The CQ Press has just released its 14th Edition of City Crime Rankings to a loud chorus of disapproval from police chiefs, mayors and others, especially in those cities that top the list of the most dangerous. The 5 most dangerous cities were, in order: Detroit, St. Louis, Oakland, Flint, and Camden. The 5 safest cities were Mission Viejo, California; Clarkston, New York; Brick Township, New Jersey; Amherst, New York; and Sugar Land, Texas.
Those lists alone are interesting: Michigan has two cities in the top 5 most dangerous; New York has two cities in the top 5 safest; California and New Jersey each have one city in the top 5 of each. And most obviously, the top 5 safest "cities" aren't really cities at all. They are political units with at least 75,000 persons living in them. These facts tend to confirm what some critics of the project note: That the variation in crime rates within a single city is often larger than the variation from city to city. A list of the most dangerous and safest STATES would be misleading with respect to California, New Jersey and other states, and so the fact that a large city is dangerous (or safe) overall can be very misleading about particular neighborhoods.
In addition, the fact that CQ Press only makes its list and methodology available for a fee suggests that this is a profit-driven gimmick, not unlike, say, the US News & World Report rankings of educational institutions. And like those rankings, this sort of effort has something of the self-fulfilling prophesy about it. Some number of people considering relocating to Detroit or St. Louis will choose not to, which will hurt property values further, which will further shrink the tax base, which will require further sacrifices in policing and social programs that fight crime, and so on in a vicious cycle. (At the safe end of the spectrum, virtuous cycles are less likely. Given the small populations, a single criminal coming to town could ruin the rating of a very safe "city.")
Still, for all the flaws of the aggregation methods, there should be little doubt that accurate measures of crime are an essential piece of a successful crime-fighting strategy. New York City (which has experienced a 75% drop in its murder rate since 1993), publishes its crime statistics, broken down by police precinct, on a weekly basis (here). This kind of fine-grained information has two principal effects: First, it enables the police to detect crime trends early and to respond; and second, it enables citizens concerned about crime to figure out whether they are simply reacting to media hype and isolated incidents, on the one hand, or to real trends on the other, and if the latter, to apply political pressure for responses.
I'm NOT saying that CompState (the NYPD's program of tracking crimes by neighborhood) is solely or even principally responsible for the dramatic decline in crime in NYC over the last decade and a half. Criminologists continue to debate the relative contributions of various factors (including, among other things, CompStat, legal abortion, the waning of the crack epidemic, the "broken windows" policing strategy, and others). What I am saying is that leaders in cities with high crime rates should not simply point to flaws in the CQ Press rankings. They should track crime by their own, transparent, and more informative methods. Perhaps Brian Leiter could help out if the NYPD isn't willing to.
Posted by Mike Dorf