Steven Levitt on Bill Bennett:

Former Education Secretary and Drug Czar Bill Bennett caused a stir when, on his talk show, he suggested that aborting all black babies would cause the crime rate to drop. Bennett made clear that such a policy would be "impossible, ridiculous, and morally reprehensible." Nonetheless, the remarks touched a nerve, insofar as they reinforced racial stereotypes.

Insofar as Bennett's remarks drew on the controversial work of economists Steven Levitt and John Donohue on the relationship between abortion and crime rates, Levitt's comments on the Bennett controversy are particularly interesting. According to Levitt, writing on his Freakonomics blog,
It is true that, on average, crime involvement in the U.S. is higher among blacks than whites. Importantly, however, once you control for income, the likelihood of growing up in a female-headed household, having a teenage mother, and how urban the environment is, the importance of race disappears for all crimes except homicide. . . . In other words, for most crimes a white person and a black person who grow up next door to each other with similar incomes and the same family structure would be predicted to have the same crime involvement. . . . He made a factual statement (if you prohibit any group from reproducing, then the crime rate will go down), and then he noted that just because a statement is true, it doesn't mean that it is desirable or moral. That is, of course, an incredibly important distinction and one that we make over and over in Freakonomics.
While Levitt seems to think Bennett's statement was generally defensible -- if ill-advised, as many other analogies could have been used to make Bennett's argument -- he does have some criticism of the Drug Czar turned talkshow host.
There is one thing I would take Bennett to task for: first saying that he doesn't believe our abortion-crime hypothesis but then revealing that he does believe it with his comments about black babies. You can't have it both ways.
Levitt's whole post is interesting, as are many of the extensive comments.

In a different vein, Scrappleface has a different take on the Bennett brouhaha.

Justin Kee (mail):
Steven Levitt's claim in his chapter on the abortion-crime hypothesis, "Where did all the criminals go?" (if I recall correctly), was that the specific bias of race was controlled for and discounted in his analysis. The salient point he raised concerned the preganant women who were not ready to have a child, i.e. single women that were the head of the household, teenage women, lower-income women, etc., who were aborted is strongly correlated with the drop in crime a decade and a half later. The fact that William Bennett, "off-the-cuff", used race to illustrate his point suggests that he did not completely understand Levitt's analysis and he read into the analysis an element that does not exist.

Yes, Mr. Bennett disavowed the premise of his hypothetical in his following statement. That said, his hypothetical could be construed as illogically following from Mr. Levitt's analysis. Nowhere does Levitt argue for abortion in general or for the abortion for specific categories of pregant women in order to reduce the crime rate. Apart from a defense of mistake, the fact that Mr. Bennett choose to add a specific instance of an irrelevant category to his hypothetical speaks volumes, does it not?
10.1.2005 5:42pm
truetanus (mail):
Ha! Bill should be taken to task not for anything he said being some endorsement of racism or segregation or drowning black children at birth, but rather for not sensing the potential disaster he was walking into. He isnt some wet behind the ears novice, or a hollywood celebrity. He is a professional at this sort of social commentary and should know better. I would say all bets are off as to what happens to him as a result of this, and I hope he sticks to his guns, while also admitting that he could have thought of fifty better ways to say it. It will be one of those gaffes that haunt you and replay in your head for the rest of your life. That will be punishment enough.
10.1.2005 5:48pm
DNL (mail):

If that's the case, remember that he hasn't apologized. It is possible therefore that he was fully aware of the reprecussions, and figured his statements would get wide media play.
10.1.2005 6:49pm
Henry Schaffer (mail):
"Half of the population is below average." is a statistical truism and a popular joke. Elimination of below average portions of the population will raise the mean of the remaining portion. (Justin Kee doesn't seem to understand that the statistical controls he mentions don't change this point.)

Bennet could have chosen any group in one half of the population to illustrate his point - that its elimination would change the mean of the rest of the population in the other direction. Was his inflammatory choice "accidental"?

Bennet (if he really didn't mean to start a controversy), and also Summers, seem to have overlooked the difference between informal musing and public utterances of well known people. I'd expect such capable, experienced people to be more aware of what they are saying.
10.1.2005 7:42pm
In the sole context of Mr. Bennett's discussion, his hypothetical makes a certain convoluted sense; even if it is not 'politically sensitive.' I would assume that most of his listeners are unaware of Levitt's work; so any critique of his usage of a hypothetical in that context would be irrelevant and misleading in that I don't recall that Bennett actually referenced Levitt in his commentary.

Could he have made his point using a simpler, more clear example? Certainly. In the end, however, he did make his point. He has encouraged some discussion on the issue. He has drawn out the knee-jerk reactionaries; e.g., the 'professional victims' who will attempt to turn any "conservative" issued sound bite into racist commentary. And, he has shown, once again, the double standard in our society - Why do people like Howard Stern and Al Franken get to make similarly racist (not to mention sexist, mysogynist, or just asnine) statements, totally without pretense to an 'intellectual' discourse, while someone like Bill Bennett cannot use such a hypothetical, clearly pointing out that it would be reprehesnible to suggest seriously or carry out, without being castigated?

Is it because all listeners understand that people like Stern or Franken really don't mean it as liberals; while Bill Bennett is obviously a racist because he's conservative? Say that I'm perhaps too cyncial, but do I detect agenda driven politics and news coverage at work again?
10.1.2005 7:44pm
orson23 (mail):
The ignorant clucking above reveals more about the writers than Bennett. Consider the Levitt's own disingenuity:

"There is one thing I would take Bennett to task for: first saying that he doesn't believe our abortion-crime hypothesis but then revealing that he does believe it with his comments about black babies. You can't have it both ways."

Excuse me? If we are dealing with deductive truths based on social science data, exactly where does "belief" enter into it? It doesn't. So why does Levitt cluck about having Bennett's "beliefs" "both ways?"

What Levitt cliams is not, in fact, what Bennet said:
"BENNETT: I would not argue for the pro-life position based on this [utilitarian argument], because you don't know [all the reasons]. I mean, it cuts both [ways].... But I do know that it's [deductively] true that if you wanted to reduce crime..., you could abort every black baby in this country, and your [overall] crime rate would go down."

It's deductively true because of the racial skew in crime.

All this reveals is that when it comes to the left's sacred cows after Katrina, we still insist on talking dishonestly about the responsibility of blacks to blacks (black Mayor Nagin to New Orleans predominantly black citizens), or of a woman governor (Blanco) to her state's people because she is a woman. Both remain protected from honest productive criticism because of Id Pol.

People died needlessly in a hurrican's aftermath because of civic leaders failures, but no - we can't discuss responsibility honestly because right-wingers (like Bennett) are still evil sexist/homophobic, oh! and racist. Never mind that no one died because of Bennett's deductively valid remarks, but many hundreds died needlessly because of the formers civic actions. Furthermore, because Bennett's a faithful Chrisitan who believes in the sacredness of all human souls, thus contradicting utilitarian judgement, we can't admit even his self-consistent premises into evidence - Just give us our sacred PC cows!

How comforting to live with only ones stereotypes ever confirmed. As the psychoanalyst at shrinkwrapped shows, it's not only Islamists who are guided by a self-confirming fantasy ideology. fantasy and delusion As he observes there, "the 'bigotry of soft expectations' is racism by any other name and has a long history among liberal supporters of the dependency culture."

As the inainity of other commenters in this thread show, the denial of PC will be our civic derapage. We no longer care enough about real people and real problems to deal seriously with them. How sad: denial has become all.
10.1.2005 8:15pm
I am a historian, and I always enjoy social scientists and their methodologies. And their pretend little worlds. This one in particular.

"Once you control for income, the likelihood of growing up in a female-headed household, having a teenage mother, and how urban the environment is, the importance of race disappears for all crimes except homicide. . ."

Sad thing is, in /real/ societies like ours, we can't actually "control for" any of these predictors. Rather, we watch them fall disproportionately on the black urban poor. These are the statistically-significant predictors of crime in a given population. Bennett didn't say that race was the determining factor, or even that it was /a/ determining factor. He implied correlation. Isn't this correlation empirically verifiable?

Another "thought experiment," as an illustration. (Answer honestly.) "Race" doesn't cause people to blow other people to smithereenies. But which group would you suspect would produce /more/ terrorists: 100,000 young Middle Eastern men, or 100,000 Korean grandmothers?

Have we lost the ability to distinguish between correlation and causation? Or are we simply denying the correlation?
10.1.2005 9:37pm
this isn't the first time Bill Bennett has talked about race and crime. It wasn't merely an "off the cuff" thoughtless remark.

For a more fully developed version, see his book "Body Count."
10.1.2005 9:40pm
DWPittelli (mail) (www):
Why does Bennett's "inflammatory" choice have to be either "accidental" or nefarious? It seems to me that when making a reductio ad absurdum you are supposed to choose an example that is particularly inflammatory or heinous.

Perhaps for that reason, Wikipedia uses the same "racist" argument as an example of reductio ad absurdum.
"if we wanted less crime we could just abort all minority babies, since adult minorities commit most of the crime. But that would be absurd. Abortion should not be used as a tool of social engineering."
10.1.2005 10:21pm
Tom952 (mail):
the importance of race disappears for all crimes except homicide. . ."

Isn't homicide too serious to sweep out of consideration as though it is some trivial leftover crime category?
10.2.2005 1:20am
Sebastianguy99 (mail):
Based on the comments, one would also be correct to assert...
"if you wanted to reduce racism, you could -- if that were your sole purpose -- you could abort every white baby in this country, and your incidents of racism, as well as the rationalization and intellectualization for racism would go down."
Sounds as logical as what Bennet and the commentors here have poisted!
10.2.2005 1:39am
Steve Sailer (mail) (www):
Levitt is trying to weasel out of getting crucified like Bennett is now that the sizable racial aspect of Levitt's popular abortion-cut-crime theory has been brought to public attention. Levitt expunged the racial eugenics aspect of his theory from his bestseller "Freakonomics," but he had made it clear in his 2001 academic paper with John J. Donohue, where he and Donohue wrote:

"Fertility declines for black women [after the legalization of abortion] are three times greater than for whites (12 percent compared to 4 percent). 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. Under the assumption that those black and white births eliminated by legalized abortion would have experienced the average criminal propensities of their respective races, then the predicted reduction in homicide is 8.9 percent. In other words, taking into account differential abortion rates by race raises the predicted impact of abortion legalization on homicide from 5.4 percent to 8.9 percent." -- Source

The funny thing is that Levitt's whole theory, both the parts based on eugenic logic and the rest, turns out to have been wrong historically, as I first pointed out to him six years ago in our Slate debate that Bennett referred to: the first cohort of black youths born _after_ the legalization abortion in 1970-1973 was not better behaved, as Levitt's theory predicted. Instead, their homicide rate as 14-17 year olds was more than quadruple the homicide rate of the last cohort of black youths born before legalizaton!

You can read all the facts Levitt doesn't want you to know about his theory at
10.2.2005 1:39am
Darleen (mail) (www):
Wow. It seems from some of the "Bennett should have known better" statements, that y'all would have agreed it was a good and proper thing that poor David Howard was forced to resign when he used the word "niggardly" in a memo.

You'll be happy, too, to note that he actually agrees with you..having said in apology Clearly what matters is not my intention but the impact of the words on others.

Yep. All conversation, discourse and debate hinges solely on the perception (regardless of reality) of the listener.
10.2.2005 3:25am
Mr. Sailer, what you've just said is one of the most insanely idiotic things I have ever heard. At no point in your rambling, incoherent response were you even close to anything that could be considered a rational thought. Everyone in on this blog is now dumber for having read it. I award you no points, and may God have mercy on your soul.

10.2.2005 3:37am
Justin (mail):
Does anyone actually find Scrappleface funny? At all? If so, you can join Steve Sailor in the Adam Sandler box.
10.2.2005 3:57am
Jody (mail) (www):
To simplify it for Greg: Sailer is calling Levitt a liar in a professional manner.

Levitt now blogs that his research showed that race had nothing to do with his abortion hypothesis. Sailer points out that in Levitt's actual research paper on abortion and crime, Levitt acknowledges that race has a very big effect.

Ergo, Levitt is misrepresenting his own work on his blog in the interests of remaining on the right side of the PC police.

Sailer then points out a serious flaw in Levitt's abortion theory.

So in two quick bullet points,
1) Sailer says that Levitt's theory (in Levitt's own words) is indeed based on race.
2) Sailer says Levitt's theory is wrong, anyways.
10.2.2005 11:49am
John Thacker (mail):
At no point in your rambling, incoherent response were you even close to anything that could be considered a rational thought.

Actually, he did. You plainly can look up Professor Levitt's 2001 article with Professor Donahue. The original 2001 article did reference race explicitly. All those who claim that Professor Levitt never did so are wrong-- he did at one point argue that the higher abortion rates among blacks reduced crime. Of course, he has removed it, subsequently arguing that race is unimportant after controlling for various variables. That is fair enough, but it doesn't change the fact that a racist could, ignoring the other variables, simply choose to have babies aborted based on race and still, according to Professor Levitt, see a reduction in crime.

That is obviously a horrible, evil policy that we would all reject. And that was Bennett's point-- there are some sorts of arguments that one could make on utilitarian grounds which would be utterly wrong morally. He was arguing it as a reason for fellow abortion opponents to not make a similar utilitarian claim, about Social Security.

Professor Levitt's last point is, I think, clearly wrong. Bill Bennett could easily be speaking in the hypothetical-- he does not have to believe the argument (and indeed, he specifically disavowed it right after saying it) in order to offer it up as the type of bad argument people could make. He is not attempting to have it both ways.
10.2.2005 11:55am
Doc Rampage (mail) (www):
Correcting for poverty and single parenthood doesn't really help Levitt because crime is a cause of poverty, and single parenthood is a cause of crime. You can't factor out causally-linked properties like that and expect to get meaningful results.

And anyway, what difference would it make? By whatever casual chain, blacks-American culture has gotten into the condition where blacks are disproportionately responsible for crime. I'll go Bennet one better: if you went around the country and killed every young black man between the ages of ten and thirty, we would see a huge drop in crime in this country. It doesn't matter why they commit the crimes, the fact is that they do.

That's a pretty shocking thing to say, right? Well that's the point in this variant of a reductio ad absurdum. You look for the most shocking conclusion to be be made from your opponent's argument. The purpose is to prevent your opponent from simply appropriating the point.

If Bennet had said "we could just abort all male babies" then his opponents could have said, "that's not a bad idea. Maybe we should offer incentives to women to abort male babies in high-risk social groups". Bennet had to pick something that his opponent would never agree to, even in principle. That's why he picked such a shocking example.

This is where you end up when you start looking for solutions to the problem of crime by killing people on the grounds that they are statistically likely to be criminals. This is what Levitt's logic will lead to. Just kill all black men. And that is why Levitt's supporters are so determined to make this argument just about what a racist Bill Bennet is. They don't want people to understand this.
Incidentally, the "absurdum" in "reductio ad absurdum" doesn't meant "absurd" in the modern sense, it means "aburd" in the original sense: logiclally contradictory. That's why I (rather pedantically) refer to this as a variant of reductio.
10.2.2005 3:18pm
Cheburashka (mail):
I thought Sailer's point was cogent and interesting.
10.2.2005 6:08pm
Richard Gould-Saltman (mail):
Hmm, "exaggeration for effect" "Swiftian satirical intent", "reducto ad absurdum" even "though experiment"... except that Bennett didn't say any of that, including the invocation of Einstein, until he attempted to clean up, after the fact. Harry Shearer's got the sound-bite, and Bennett sounds like the smarmy, Limbaugh-with-a-degree that he occasionally does when he gets cranky, and doen't have someone editing him.

10.3.2005 12:50pm
I think a point many of us have neglected is that there is no adequate way to measure who is committing more crime. Many statistics are based on how many persons are convicted or charged; however, if prosecutors tend to focus on a specific group, this group will have higher crime stats. For instance, if there was policy shift to spend less time prosecuting crack in the city and more time prosecuting cocaine in surburban neighborhoods, stats would change and there would be "higher white crime rate," whatever that means. Everyone here seems convinced that black males presumably commit more crimes. Well, I pose a different question. If we killed all black men, would there be a reduction of crime, or would more white criminals get prosecuted? Let us not become the victims of "pop science" and "pop statistics" that are clearly flawed and skewed. Please, tell me one valid source to prove this. I have searched and have seen none. There is no way of quantifying uncharged crimes.
10.30.2005 6:56pm