There’s some discussion over at Catallaxy of the new book by Steven Levitt and Stephen Dubner, Freakonomics: A Rogue Economist Explains the Hidden Side of Everything. I don’t know if Levitt could actually be described as a rogue economist, having picked up awards from the profession. One of his more interesting theses has been highly controversial, and will no doubt be more so with the book heading up the bestseller lists:
A few years ago, a young economist named Steven D. Levitt became briefly notorious for collaborating on a research paper that contained a strikingly novel thesis: abortion curbs crime. What Levitt and his co-author claimed, specifically, was that the sharp drop in the United States crime rate during the 1990’s — commonly attributed to factors like better policing, stiffer gun laws and an aging population — was in fact largely due to the Roe v. Wade decision two decades earlier. The logic was simple: unwanted children are more likely to grow up to become criminals; legalized abortion leads to less unwantedness; therefore, abortion leads to less crime.
This conclusion managed to offend nearly everyone. Conservatives were outraged that abortion was seemingly being promoted as a solution to crime. Liberals detected a whiff of racist eugenics. Besides, what business did this callow economist have trespassing on the territory of the criminologist? Economics is supposed to be about price elasticities and interest rates and diminishing marginal utilities, not abortion and crime. That is what makes it so useful to undergraduates seeking relief from insomnia.
Levitt has strayed far from the customary paddock of the dismal science in search of interesting problems. How do parents of different races and classes choose names for their children? What sort of contestants on the TV show ”The Weakest Link” are most likely to be discriminated against by their fellow contestants? If crack dealers make so much money, why do they live with their moms? Such everyday riddles are fair game for the economist, Levitt contends, because their solution involves understanding how people react to incentives. His peers seem to agree. In 2003, Levitt was awarded the John Bates Clark Medal, bestowed every two years on the most accomplished American economist under 40.
Interestingly, Levitt has a blog attached to the book’s website, where he posts on his abortion research.
As a sociologist, I’m not professionally competent to evaluate Levitt’s methodology, so I’m grateful to Jason Soon for this link to Steve Sailor’s criticisms. It did strike me, particularly after reading a post at John Quiggin’s place about physicists re-inventing the sociological wheel by doing social network analysis, that this is an example of an economist dealing with problems that would normally be of interest to sociologists. This is of course more usual in the application of rational choice methods (which originated in economics) to the other social sciences, and Levitt’s methodology appears to be on a quick inspection more characteristic of social science research more broadly than economic research per se.
What Levitt is probably picking up on is the consensus among American criminologists that the fall in the rate of violent crime over the last 15 years (though it’s important to remember that it’s still historically high) is not explained by things like three strikes laws or mandatory sentencing or zero-tolerance policing. The most powerful explanation appears to be demographic - the generation of young men aged between 18 and 24 (the time in the lifecourse where people are most prone to commit crimes against the person) is simply smaller than it was in the 80s when US crime rates peaked. This conclusion, of course, is reached using regression analysis, and as with all causal correlations from statistical research, only tells part of the story. There are underlying social and cultural shifts going on as well which are better brought out by social theory and qualitative work. However, there is no doubt that this demographic factor has a lot of explanatory power.
Levitt’s analysis seems to me to be parasitic on this research, applying it to cross-state comparisons (an attempt to reproduce the methodology of the experiment in the natural sciences). However, this is a very problematic method and I doubt that there’s a straight causal direction from the legalisation of abortion to falling crime rates. I’m grateful to Nicholas Gruen for advising me via email there’s some doubt among economists about the validity of this method as well. Nevertheless, it’s a very interesting correlation, to say the least.
Rafe Champion seems to suggest at Catallaxy that there’s some sort of general truth demonstrated from Levitt’s research that “incentives and motivations” can explain human behaviour. No doubt this is true to some degree, but I’d argue that the lack of sociological context in the application of economic research to sociological problems casts doubt on such an ambitious claim.
The best thing about abortion is that it reduces the number of Democrat voters.
I suppose there’s something to be said for this type of eugenics after all.
I’m only guessing here but maybe the theory is that abortion curbs the number of people that didn’t want children that would have been forced to bring them into this world and didn’t want to give them a complex by finding out they were adopted.
Thus abortions curb the decay of social moral fabric which is the catalyst for some crimes thus it lowers crime levels.
I’m still guessing of course, and maybe the economic factor cuts in at the crime level and you work backward for there.
As I said, just a guess but how did I do?
Pretty close, Vee.
Here’s the basic statement of his theory:
The reality is that there is about as much empirical evidence that legalizing abortion drove _up_ the serious violent crime and murder rates in America as that it drove them down. It appears to be beyond the current capability of social science to determine a reliable answer to the question of what was the impact of legalizing abortion on violence.
Levitt covers up in “Freakonomics” what actually happened: the first generation born after the legalization of abortion went on the worst teenage murder and serious violent crime spree in American history. Indeed, the teen crime rate went up in the late 1980s first in those big cities — NYC, LA, and DC — that were the first to legalize abortion, de jure or de facto.
Levitt’s simplistic “unwantedness” model of how abortion might have cut crime is badly undermined by his own admission that the birthrate only dropped 6% but the pregnancy rate went up by almost 30%. That legalizing aboriton greatly increased the unwanted pregnancy rate makes his “unwantedness” model obviously inadequate. Levitt knows that but he keeps pushing it because nobody else thinks about the impact of legalization on unwanted pregnancies.
You can read the whole story on the abortion-cut-crime-theory controversy at http://www.iSteve.com/abortion.htm
Thanks very much, Steve, that’s most informative.
For a more thorough discussion of Sailer’s criticisms of Levitt, see http://crookedtimber.org/2005/05/23/response/.
In the Australian context, Justin Wolfers and I found some evidence consistent with Levitt’s theory: http://econrsss.anu.edu.au/~aleigh/opinion_fulltext.htm#Abortion