A Response to “The Impact of Legalized Abortion on Crime”
The May 2001 issue of Harvard's Quarterly Journal of Economics published the controversial study entitled "The Impact Of Legalized Abortion On Crime," http://mitpress.mit.edu/journals/QJEC/Donohue.pdf, by Stanford's John J. Donohue III and the University of Chicago's Stephen D. Levitt. The authors conclude that legalized abortion accounts for as much as 50% of the drop in crime experienced in the 1990's. This, they say, has a social benefit of $30 billion annually. They note that crime began to fall in 1991 roughly eighteen years after abortion became legalized. Donohue and Levitt claim that research indicates that "unwanted" babies are more likely to become criminals. They claim that since abortion cuts down on the number of "unwanted" babies, it tends to eliminate those most likely to commit crime when they are 18 to 24 years old.
Donohue and Levitt’s ideas were circulated first in August of 1999 and got a lot of attention then. The decision by the Quarterly Journal of Economics to publish this article now seems to be generating renewed interest in the theory. Unfortunately, the publication of this study in such a prestigious journal gives the theory more credibility. Perhaps most disturbing to pro-lifers should be the attention this study has already obtained at major universities. For example, a simple Internet search using the title of the study shows that this study is being referenced in various university publications and is a subject of discussion in university classes. At least one of the authors, Donohue, was invited to speak about the study at a major university (Cornell, 9/25/2000). Many seem to be uncritically accepting the study. Some, to describe their reaction, are using words like fascinating, novel, remarkable.
The purpose of this response is to review the current status of the study and to focus on some authors who have either directly challenged the Donohue & Levitt theory from a statistical perspective or have provided research showing that other factors really explain the crime drop in the 1990’s. Hopefully this will give pro-lifers some ideas for their own responses to the study.
An example of just how persuasive this study seems to be to some people is contained in what I found to be a very disappointing column by George Will in the April 30, 2001 issue of Newsweek (More Abortions, Fewer Crimes? http://stacks.msnbc.com/news/562756.asp?cp1=1#BODY ). He says "Now, nothing in the Donohue-Levitt paper is shocking, or even counter intuitive." Later he writes, "But Donohue and Levitt are no more advocating abortion than Galileo was 'advocating' planetary motion." Will also says, “And Donohue and Levitt come to common-sense conclusions like this: ‘Given that homicide rates of black youths are roughly nine times higher than those of white youths, racial differences in the fertility effects of abortion are likely to translate into greater homicide reductions.’”
Will' s article seems more like a defense of Donohue and Levitt than a critical analysis. Will seems to agree with their basic conclusions. We should note, however, that Will does end on a bit of a high note by questioning: “Furthermore, here is a pertinent question, albeit one difficult to research: Does the policy of abortion-on-demand, which reduces children to “choices” and pregnancies to casually disposable inconveniences, contribute to the mentality that does make many children-not just pregnancies-"unwanted" by their mothers? In which case, the abortion culture itself is an incubator of crime. If this and other issues raised by the Donohue-Levitt paper make people uncomfortable, good.”
It is true that in the last paragraph of their paper, Donohue and Levitt warned that their linking of falling crime to legalized abortion should not be misinterpreted as an endorsement of abortion. They suggest that equivalent reductions in crime could “… be obtained through alternatives for abortion, such as more effective birth control, or providing better environments for those children at greatest risk for future crime.” While a true pro-lifer must strongly reject any promotion of birth control, the second suggestion about better environments has merit, if, by it, Donohue and Levitt are implying such things as adoption and the strengthening of families. Nonetheless, Planned Parenthood’s spin on this whole study will likely be that we need to invest even more money into contraceptive programs (Title X) to prevent “unwanted” children because “unwanted” children are more likely to become criminals. Notice that Donohue and Levitt do not specifically mention such things as abstinence for the unmarried, adoption, or the strengthening of family life in this critical summation. They do, however, suggest “more effective birth control” which, of course, includes many so-called contraceptives, which frequently act to prevent implantation; thus, they are really causing very early deaths.
Donohue made comments during a May 14 , 2001 FOX Special Report with Brit Hume which neatly sum up his theory and agenda. Donohue said: “Children who are unwanted do badly across the board in many life outcomes, and in general, we find that unwanted children have a much higher risk of being involved in criminal activity… If one is uncomfortable with legalized abortion, maybe focusing one's efforts on reducing unwanted pregnancies would be the way to go.” (Transcript # 051403cb.254) This is classic Planned Parenthood-like rhetoric. Certainly Donohue must realize that many originally "unwanted" children are subsequently loved and properly cared for by their parents. It is the parents that should bear the responsibility and the consequences, not the poor innocent children who are killed under the selfish banner of "unwantedness" either by surgical abortion or by the often abortifacient effects of many of the most common types of birth control drugs and devices.
Donohue and Levitt’s study, though not explicitly focused on race, should be of special concern to blacks. In 1997 blacks accounted for 35.9% of the induced abortions reported to the CDC while they only comprised about 12 % of the population. Blacks are being disproportionally aborted as it is. When one considers that the idea that blacks are more likely to be involved in crime (see Will’s statement above) is being combined with the Donohue & Levitt thesis, one should be concerned that there will be increased efforts to have black women abort their children. It is not unreasonable to predict that those with racist or eugenic leanings will use this study to promote their agenda.
Perhaps the earliest and most vocal critic of the Donohue and Levitt study who focused on its statistical defects is writer and businessman Steve Sailer. On August 23-24, 1999 Sailer and Levitt exchanged a series of letters on slate.msn.com.
On August 24, 1999 the National Post of Canada ran Sailer's "No truth to the rumour abortion reduces crime" (http://www.isteve.com/np-abort.htm). There Sailer says, in part:
“...Since moralizing is easy while analyzing is hard, almost all pundits have merely assumed Levitt's and Donohue's highly statistical 63-page paper is correct, and then went on to chatter about the meaning of it all. This theory is so important, though, that it demands intense scrutiny. Fortunately, we can easily test the most obvious prediction of their logic: that babies born soon after 1973 should have grown up to be especially law-abiding teens in the early '90s. Did they?”
“Not exactly. Instead, they went on the worst youth murder spree in U.S. history. According to FBI statistics, the murder rate in 1993 for 14- to 17-year-olds (born in the high abortion years of 1975-1979) was a horrifying 3.6 times higher than that of the kids who were the same age in 1984 (who were born in the pre-legalization years of 1966-1970). In dramatic contrast, over the same time span the murder rate for those 25 and over (all born before legalization) dropped 6%.”
“What about just black male youths? Since their mothers were having abortions at three times the white rate, their murder rate should have fallen spectacularly from 1984 to 1993. Instead it grew an apocalyptic 5.1 times.”
“Why, then, is this generation born in 1975-1979 now committing relatively fewer crimes as it ages? It makes no sense to give the credit to abortion. Instead, it's the rise and fall of the crack cocaine epidemic that largely drove crime first up, then down. Thus, the crime rate has fallen fastest exactly where it had previously grown fastest due to crack -- in the biggest cities and among young black males. This generation born right after legalization is better behaved today in large part because so many of its bad apples are now confined to prisons, wheelchairs, and coffins..."
That data that Sailer cites from the FBI is available in chart form at the US Department of Justice Web site at: http://www.ojp.usdoj.gov/bjs/homicide/ageracesex.htm#oars. Scroll down to the chart titled "homicide offending by age, gender, and race 1976-99." This chart clearly supports Sailer and shows simply that Levitt and Donohue's theory just doesn't predict what really happened in 1993. Sailer's argument is simple, and he makes several other excellent points as well. Anyone serious about opposing Levitt and Donohue’s conclusions should visit all three of the web sites noted previously which contain Sailer's arguments. Now that the study has been rehashed, it is time to also do the same for the criticisms.
Levitt and Donohue actually thank Steve Sailer, and several others, in the current version of the study published in May 2001. Yet they do not adequately address his criticism noted above. In footnote 21, Donohue and Levitt admit to the inconsistency between their predications and the disaggregated time-series data that has been pointed out by their critics, but Donohue & Levitt gloss over that fact with a very weak response.
The Donohue and Levitt study contains a graph entitled "Figure II" which plots per capita violent crime, property crime, and murder 1973-1999. (This graph is on page 392 in the May issue of Quarterly Journal Of Economics, but it has been clipped out and placed at http://www.all.org/stopp/images/figure2.gif as well for quicker viewing.)
The graph clearly shows a wave of crime that increased in the late 80's and peaked in the early 90's and then declined. Sailer, and others (below) have pointed to the crack epidemic to explain both the rise and the fall. Levitt and Donohue offer no convincing explanation for the rise. Had the graph been relatively steady prior to 1991 (1973 + 18) one might be more inclined to accept their theory.
Another good source who casts doubt upon the Donohue and Levitt theory is Alfred Blumstein, a Carnegie-Mellon University professor who is widely recognized as being among the leading criminologists in the nation. He is also the director of the National Consortium On Violence Research. On May 17, 2001 I spoke with Dr. Blumstein about the Donohue and Levitt thesis. Dr. Blumstein said they were “off the wall” with their claim that abortion accounted for up to 50% of the crime drop in the 1990’s. We spoke about Donohue & Levitt’s footnote 21, and he agreed that it was a weak response to criticism raised by Sailer and others. Dr. Blumstein cited several factors for the crime drop including the waxing & waning of the crack cocaine epidemic. Dr. Blumstein also recommended that I speak to Professor Ted Joyce, (see below) who has a counter study using the same data that shows no effect on the crime rate from legalized abortion.
Professor Alfred Blumstein and Joel Wallman edited a work entitled "The Crime Drop In America," which was published in the fall of 2000 by Cambridge University Press. This book is another useful resource for anyone serious about opposing the Donohue and Levitt theory. The book consists of nine chapters authored by experts in the field who review several factors that combined to cause the crime drop. Dr. Willard Oliver of Radford University has a review of the book at: http://www.scja.net/oliver2.html . Therein he says, in part:
“The reader is left with the assessment that 25 percent of the drop was the result of the prison expansion movement and that the significant drops in crime during the 1990s was also a result of changing drug use, increased gun control/intervention efforts, changes in adult and juvenile homicide rates, changes in policing, the labor market, and basic demographics. In other words, the crime drop was caused by a host of key factors mostly resulting from changes in political, economic, and social conditions, thus creating an intricate web of causes affecting the crime rates since 1992.”
On May 18, 2001 I spoke with Dr. Ted Joyce of Baruch College and the National Bureau of Economic Research. He obtained the data sets from Levitt & Donohue and recently completed his own 74 page analysis entitled “Did Legalized Abortion Lower Crime?” which refutes the Donohue & Levitt theory. He will be debating Donohue and Levitt at the November, 2001 meeting of the American Society of Criminology. Dr. Joyce kindly sent me a copy of his, as yet, unpublished study. He also told me that he is “pro-choice.” This, obviously, makes unlikely any dismissal of his findings based upon claims of pro-life bias. Here is what he says, in part, in the abstract:
“... I analyze changes in homicide and arrest rates among teens and young adults born before and after 1970 in states that legalized abortion prior to Roe v. Wade. I compare these changes with variation in homicide and arrest rates among cohorts from the same period but who were unexposed to legalized abortion. I find little evidence to support the claim that legalized abortion caused the reduction in crime. I conclude that the association between abortion and crime is not causal, but most likely the result of confounding from unmeasured period effects such as changes in crack cocaine use and its spillover effects.”
Others critics have pointed out that if what Donohue and Levitt say is true, there should be some evidence of the same pattern in other countries. Here’s a bit of a report from the May 17, 2001 issue of the National Post Online:
“David Murray, the director of the Statistical Assessment Service, a non-partisan think-tank in Washington, said the study poses an intriguing argument, but does not stand up to scrutiny. He said using the authors' hypothesis, crime rates in other countries with abortion access should have seen a similar dip in crime. But in Great Britain, which liberalized abortion in 1968, violent crime has risen dramatically in the past decade. 'And the country with the highest abortion rate in the world is Russia. It should be a worker's paradise of safety by their argument,' he said. 'Well, they've got all kinds of crises happening and it's not a safe place.' ” http://www.nationalpost.com/home/story.html?f=/stories/20010517/565236.html
Steve Sailer refers to legal abortion, in light of the Levitt and Donohue thesis, as "prenatal capital punishment." It is relatively rare that convicted adult criminals receive the death penalty even after actually committing very serious crimes. Yet, sadly, if one follows the Donohue and Levitt train of thought, abortion is a means of executing large numbers of potential criminals in the womb some 18-24 years before they might commit crime. The whole concept reeks of eugenics. Margaret Sanger, a proponent of eugenics and the founder of Planned Parenthood, certainly would welcome the Donohue and Levitt study if she were alive today.
Hopefully some of the resources above may be of use to pro-lifers looking for ways to respond to “The Impact of Legal Abortion on Crime.”