Abortion, "Broken Windows", Or Epidemiology

| 1 Comment

There's a bit of a back-and-forth going on between Malcom Gladwell and Steves Levitt/Dubner on the causes behind the drop in crime rates during the 1990s.

Gladwell serves. Dubner on the return. Gladwell's rally.

Read it all; it's time well spent. In sum, though, it can be said that both sides seem to be sticking to their guns. Gladwell puts a lot of emphasis on the "broken windows" notion (not the broken windows fallacy), while Dubner reaffirms belief that legalized abortion and the fall-off in violent crack trade played large roles.

To add to the debate (though unnoticed by the debaters, but hey, whatcanyado?), I'd offer up another perspective, this time from Paul Ormerod and Michael Cambell (PDF). The authors focus on the decline in crime rates by using methods from mathematical biology, and considering crime as something akin to an epidemic. (Note: the paper is from 1996, though this is newer than the original Broken Windows article.) The point of the article isn't to point to one or two causal elements, but rather propose a new model for examining the phenomenon:

We have proposed in this paper an approach towards understanding the dynamics of crime which is similar to that used in mathematical biology to model the spread or containment of epidemics. A population, however defined, can be split conceptually at any point in time into three groups, those who are not interested at all in committing crime, those who are susceptible to become criminals, and those who are criminals. A system of differential equations is set up to model the flows between these groups over time.

The strengths of the flows can be interpreted as corresponding to the main factors identified in the empirical literature as the causes of crime, such as demographic movements, general social and economic conditions, and the positive and negative deterrence effects of the criminal justice system. Despite a voluminous literature, no firm quantitative conclusions have emerged, and the model can be used to explore the consequences for crime of variations in the respective strengths of such factors. In principle, the model could be calibrated on a data set of particular crime rates.

Of course, this leaves the agent-based method open to problems stemming from the data set chosen, but the same can be said of the Dubner/Levitt and Gladwell arguments. Given that neither of those explanations is wholly compelling to me, I'd be interested to see a rigorous implementation of this third method. Here's one paper (PDF) from the Society for Computational Economics that gets close, though it models the effects of shocks in mortality rates (the drop in violent crack trade?) on crime rates. I'm sure there are plenty of others that I simply haven't seen.

1 Comment

I thought both Gladwell (in Tipping Point) and Ormerod (in Death of Economics) makes similar sort of arguments;'how little things can make a big difference'. The popularity of Tipping Point over the book 'Death of Economics' may be one example. May be economists should study under epidemologists.-Paul


Powered by Movable Type 5.02

About this Entry

This page contains a single entry by published on March 13, 2006 5:03 PM.

What Does Gold Farming Look Like? was the previous entry in this blog.

Good Economic News for Bin Laden is the next entry in this blog.

Find recent content on the main index or look in the archives to find all content.