Abortion and crime: who should you believe?

Two very vocal critics, Steve Sailer and John Lott, have been exerting a lot of energy lately trying to convince the world that the abortion reduces crime hypothesis is not correct. A number of readers have asked me to respond to these criticisms. First, let’s start by reviewing the basic facts that support the Donohue-Levitt hypothesis that legalized abortion in the 1970s explains a substantial part of the crime decline in the 1990s:

1) Five states legalized abortion three years before Roe v. Wade. Crime started falling three years earlier in these states, with property crime (done by younger people) falling before violent crime.

2) After abortion was legalized, the availability of abortions differed dramatically across states. In some states like North Dakota and in parts of the deep South, it was virtually impossible to get an abortion even after Roe v. Wade. If one compares states that had high abortion rates in the mid 1970s to states that had low abortion rates in the mid 1970s, you see the following patterns with crime. For the period from 1973-1988, the two sets of states (high abortion states and low abortion states) have nearly identical crime patterns. Note, that this is a period before the generations exposed to legalized abortion are old enough to do much crime. So this is exactly what the Donohue-Levitt theory predicts. But from the period 1985-1997, when the post Roe cohort is reaching peak crime ages, the high abortion states see a decline in crime of 30% relative to the low abortion states. Our original data ended in 1997. If one updated the study, the results would be similar.)

3) All of the decline in crime from 1985-1997 experienced by high abortion states relative to low abortion states is concentrated among the age groups born after Roe v. Wade. For people born before abortion legalization, there is no difference in the crime patterns for high abortion and low abortion states, just as the Donohue-Levitt theory predicts.

4) When we compare arrest rates of people born in the same state, just before and just after abortion legalization, we once again see the identical pattern of lower arrest rates for those born after legalization than before.

5) The evidence from Canada, Australia, and Romania also support the hypothesis that abortion reduces crime.

6) Studies have shown a reduction in infanticide, teen age drug use, and teen age childbearing consistent with the theory that abortion will reduce other social ills similar to crime.

These six points all support the hypothesis. There is one fact that, without more careful analysis, argues against the Donohue-Levitt story:

7) The homicide rate of young males (especially young Black males) temporarily skyrocketed in the late 1980s, especially in urban centers like Los Angeles, New York City, and Washington, DC, before returning to regular levels soon thereafter. These young males who were hitting their peak crime years were born right around the time abortion was legalized.

If you look at the serious criticisms that have been leveled against the Donohue-Levitt hypothesis, virtually all of them revolve around this spike in homicide by young men in the late 1980s-early 1990s. (There are also some non-serious criticisms, which I will address below.) This is the point that Sailer is making, and also the point made far more rigorously by Ted Joyce in an article published in the Journal of Human Resources.

So, a reasonable thing to ask yourself is: Was there anything else going on in the late 1980s that might be causing young Black males to be killing each other at alarming rates that might be swamping the impact of legalized abortion over a short time period? The obvious culprit you might think about is crack cocaine. Crack cocaine was hitting the inner cities at exactly this time, disproportionately affecting minorities, and the violence was heavily concentrated among young Black males such as the gang members we write about in Freakonomics. So to figure out whether this spike in young Black male homicides is evidence against legalized abortion reducing crime, or even evidence legalized abortion causes crime, one needs to control for the crack epidemic to find the answer. This is the argument that I have been making for years. First in the Slate exchange with Steve Sailer back in 1999, then in the Donohue and Levitt response to Ted Joyce, and now in a recent paper by Roland Fryer, Paul Heaton, me, and Kevin Murphy.

The key points I mentioned in Slate five years ago in debating Sailer are reprinted below:

Your hypothesis that crack, not abortion, is the story, provides a testable alternative to our explanation of the facts. You argue:

The arrival of crack led to large increases in crime rates between 1985 and the early ’90s, particularly for inner-city African-American youths. The fall of the crack epidemic left many of the bad apples of this cohort dead, imprisoned, or scared straight. Consequently, not only did crime fall back to its original pre-crack level, but actually dropped even further in a “overshoot” effect.
States that had high abortion rates in the ’70s were hit harder by the crack epidemic, thus any link between falling crime in the ’90s and abortion rates in the ’70s is spurious.

If either assumption 1 or 2 is true, then the crack epidemic can explain some of the rise and fall in crime in the ’80s and ’90s. In order for your crack hypothesis to undermine the “abortion reduces crime” theory, however, all three assumptions must hold true.

So, let’s look at the assumptions one by one and see how they fare.

1)Did the arrival of crack lead to rising youth crime? Yes. No argument from me here.

2) Did the decline in crack lead to a “boomerang” effect in which crime actually fell by more than it had risen with the arrival of crack? Unfortunately for your story, the empirical evidence overwhelmingly rejects this claim. Using specifications similar to those in our paper, we find that the states with the biggest increases in murder over the rising crack years (1985-91) did see murder rates fall faster between 1991 and 1997. But for every 10 percent that murder rose between 1985 and 1991, it fell by only 2.6 percent between 1991 and 1997. For your story to explain the decline in crime that we attribute to legalized abortion, this estimate would have to be about five times bigger. Moreover, for violent crime and property crime, increases in these crimes over the period 1985-91 are actually associated with increases in the period 1991-97 as well. In other words, for crimes other than murder, the impact of crack is not even in the right direction for your story.

3) Were high-abortion-rate states in the ’70s hit harder by the crack epidemic in the ’90s? Given the preceding paragraph, this is a moot point, because all three assumptions must be true to undermine the abortion story, but let’s look anyway. A reasonable proxy for how hard the crack epidemic hit a state is the rise in crime in that state over the period 1985-91. Your theory requires a large positive correlation between abortion rates in a state in the ’70s and the rise in crime in that state between 1985 and 1991. In fact the actual correlations, depending on the crime category, range between -.32 and +.09 Thus, the claim that high-abortion states are the same states that were hit hardest by crack is not true empirically. While some states with high abortion rates did have a lot of crack (e.g., New York and D.C.), Vermont, Kansas, Hawaii, Massachusetts, and Washington were among the 10 states with the highest abortion rates in the ’70s. These were not exactly the epicenters of the crack epidemic.

So, what is the final tally? Two of the key assumptions underlying your alternative hypothesis appear to be false: The retreat of crack has not led to an “overshoot” in crime, causing it to be lower than 1985, and even if it had, the states with high abortion rates in the ’70s do not appear to be affected particularly strongly by the crack epidemic. Moreover, when we re-run our analysis controlling for both changes in crime rates from 1985 to 1991 and the level of crime in 1991, the abortion variable comes in just as strongly as in our original analysis.

Re-reading this response five years later, it still sounds pretty good to me. Interestingly, at the time, Sailer refused to respond directly to my arguments. His response in Slate completely side-stepped the fact that I had destroyed his core argument. He wrote, for instance, “…rather than mud wrestle in numbers here, I’ll privately send you my technical suggestions. In this essay I’ll step back and explain why this straightforward insight [that abortion reduces crime] might not work in practice.” I should note that I am still waiting for those technical suggestions he promised to arrive!! And if you compare his Slate arguments to his “new” article in the American Conservative, you will see that his thinking has not progressed very far on the issue. In contrast, I spent two years working on that paper on crack cocaine, which provides hard, quantitative evidence in favor of those earlier conjectures I had made.

Now let’s talk about John Lott for a minute. Along with John Whitley, he wrote a paper on abortion and crime. It is so loaded with inaccurate claims, errors and statistical mistakes that I hate to even provide a link to it, but for the sake of completeness you can find it here. Virtually nothing in this paper is correct, and it is no coincidence that four years later it remains unpublished. In a letter to the editor at Wall Street Journal, Lott claims that our results are driven by the particular measure of abortions that we used in the first paper. I guess he never bothered to read our response to Joyce in which we show in Table 1 that the results are nearly identical when we use his preferred data source. It is understandable that he could make this argument five years ago, but why would he persist in making it in 2005 when it has been definitively shown to be false? (I’ll let you put on your Freakonomics-thinking-hat and figure out the answer to that last question.) As Lott and Whitley are by now well aware, the statistical results they get in that paper are an artifact of some bizarre choices they made and any reasonable treatment of the data returns our initial results. (Even Ted Joyce, our critic, acknowledges that the basic patterns in the data we report are there, which Lott and Whitley were trying to challenge.)

To anyone who actually made it this far, I applaud you for your patience. Let me simply end with an analogy. Let’s say that we are living in a world in which global warming is taking place, but also a world in which El Nino occasionally leads to radical, short run disruptions in normal weather patterns. You wouldn’t argue that global warming is false because for a year or two we had cold winters. You’d want to figure out what effect El Nino has on winter weather and then see whether controlling for El Nino it looks like global warming is taking place. The impact of legalized abortion on crime is a lot like global warming — it is slow and steady and grows a little year by year. Crack is like El Nino, it comes in with a fury and then largely disappears. That is why I have invested so much time and effort in understanding both abortion and crack, and why the criticisms made against the abortion-reduces-crime hypothesis to date have not been very compelling.