Mr. Levitt was trying to figure out why crime rates had fallen so dramatically
in the previous decade. He was struck by the fact that crime began falling nationwide just 18 years after the Supreme Court
effectively legalized abortion. He was struck harder by the fact that in five states crime began falling three years earlier
than it did everywhere else. These were exactly the five states that had legalized abortion three years before Roe v. Wade.
Did crime fall because hundreds of thousands of prospective criminals had been aborted? Once again, the pattern
by itself is not conclusive, but once again Mr. Levitt piles pattern on pattern until the evidence overwhelms you. The bottom
line? Legalized abortion was the single biggest factor in bringing the crime wave of the 1980s to a screeching halt.
Mr. Levitt repeatedly reminds us that economics is about what is true, not what ought to be
true. To this reviewer's considerable delight, he cheerfully violates this principle at the end of the abortion discussion
by daring to address the question of whether abortion ought to be legal or, more precisely, whether the effect on crime rates
is a sufficient reason to legalize abortion. He doesn't pretend to settle the matter, but in just a few pages he constructs
exactly the right framework for thinking about it and then leaves the reader to draw his own conclusions.
Economists, ever wary of devaluing their currency, tend to be stinting in their praise. I therefore tried
hard to find something in this book that I could complain about. But I give up. Criticizing "Freakonomics" would be like criticizing
a hot fudge sundae. I had briefly planned to gripe about the occasional long and pointless anecdotes, but I changed my mind.
Sure, we get six pages on the Chicago graduate student who barely escaped with his life after his adviser sent him into the
housing projects with a clipboard to survey residents on how they feel about being black and poor. Sure, there is no real
point to the story. But a story that good doesn't need a point.
The cherry on top of the sundae is Mr. Levitt's co-author, Stephen Dubner, a journalist who clearly understands
what he is writing about and explains it in prose that has you chuckling one minute and gasping in amazement the next. Mr.
Dubner is a treasure of the rarest sort; we are fortunate that Mr. Levitt managed to find him. I think I detect a pattern.
Mr. Landsburg, an economics professor at the University of Rochester, is the author of "Armchair Economist:
Economics and Everyday Experience." You can buy "Freakonomics" from the OpinionJournal bookstore.
For the last decade, the crime rate for serious
violent and property offenses has been falling throughout the country -- in some cases, reaching levels not seen since the
mid-1960s. Homicide rates, for instance, decreased a staggering 42 percent; robbery by approximately 40 percent. Rates for
other types of street crime also fell by substantial amounts. Not only are these declines widespread; they are also real,
in the sense that they cannot be ascribed to changes in reporting or in record-keeping.
Ignoring the adage "Don't
look a gift horse in the mouth," experts have offered many explanations for this welcome development. These include "smarter"
policing, greatly increased incarceration, a drop in the crack trade (with its attendant use of firearms), a booming economy
and various demographic factors. But none is as startling or controversial as the thesis of Stanford Law School Professor
John J. Donohue III and University of Chicago Professor Steven D. Levitt.
In a recent article in the Quarterly Journal of Economics, they maintain that legalized
abortion may account for up to 50 percent of the shrinkage in the crime rate. At one level, the impact on crime (if not its
size) is unsurprising. Abortion reduces the number of people in any given age cohort -- including, most significantly, young
men in their peak years for violent crime, about 18 through 24. Donohue and Levitt, however, make the much more sweeping claim
that children born to mothers for whom abortion was an available option have lower per capita rates of offending.
posit two reasons for this phenomenon. First, research has demonstrated that teens and unmarried and poor women are likeliest
to seek abortions. "[T]he early life circumstances of those children on the margin of abortion are difficult along many dimensions,"
such as "growing up in a single-parent family and experiencing poverty." These adverse conditions, including the fact of being
unwanted, have been shown to be strongly linked to future criminality. Second, access to abortion gives all women the choice
of delaying childbearing until they can provide their offspring with as good (and hence, noncriminogenic) an environment as
In support of their thesis, the authors note that legalization -- which occurred in five states in 1970
and three years later in the rest with Roe v. Wade -- dramatically raised the number of abortions. Crime rates began to fall
in the early 1990s (somewhat before, in the earlier-legalizing states), just when the first group born after Roe would attain
"its criminal prime." Donohue's and Levitt's more formal analysis, which controlled for a number of factors influencing crime,
indicates that states with high abortion rates in the 1970s and early 1980s experienced a decline in crime in the 1990s roughly
30 percent greater than that of low-abortion jurisdictions.
Moreover, almost all the decrease occurred in offenses
committed by the post-Roe cohort.
Although the authors emphatically disclaim any normative stance, their study, if
valid, lends ammunition to pro-abortion rights advocates.
STILL AN UNSETTLED QUESTION
may draw comfort from the fact that some scholars have challenged Donohue's and Levitt's conclusions. Professor Ted Joyce
of BaruchCollege, using different data and modes of analysis, sees no correlation between abortion and violent crime. In fact, he has
found that states with higher abortion rates have higher, not lower, fertility rates, which conflicts with the argument that
legalized abortion reduces the number of unwanted births.
So, too, Professors John R. Lott Jr. of YaleLawSchool and John Whitley of Australia's AdelaideUniversity, also using some different data and altering some Donohue-Levitt assumptions, find no negative
correlation between abortion and homicide rates. Indeed, on rather weak evidence, they contend that legalization actually
increases homicides. Their reasoning is that women who won't abort have to compete as sex partners with those who will --
thereby augmenting the number of out-of-wedlock children at risk for future criminal behavior.
and Levitt note that even their view of the data need not support an abortion program; society could, among other things,
furnish better environments for youngsters who might otherwise break the law. That, however, would force politicians to address
intractable "root causes." Sadly, politicians know it is easier to call for nostrums like three-strikes laws and the death
penalty, and so pander to baseless -- but pervasive -- fears that crime rates are continually climbing. In any event, Donohue's
and Levitt's intriguing thesis will surely generate further discussion in years to come. One can only hope that scholarly
research and analysis will shed more light on the subject before zealots on either side of the abortion debate hijack the
issue for partisan ends.
Vivian Berger is a professor emerita at ColumbiaUniversityLawSchool.
From: Steven Levitt To:
Steve Sailer Monday, Aug. 23, 1999, at PT
In recent weeks there
has been a lot of media coverage of a paper John Donohue and I recently wrote connecting the legalization of abortion in the
1970s to reduced crime in the 1990s. (A preliminary version of the paper is posted here.) The purpose of the study is to better understand the reasons
for the sharp decline in crime during this decade, which, prior to our research, had largely eluded explanation. While there
are many other theories as to why crime declined (more prisoners, better policing, the strong economy, the decline of crack,
etc.), most experts agree that none of these very convincingly explains the 30 percent to 40 percent fall in crime since 1991.
The theoretical justification
for our argument rests on two simple assumptions: 1) Legalized abortion leads to fewer "unwanted" babies being born, and 2)
unwanted babies are more likely to suffer abuse and neglect and are therefore at an increased risk for criminal involvement
later in life. The first assumption, that abortion reduces the number of unwanted children, is true virtually by definition.
The second assumption, that unwanted children are at increased risk for criminal involvement, is supported by three decades
of academic research. If one accepts these two assumptions, then a direct mechanism by which the legalization of abortion
can reduce crime has been established. At that point, the question merely becomes: Is the magnitude of the impact large or
research suggests that the effect of abortion legalization is large. According to our estimates, as much as one-half of the
remarkable decline in crime in the 1990s may be attributable to the legalization of abortion. We base our conclusions on four
separate data analyses.
First, we demonstrate
that crime rates began to fall 18 years after the landmark Supreme Court decision Roe vs. Wade legalized abortion across the
nation, just the point at which babies born under legalized abortion would be reaching the peak adolescent crime years. In
my opinion, this is the weakest of our four data analyses. In a simple time series, many factors are negatively correlated
with crime. Furthermore, the world is a complicated place and it would be simplistic to believe that legalized abortion could
overpower all other social determinants of crime.
Second, we show
that the five states that legalized abortion in 1970--three years before Roe vs. Wade--saw crime begin to decrease roughly
three years earlier than the rest of the nation. This is a bit more convincing to me but still far from conclusive.
Third, we demonstrate
that states with high abortion rates in the mid-1970s have had much greater crime decreases in the 1990s than states that
had low abortion rates in the 1970s. This relationship holds true even when we take into account changes in the size of prison
populations, number of police, poverty rates, measures of the economy, changes in welfare generosity, and other changes in
fertility. This is the evidence that really starts to be convincing, in my opinion.
Fourth, we show
that the abortion-related drop in crime is occurring only for those who today are under the age of 25. This is exactly the
age group we would expect to be affected by the legalization of abortion in the early 1970s. That is where our paper stops.
Our paper is a descriptive exercise attempting to explain why crime fell. While our paper highlights one benefit of allowing
women to determine whether or not to bring pregnancies to term, we make no attempt to systematically analyze the many possible
costs and benefits of legalized abortion. Consequently, we can make no judgment as to whether legalized abortion is good or
bad. In no way does our paper endorse abortion as a form of birth control. In no way does our paper suggest that the government
should restrict any woman's right to bear children. Although these are the most interesting issues for the media to discuss,
our paper actually has very little to say on such topics.
I think the crux
of the misinterpretation of our study is that critics of our work fail to see the distinction between identifying a relationship
between social phenomena and endorsing such a relationship. When a scientist presents evidence that global warming is occurring,
it does not mean that he or she favors global warming, but merely that the scientist believes such a phenomenon exists. That
is precisely our position with respect to the link between abortion and crime: We are not arguing that such a relationship
is good or bad, merely that it appears to exist.
As an aside, it
has been both fascinating and disturbing to me how the media have insisted on reporting this as a study about race, when race
really is not an integral part of the story. The link between abortion and unwantedness, and also between unwantedness and
later criminality, have been shown most clearly in Scandinavian data. Abortion rates among African-Americans are higher, but
overall, far more abortions are done by whites. None of our analysis is race-based because the crime data by race is generally
not deemed reliable.
I look forward to
hearing your thoughts. I am interested in your views on the paper and its analysis, but also on the broader topic of the coverage
of scientific research in the popular press, particularly when it relates to sensitive subjects like abortion, crime, or race.
Do you think any good comes from a public discussion of academic studies such as this one? What, if anything, could be done
to make such public debates more productive?
Steven Levitt is a professor
of economics at the University of Chicago
and a research fellow at the American Bar Foundation. His primary research focus is the economics of crime and the criminal