The Study That Could Upend Everything We Thought We Knew About Declining Urban Crime
Bill Bratton took the job as commissioner of the New York Police Department in 1994 under Mayor Rudy Giuliani, setting the stage for a Cinderella story in urban law enforcement that went on to change how virtually every major U.S. city tackles crime. "He made a vow that he was going to bring down New York crime, he specified a certain amount," says New York University sociologist David Greenberg (that amount: a 40 percent reduction in crime within three years). "And he delivered on that and took credit for the decline."
The story of exactly how that happened has since proliferated, in Bratton’s own accounting, in think tanks celebrating his tactics, in op-eds cementing a popular consensus. During the 1990s, New York City began to deploy a suite of new policing tactics, and – clearly – they worked! The city cracked down on misdemeanors thought to contribute to neglected environments in which felonies then occur (this was the "broken windows" theory of policing popularized, in large part, by The Atlantic). New York rolled out the now-vaunted CompStat computer system designed to better track crime. And stop-and-frisks (a less savory strategy) increased.
"Part of the history of New York City’s revival as a city is that people are safe," says Greenberg, who has studied the city’s crime data from this period. "Our findings suggest they are much safer, but not because of what the New York Police Department is doing."
That conclusion, based on a paper recently published in the journal Justice Quarterly, is a pretty jarring one. It challenges widely held narratives of how New York won its war on crime. But it also raises awkward questions about the efficacy of certain police tactics everywhere, particularly "broken windows." ("It’s a curiosity," Greenberg adds, "that this name got attached to what the New York Police Department was doing, because the police never went after broken windows").
In the canon of criminology research about New York’s historic crime decline, some data have supported the NYPD's claims, while other studies have been less conclusive. "I found as I reviewed these different studies," Greenberg says, "that all of them – particularly the ones that claimed the policy was working – had statistical flaws." Past research on the topic, he adds, has generally been conducted with incomplete or "limited" police data.
Greenberg has analyzed instead a new and richer database of precinct-level data from New York, collected from 1988 through 2001 by Missouri researcher Richard Rosenfeld. This data contains deeper information about the demographic composition and socioeconomic status of neighborhoods where crime occurred, the presence and practices of local police officers there, and the prison admission rates of people who were arrested.
During this 14-year span, homicide rates dropped in every one of the city’s 75 precincts, and cases of assault and robbery dropped in nearly all as well. But these long-term trends predated the implementation of CompStat and appeared to be unaffected by its arrival. In this graph of violent crime trends from 1988-2001, from Greenberg's paper, the vertical red line down the center marks the introduction of CompStat:
"Studying New York City’s Crime Decline: Methodological Issues," D. Greenerg in Justice Quarterly
Greenberg also found no causal connection between officers per capita at the precinct level and reductions in violent crime, or between an increase in misdemeanor arrests and a drop in felonies (as "broken windows" implies might happen). Other research [PDF] conducted by Rosenfeld and Robert Fornango found that the policy of stop-and-frisks had scant impact as well. There's also the sticky issue that crime fell dramatically during this period in cities everywhere. New York wasn’t alone (although its story was admittedly dramatic).
Some critics of the police department have gone so far as to suggest that the city’s dramatic drop in crime never really happened, that it was the result of accounting tricks created by reclassifying homicides as suicides, or by downgrading felonies to misdemeanors. The overall crime trend, however, has been validated by researchers using outside sources such as medical examiner records and auto insurance claims. (A rise in homicides reclassified as suicides would also imply that suicides in this era would have increased, and that did not happen.)
"So as best as we can check these results, they seem to be believable," Greenberg says. "That’s not to say they’re perfect. Crime statistics are never perfect."
Which brings us back to the central question: Why did the crime rate tumble? If it wasn’t thanks to touted police tactics, what then?
In a particularly zinging passage from his latest paper, Greenberg puts it this way:
For some other, unstudied, police activity to have been responsible for a large part of the drop, one has to assume that the police inadvertently stumbled on highly effective crime prevention strategies and pursued them for 20 years without knowing what they were, or without ever publicizing them. One would also have to assume that they escaped the notice of defenders and critics of the police over all these years. These possibilities are not plausible.
The more plausible assumption, he argues, is that other social or cultural changes were at play in New York (and the many other cities where crime declined). University of Chicago economist Steven Levitt – of Freakonomics claim – has famously suggested that a post-Roe v. Wade rise in abortions helped lead to a drop in crime. Others have proposed the “lead hypothesis,” that crime has fallen along with exposure to the potentially toxic material. Greenberg allows that both explanations could play a part, but he has no definitive answer to fill the void his research has created debunking police strategy.
His study does also raise one other big-picture question. If creative policing played no discernible role in one of the most dramatic crime reductions in recent history, does that mean police tactics don’t much matter anywhere?
"No one would claim that crime levels would be unchanged if the police were taken off the streets totally," he says (and, being a researcher he cites some hard evidence: during the 1919 police strike in Boston, and the 1977 blackout in New York, looters took advantage of otherwise preoccupied cops). "But it does mean that at the margin, the things they are doing may not be productive, and maybe in the long run they’re counterproductive."
The policy of stop-and-frisks, for one, may alienate residents from the police, and law enforcement needs good relationships with communities to effectively stop crime (as, well, a whole different set of research suggests).
Top image: Shutterstock.com/