« Damaging the supply-side | Main | Ruth Kelly: a defence »

January 06, 2007



Good post. Now if Nick Hill and the others could just bring the Vatican and evangelical Christians on board...


isn't there a similar theory to the effect that abortions reduced US crime ?


"Donohue and Levitt say that the increase of abortions in the 1970s corresponds to the decline in crime seen in the 1990s, just about the time those babies would have turned 17 and entered an age group characterized by especially high crime rates. Even more significantly, Donohue and Levitt argue, that drop in crime began to show up first in the five states that legalized abortion prior to Roe v. Wade. Further, they contend that crime dropped most dramatically in those states with the highest abortion rates."


Maybe I should read the post more closely before commenting ...

I don't think the thesis would necessarily explain the drop in UK crime rates in Victorian times, though (It might if the London Rubber Company was doing lots of technical innovation and price reductions - which they may well have been). It's probably one variable among many.

Dave Petterson

Now all we have to do is change the welfare state so those who get pregnant to make a living decide it's better to remain child free.


I would say also that increased use of contraception by women is associated with better education, as far as I know. And since lack of education is associated with poverty and poverty is associated with higher crime rates...

I'm just pulling that out of nowhere on a "common sense" sort of basis, I might be talking utter shite.

Chris Williams

I'm not so sure that crime (as opposed to public disorder) really fell in the Victorian period. Chapter and verse OR an argument from authority are available on request to back that statement up.


[Hill's skill is that he's found the empirical evidence]

you are kidding. That is one of the worst microeconometric papers I have ever seen. I am amazed that anyone had the balls to present it. The "Contraception" parameter is barely significant most of the way through, and half of the models presented are badly misspecified (in fairness, the paper does admit this, but why waste our time on them then?). There is no empirical evidence of anything at all in that paper; it doesn't even have standard demographic controls so we don't know if there is any effect of "wantedness" over and above the known effect of the baby slump of that period. It's dire. In general, Chris, there is a worrying tendency on your blog to take econometrics working papers at face value - even journal articles are often nothing like as definitive as they claim to be, and a lot of really awful stuff makes it to the working paper stage.

Mark Wadsworth

We are all hopefully familiar with the concept from Freakonomics re legalisation of abortion and corresponding fall in crime 15-20 years later; the ban on abortion in Romania in the seventies that led to Ceausescu's downfall 15-20 years later?

There was a cracking column in The FT of 6/7 Jan reviewing a German study* that says that there is a clear correlation between fertility and lawlessness, in extremis, Palestian mothers have 9 children - since the Israelis got out of Gaza, these people have just turned on each other. Countries with low birth rates, in extremis Germany and Japan are in comparison the most peaceful nations on earth.

*"Söhne und Weltmacht: Terror im Aufstieg und Fall der Nationen", by Gunnar Heinsohn, publ. Orell Füssli Verlag.


Mark, that "connection" is a theory, not a fact. The empirical evidence in favour of it is decidedly less conclusive that Levitt claims in Freakonomics and there are quite a few reputable people in the profession who are really quite angry at the way he's overstated his case.

Mark Wadsworth

I didn't say "connection" I said "correlation". Correlation are facts, but they are not of course in themselves explanations (e.g. widely accepted that left handed people have shorter life expectancy) and do not say what is cause and what is effect.

You are free to argue that people somehow anticipate lawlessness and civil war to come in 15 - 20 years time and, as a defensive mechanism, decide to have more babies NOW to ensure that at least some of their children survive. Doesn't sound very plausible to me, but hey...

john b

Dan - perhaps someone should write some kind of blog article about Levitt's claims, possibly based around a review of his book Freakanomics?

Mark Wadsworth

You can look up population pyramids for most countries on this most helpful website and do your own correlation between "youth bulge" and violence.



[You are free to argue that people somehow anticipate lawlessness and civil war to come in 15 - 20 years time and, as a defensive mechanism, decide to have more babies NOW to ensure that at least some of their children survive. Doesn't sound very plausible to me, but hey...]

well, since I have the alternative line of argument of suggesting that what you have here is a theory of a very small number of data points which doesn't prove anything, and that your strawman suggestion wouldn't actually make my point, I think I won't bother, thanks.

John: check out my blog ...

Mark Wadsworth

Dsquared, I am not taking about a couple of data points, look up that link and you have 200 data points. I am not sure if there is an official "violence" rating for all these countries, but no doubt somebody knows of one somewhere...

Chris Williams

Part of my job involves checking out 'official violence ratings' for various countries. They are worthless in the extreme.

Mark Wadsworth

Chris, that is a shame, but what would you say is the least unreliable source? Then I can plot a graph, the X axis being "youth bluge" and the Y axis being "relative violence rating".

Chris Williams

The Y axis would be 'extent to which this society's governmental forms generate statistics of violence.' Waste of time. There are a few international victim surveys which _might_ get the rates in the right ball-park, but these haven't been running for long enough to generate any reliable longitudinal data.

Steve Sailer

Levitt's abortion-cut-crime theory has severe problems such as that the first cohort born after legalization had a teen murder rate triple that of the last cohort born before legalization, and that his original paper was based on two technical errors he had made.

Similarly, Levitt used only European data to support his theory that women who have abortions would make worse mothers because Americans studies had reached the opposite conclusions -- all else being equal, abortion was used by the better organized, more effective women, while disorganized, present-oriented women didn't bother having abortions.

But Levitt told people what they wanted to hear, so, as they say, a lie goes halfway around the world before the truth gets its boots on.

For details on problems with the theory, see:


The comments to this entry are closed.

blogs I like

Blog powered by Typepad