Does A Broken Windows Policy Work?

Denver is currently working on implementing a "broken windows" policy of policing with one of its main proponents, George Kelling, brought on board by Mayor Hickenlooper as a consultant to make it happen. The idea behind the proposal is that aggressive enforcement of even minor violations reduces a sense of disorder that can cause crime.

Support for such policies is not universal. A few days ago, Em Rosa pointed out some of the main problems with it (such as the fact that it comes down hard on roudy by lawful activity and leads to stereotyping), and notes that San Francisco got similar or better results with an opposite policy a less aggressive policing of quality of life offenses (itself potentially a product of San Francisco's increasing gentrification, rather than policing policies).

Today, the Denver Post notes another opponent of the policy who notes that:

Bernard Harcourt, a University of Chicago law professor, and Jens Ludwig, associate professor of public policy at Georgetown University, in their upcoming article argue: "From a public policy perspective, the faith that many policymakers place in the efficacy of broken windows is in the end just faith, rather than the result of convincing empirical evidence."

They said that a 2001 study Kelling co-wrote on broken windows is flawed. Kelling, who helped implement the broken-windows theory in New York during the 1990s, said in the study that a crackdown on misdemeanor crimes in New York reduced felony crimes.

Harcourt said the real reason for the decline was that the city's crack-cocaine epidemic, which ravaged New York's neighborhoods, had subsided.

He also reviewed a federal initiative that moved about 4,800 low-income families in high-crime public-housing areas to less disadvantaged and less disorderly communities.

The review found that those who moved to better neighborhoods failed to have a lower arrest rate than those who stayed behind.

I honestly don't know who's right in this debate. Conclusions about the causes of crime are always difficult to prove because there is no such thing as a perfectly controlled experiment in criminology. I can see the intuition behind both approaches and suggestive, but not conclusive, evidence in support of each. The null hypothesis that policing policies have little impact on crime for better or for worse is also a strong contender in my opinion.

I suspect that any coherent policy for addressing crime in Denver, whatever its focus, is likely to have a positive effect, simply because it requires a fresh look at a stagnant set of policing policies and will correct something that is wrong, and because placing a spotlight on the police managers involved gives them a strong incentive to do a good job. To some extent, the particular tenents of the policy to be implemented doesn't matter all that much, in terms of the results that they are likely to produce.

For now, I will sit back, see what happens and hope for the best, as I have no other real choice in the matter, and don't have a magic bullet solution to propose myself.