(Cross-posted to property professors' blog, see Barros blog in list of blogs to left)
In my experience, most people who discuss the relationship between urban decay and crime treat the relationship as a one-way street: city crime causes people to leave cities, period.
But in Nicole Garnett's review of Bruegmann's and Joel Kotkin's new books (posted on SSRN, and referred to in a post on this blog a few days ago) she suggests that lower urban densities might induce crime by making cities less "busy" and more deserted- and thus that (to oversimplify the point into a sound bite) that sprawl might even cause crime in a sense.
I'm not sure there's any way to prove or disprove the theory- but if the argument is verifiable, it certainly leads to some interesting results.
Let's go back to the 1930s, when the FHA started to bribe people to move to suburbs with mortgage subsidies and all levels of government were beginning to make suburban commutes easier through road-building. A few people leave the (now safe) cities. Over the next few decades, a few city neighborhoods here and there become less busy and thus more dangerous, and the most risk-averse people start to trickle out. This causes neighborhoods to become even less busy and more dangerous which cause even more people start to trickle out, and eventually we have a vicious circle on our hands- a vicious circle that spirals out of control in the 60s (when for reasons unrelated to urban policy, crime increases everywhere in the United States).
And depopulation causes other problems that independently might increase crime. A city without a large middle and upper class might support more lenient policing policies which in turn might lead to more crime - another respect in which sprawl (or more accurately, the type of sprawl that depopulates cities, as opposed to sprawl in growing regions where there is enough population growth to build up city and suburb alike) might increase urban crime.
Two caveats:
1. All of this is pretty speculative.
2. I think it is easy (but mistaken) to assume that crime is a problem that can be resolved solely through more enlightened city government. Even if you assume for the sake of argument that the criminal justice system has a major effect on crime (as opposed to, say, liberal morality, economic inequality, or family breakdown), criminal justice is more of a federal and state responsibility than a local responsibility. Cities may hire police, but states decide whether to build enough prisons to house the people arrested by city police, and both federal and state courts set the rules that decide how crowded those prisons can be and how easy it is to convict people arrested by the city police.
