2/14/2007

Freakonomics

One of the most interesting books I read recently was Freakonomics. The introduction describes how common predictions for the 1990s foretold crime rates spiraling out of control. When these predictions failed to materialize, and, in fact, crime rates dropped drastically nation-wide, politicians cited their own public policies as the reason. The authors contend that it was, in fact, the legalization of abortion twenty years earlier that was the cause of this drop in crime. Leaving ethical debates about abortion out, it certainly makes sense that legalized abortion contributed to the drop in crime (see the student study guide at the previously mentioned website for details). However, the authors themselves assert that correlation is not causation. In economics, the “all other things being equal” clause never happens, so, in effect, many variables impact everything. What other variables may have impacted this drop in the crime rate? What about a healthy economy, increased access to higher education, or outsourcing many 'blue-collar' jobs overseas? The reality is that all these things likely contributed to the drop in crime.

6 comments:

joseph said...

Sounds like an interesting book, since the arguments presented by the politicians seem rather convincing. To be able to link the legalization of abortion to the drop in crime rates, 'all other things being equal', seems to be a difficult job for me. Just plugging in the numbers and doing a statistical analysis is not enough. How does the author of the book present his case? And how does 'outsourcing many 'blue-collar' jobs overseas' relate to drops in crime rates?

TR said...

I was unable to view the student study guide to see the information on the link between legalized abortion and the drop in the crime rate. I do agree that in the real world there is never a time when “everything else is equal.” However, I do think that some of the politician’s policies might have had an effect on the low crime rate. I also agree that a good economy helped out. If unemployment is way low than there won’t be as many people who are desperate for money. If there are less people desperate for money there will be fewer crimes because people’s lives are being sustained. There are many other events that could contribute to lower crime rates, but I doubt we will ever know the exact impact of each one.

Jacob said...

That is quite a claim, to link a drop in crime to the legalization of abortion. I don't know if the legalization of abortion was a factor in the drop in crime. It is hard for me to swallow that idea. I think it is easy to make a mistake when trying to find a correlation between events in history and the circumstances that could have caused them. For example, the summer months are typically hotter than winter months. In addition, people tend to by more ice cream in the summer than in the winter. If a study found that more murders occur in the summer months than in the winter months, is it safe to assume that more murders happen in the summer months than in the winter months because people buy more ice cream in the summer? Does that make sense? My point is that we should be careful when promoting coorelations between events.

Alex said...

Read a critique of the book here.

http://alexmthomas.wordpress.com/2006/08/21/freakonomics-in-the-morgue/

Jada said...

There are a few education studies that have been published proving that a $1 spent on early childhood development will save thousands on adult incarceration. I think the increase in government programs such as head start and furthering educational opportunities are more of a 'direct link or cause' for the declining crime rate, rather than the abortion theory. I also believe that Affirmative Action had its time and place as well by giving minorties a chance to prove they would rather earn a living than eke out an existence. It would be interesting to compare crime rates to education levels and to determine if they are substitutes for each other (which is my belief).

Dr. Tufte said...

-1 on Madeline and Jacob for no title, and poor spelling.

There is still considerable debate on Levitt's conclusion that legalized abortion led to lower crime rates later on. However, that claim has not been generally dismissed or rejected after quite a lot of examination.

An important corollary to that result was Levitt showing that the other explanations don't hold (much) water. That's still standing firmly.

Jada: the evidence of which you speak is not strong (but it is there).

I think Joseph, TR, and Jacob ought to step back and take a deep breath. It's fine to be suspicious, but I don't think you recognize the closed-mindedness of your comments. Consider:

1) "Just plugging in the numbers ... is not enough". Then what is? I hasten to point out that if anyone plugged in the numbers and found out that abortion was not linked to crime, then we probably wouldn't be suspicious at all.

2) "... I doubt we will ever know the exact impact...". This is often used as a justification for not investigating further. The point is to measure something better, not perfectly. Levitt did improve the quality of estimates. To point out that this isn't perfect is not an open-minded strategy.

3) "... it is easy to make a mistake when trying to find a correlation ..." No, it isn't. It is easy to draw an incorrect conclusion from a correlation. One important feature of Levitt's results is that he has a prima facie correlation: one in which one event precedes another. It is easy to say that correlation is not causality, but the problem is that sometimes it is. A step in the right direction to help make the claim of causality is to establish prima facie causality. Levitt did this.