On Reason online, Randall O' Toole bashes grid streets, asserting that cul-de-sacs and separation of land uses control crime.
O'Toole asserts that the safest neighborhoods are dominated by cul-de-sacs that are as isolated as possible from stores and from other streets, implying that any neighborhood in which residents can walk anywhere at all is "custom made for easy crime" because of the possible influx of strangers. In other words, the way to stay safe is to imprison yourself.
But the experiment of imprisoning residents in order to protect them has been tried in the American Sun Belt, with dismal results. In my former home town of Atlanta, the traditional street grid disappears about two or three miles from downtown, to be replaced by a maze of cul-de-sacs. Yet in 2002, Atlanta had 1964 burglaries per 100,000 people - more than five times as many as New York City, and more than twice as many as San Francisco. (I think burglaries are the most relevant crime to this argument because that crime is most likely to occur inside a house, and thus most likely to be connected to street design and land use - and also because O'Toole has, in another article, used the term burglar-friendly to describe pedestrian-friendly neighborhoods).
The rest of metropolitan Atlanta has tried similar techniques with equally dismal results. According to a Smart Growth America survey, Atlanta has the third lowest "street connectivity" score in America - that is, its streets do not connect with each other, just as O'Toole would like.
Yet the region boasted 924 burglaries per 100,000 people in 2003 - more than twice as many as in metro New York City (356 burglaries/100,000), the region with the highest level of street connectivity for which 2003 regional crime statistics are available. Rochester and Syracuse, the two regions with the lowest level of street connectivity, also have burglary rates higher than metropolitan New York City.
Segregating housing from commerce is no panacea either. The three regions with the lowest mix use levels are Raleigh (854 burglaries/100,000), Greensboro (1198/100,000) and Riverside (853/100,000)- all more burglary-prone than the national average of 757 burglaries per 100,000 people. By contrast, the three most mixed-use areas for which crime statistics were available (Providence, Allentown and Oxnard) all had fewer than 600 burglaries per 100,000 people.
To be sure, regional crime statistics are of limited value, because regions include a wide variety of neighborhoods.
But plenty of small suburbs have a gridded, mixed-use core and extremely low crime rates. For example, East Aurora, New York is dominated by a gridded 19th-century downtown - yet it had fewer than 150 burglaries per 100,000 people (20 in a town with just under 14,000 residents). Ditto for Haddonfield, New Jersey, which sits on a subway line running from Philadelphia and crime-infested Camden, and had just over 11,000 residents and only 10 burglaries.
Concededly, these neighborhoods are high-income enclaves. But how do they stay that way? If East Aurora-style urbanism was so inherently undesirable, high-income people would stop living there.
More importantly, if high-income suburbs have low crime rates no matter what their geographic form, it logically follows that the main reason one town is more dangerous than another isn't urban form: its demography. Places with lots of poor people have a lot more crime, whether they look like East Aurora or like a typical Atlanta suburb. Places that don't have a lot of poor people don't have a lot of crime.
Posted by lewyn
at 10:57 AM EST
Updated: Sunday, 13 February 2005 10:57 AM EST