At a time when the right to abortion is on the line, two prominent economists argue that legalized abortion has been the nation’s single most powerful antidote to crime over the last 30 years.
John J. Donohue, a Stanford professor of economics and law, and Steven D. Levitt, a University of Chicago economist, first put forth that innovative — and incendiary — proposition in a paper nearly 20 years ago. Using reams of data and statistical modeling, they deduced that after the U.S. Supreme Court’s 1973 decision legalizing abortion, it was disproportionately used by women at high risk of having children who would grow up to be criminals.
The economists estimated that abortion accounted for up to half of the drop in violent and property crime between 1985 and 1997 — and they predicted the pattern would continue as the full effects of legal abortion access played out.
Now, they say, they were so right. By updating their analysis to look at 1997 through 2014, they credit legalized abortion for 45 percent of the decline in U.S. crime rates, which have fallen by about half since the early 1990s.
“It is rare for an economic theory to make predictions for 20 years into the future that are both bold and precise. The abortion-crime hypothesis, however, did just that,” Donohue and Levitt write in a “working paper” circulated last month through the National Bureau of Economic Research.
More than a few scholars disagree. The methodology and the merits of the hypothesis have been subject to intense criticism for years.
“Their story doesn’t add up,” said Theodore Joyce, a professor of economics at Baruch College of the City University of New York. “Criminologists are struggling to explain it. Here in New York City, the homicide rate is now down to 3 per 100,000 [people]. That is nothing less than miraculous. The national rate is 4 per 100,000. But it’s not because of abortion.”
In an interview, Donohue said, “Ted is someone who has worked in this area for a long time. He feels people infringing on his turf, but that’s not a good basis for evaluating the data.”
Outside academic circles, the abortion-cuts-crime theory was popularized when Levitt included it in the 2005 bestseller Freakonomics.
Both Levitt and Donohue have said they are not advocating abortion or pushing a social, political, or racial agenda. They just wanted to understand a puzzling trend: U.S crime rates began falling in the early 1990s, defying respected criminologists who warned America was a ticking crime bomb.
“We were and have continued to be interested in explaining one of the most profound mysteries of American life,” Donohue said.
In addition to crime and abortion data, the economists drew on reproductive health research done in the U.S. and other countries showing that young, disadvantaged, and minority women have higher abortion rates, and that unwanted children are more likely to have troubled lives.
As they recap in the new paper: “Decades of social scientific research have demonstrated that unwanted children are at an elevated risk for less favorable life outcomes including criminal involvement, and the legalization of abortion appears to have dramatically reduced the number of unwanted births.”
Their updated analysis, like the original one, found that crime rates fell earlier and further in the five states that legalized abortion before 1973, and crime fell more in states with high abortion rates than in states with low abortion rates.
Those who doubt abortion drove down crime acknowledge that other explanations — including better policing, more imprisonment, the war on drugs, a strong economy, gentrification, immigration — also have weaknesses.
Besides pointing to technical flaws in the original paper’s methods, critics say the correlation between abortion and crime could simply be circumstantial.
“Correlation isn’t causation,” said Douglas Baird, a law professor at the University of Chicago. As an example of a dubious link, he cited the so-called hemline index, which says dresses get shorter or longer depending on the stock market.
Christopher Foote, senior economist at the Federal Reserve Bank of Boston, echoed: “I see people who eat broccoli are much healthier than those who don’t. I don’t know if it’s the broccoli or something else about those people. I don’t believe [Donohue and Levitt] made a compelling case that crime was falling because of abortion.”
Another issue: U.S. abortion rates have been declining for decades and are now at historic lows. In 2015, the rate was about 12 abortions per 1,000 women, down from the 1980 peak of 29 abortions per 1,000 women, according to federal data. Logically, if abortions reduce crime 15 to 20 years later, then dramatically reducing abortions should produce a spike in crime 15 to 20 years later.
But from 1997 to 2014, property crime rates fell despite declines in abortion, an “anomalous” finding that Donohue and Levitt speculate reflects the birth of fewer unwanted children.
“Drops in abortion are starting to muddy the initial picture for property crime,” Donohue conceded. “The question is, is that inconsistent with the theory. The critics will say, ‘Yeah, that’s inconsistent.’”
“Now that abortion is falling and crime is still falling, they have to explain this,” Joyce said. “They say it’s because of less unwanted pregnancy. In other words, they don’t have an explanation.”
The updated analysis comes as more and more states — including Kentucky, Mississippi, Ohio, Missouri, Georgia, and Alabama — are passing ultra-restrictive abortion laws that violate the constitutional protections laid out by the Supreme Court in Roe v. Wade. Although the laws are not being enforced, abortion foes hope this sets the stage for the high court to revisit and overturn that 1973 landmark. Then abortion legalization would again be up to individual states.
Donohue said the loss of abortion access could show up in future crime rates.
“We could conceivably see the unwinding of the [anti-crime] effect because of restricted abortion access today,” he said. “But it won’t show up for 20 years, so it makes teasing out the effects complicated.”