Sorry, Malcolm Gladwell: NYC’s Drop in Crime Not Due to Broken Window Theory

We have no idea why crime dropped, but it had nothing to do with broken windows or police strategy

The “broken window theory” has had its day. This criminological theory, which argues that keeping urban environments neat and tidy deters would-be criminals, first popped up in social science in 1969, with a famous experiment detailing the fates of two different cars left out on the street in the Bronx and in Palo Alto with their hoods open and license plates removed. (Spoiler: the car in Palo Alto fared better—until the researcher broke its window, after which it was quickly stripped down.) The theory gained popularity through the ’80s, when The Atlantic first covered it, and ’90s, when New York City used it to design policing strategy, before, in 2000, it helped journalist Malcolm Gladwell make his career with The Tipping Point. The book earned the author whopping $1 million advance, and introduced to the theory to a much wider audience—many readers remember  most vividly the broken window section of Gladwell’s best-seller.

Turns out, however, that the broken window theory doesn’t really apply that well to reality. New research shows that New York City’s historic decline in crime rates during the 1990s cannot be attributed to CompState, the NYC police department’s dynamic approach to crime, introduced in 1994, that included carrying out operations in accordance with the broken window theory. The crime decline has nothing to do with enhanced enforcement of misdemeanors, the research published in Justice Quarterly by New York University professor David Greenberg reports, nor is there any link between arrests in misdemeanors and drops in felony charges, including robberies, homicides and assaults.

“While the 1990s drop in felonies is undeniable, what remains unsolved is the cause, or causes, behind this significant change in New York City’s crime rates,” Greenberg said in a statement.

In addition, neither the number of police officers per capita nor the rate of prison sentences doled out to criminals turned out to be related to a reduction in violent crime. To arrive at these findings, Greenberg examined crime data across NYC’s 75 precincts from 1988 to 2001. During this time, crime rates fell nearly uniformly across the city. (Incidentally, Los Angeles, San Diego and other major cities underwent a similar shift in crime during this time.)

His analysis found no relationship between the decline in violence crime and CompStat or any of the other actions inspired by broken window theory. While violent crime decreased over the 13 year period, Greenberg found, misdemeanors increased in all but 11 precincts, invalidating the theory’s basis. Felonies, on the other hand, dropped consistently across the city, independently of a shrinking police force, imprisonment rates or enforcement levels, which varied greatly around the five boroughs.

“While many may point to greater enforcement of lower-level offenses as a factor in curbing more serious crimes, the data simply don’t support this conclusion,” Greenberg said. He added that he has no idea why the crime dropped—only that it had nothing to do with broken windows.

More from

Can Computers Predict Crimes? 
Check Out the Milwaukee Police’s Mind-Blowing, Crime-Busting New Site 

Get the latest stories in your inbox every weekday.