CompStat vs. community policing

Pub date June 22, 2010

By Alex Emslie

Two competing visions for the San Francisco Police Department are central to a looming debate involving the mayor and his police chief, who favor the high-tech yet impersonal CompStat model, and progressive members of the Board of Supervisors who are pushing for a community-based, cops-walking-beats blueprint for SFPD.

District 5 Sup. Ross Mirkarimi introduced a proposed ballot measure on June 7 that would require the police chief to institute foot patrols in all districts and ask the Police Commission to establish a written community policing policy. SFPD Chief George Gascón opposes the initiative, instead favoring a reliance on the new CompStat system to determine how best to use police resources.

The terms “CompStat” and “community policing” have become trendy buzz words, UC Berkeley law professor Franklin Zimring told the Guardian, so they mean different things to the police departments that employ them, muddying the waters of the current debate.

“When labels get popular, they get pasted into lots of different things,” said Zimring, who wrote The Great American Crime Decline (Oxford University Press, 2006) and is working on a second book about the crime rate drop in the 1990s in New York City, where CompStat orginated. Yet the two models point to differing law enforcement philosophies.

At its most basic, CompStat uses computerized crime mapping software to drive police deployment decisions. It emphasizes lowering a city’s crime rate by centralizing authority, spotting statistical trends, and targeting crime hot spots. Community policing, a model embraced by many U.S. police departments in the 1980s and ’90s before CompStat swept the nation, grounds police officers in the neighborhoods they serve, decentralizing authority. The model seeks to prevent crime with regular patrols that develop relationships on their beats and lets the community help set law enforcement priorities.

“There is not community policing in San Francisco,” Mirkarimi — the only member of the board to go through the police academy — told the Guardian. “I don’t care what anybody says. If they say there is, then it is isolated. It’s unique to that particular experience or location.”

Proponents of CompStat insist the new model is really just a part of community policing. Gascón wrote a letter to the Board of Supervisors in February saying the proposed legislation “oversteps the jurisdiction of the legislative branch,” “attempts to give district station captains authority and discretion that rightfully belong to the chief of police,” and “will deprive the department of the flexibility it needs to address public safety throughout the city.”

Mirkarimi doesn’t oppose CompStat and said he sees merit in the program’s statistical collection, which has long been a shortcoming in the SFPD. “But I caution against any over-reliance on CompStat as a method that dictates how policing and public safety should be applied,” Mirkarimi told us. “Because the casualty of this over-reliance will be a compromising of any hopes of having true community policing.”

The SFPD website portrays CompStat as starting with data collection and then, similar to community policing, encouraging officers to find creative solutions to ongoing problems, anything from singular incidents of burglary to repeated graffiti or even a spike in murders. The crime triangle, a lasting symbol of community policing, illustrates that victims, suspects, and locations are all necessary for crime to thrive, and successfully policing even one of those factors can prevent crime. But CompStat programs often lack sustained commitment to building relationships with neighborhoods.

“Compstat seemed to engender a pattern of organizational response to crime spikes in hot spots that was analogous to the Whack-a-Mole game found at fairs and carnivals,” argued a 2003 study commissioned by the national Police Foundation titled “CompStat in Practice: An in-depth Analysis of Three Cities.”

The study found immediate contradictions in Lowell, Mass.; Minneapolis, and Newark, N.J. between beat officers’ new responsibility to “simply follow their superiors’ orders” and the community policing model that cast them as individual, authoritative protectors of their neighborhoods. CompStat centralizes authority with the higher echelons of SFPD. It includes bimonthly meetings in which station captains are grilled by SFPD brass and are expected to answer for the statistics in their district.

“Given the gap between the two models of policing, CompStat naturally tends to encounter the greatest resistance in departments that are most committed to community policing,” the study found.

Understaffed and poorly trained crime analysis units tasked with deciphering data patterns into useful correlations (for example, between drug crimes and murder) was another barrier to the success of CompStat outlined in the study. SFPD’s crime analysis unit consists of three civilians housed at the Hall of Justice, SFPD spokesperson Lt. Lyn Tomioka told us. They are not deployed to district stations and are supervised by a lieutenant who also has other responsibilities.

“There are a lot of rough edges. There’s a lot of non-fit there,” Zimring told the Guardian. “Who sets the priorities? CompStat priorities are always crime prevention, and they are set, and tactics are provided, by the chief of police. He is, in the immortal words of George W. Bush, ‘the decider.’ Community policing is supposed to be more cooperative and organic.”

Gascón initiated CompStat in San Francisco in October 2009, although Mayor Gavin Newsom has been touting the CompStat model since he first ran for mayor in 2003, when a campaign policy brief gushed about its “accurate and timely intelligence, rapid deployment, effective tactics, and relentless follow-up and assessment.” Initially, however, SFPD only took baby steps, using a confusing plot system to map crimes. That changed when Gascón took over as police chief last August, bringing experience in the program with him from the Los Angeles Police Department.

SFPD officials say vendor contract costs to start the system’s electronic crime mapping were less than $1 million, and an additional $1 million has been proposed for next year’s budget for technology upgrades in the CompStat unit. But the numbers so far haven’t backed up the boldest claims. SFPD reports 24 homicides this year as of June 12, up 20 percent from last year’s rate for early June. Homicide arrests are down from 12 last year to eight this year. Occurrences of rape are also up by 12 percent, but overall violent crime is down 2 percent compared to this time last year.

Gascón wrote that foot patrols are a valuable tool for community policing in San Francisco, but he doesn’t want to be forced to maintain them with limited staffing. Newsom’s proposed budget maintains current SFPD staffing, 2,317 sworn officers, while many other city departments received deep staffing cuts. Progressive supervisors have pledged to closely scrutinize SFPD’S budget.

Community policing was law enforcement’s response to civil unrest in the 1960s and ’70s, when police were seen as the enforcers of institutional power. Previous beat patrol methods largely ended when the 911 system came along, and the emphasis was placed on calls for service, statistics, and response times, leaving officers with little time to patrol and prevent crime.

The change to community policing emphasized neighborhood input and officers becoming an organic part of the community they served. Citizen contributions, generally through community meetings, began to drive decision-making. Foot patrols were revived and officers were once again expected to have a physical presence and a connection to the community they served.

That change was seen as particularly important in poor neighborhoods and communities of color, where police can sometimes be seen as an occupying army and residents were reluctant to cooperate with investigations. Officials hoped to prevent crimes by showing a presence in neighborhoods rather than simply reacting to them when someone called.

Mirkarimi says a CompStat-driven police force would be a return to that reactive model, potentially sacrificing the long-term commitment required to build trust between a neighborhood and its police department, which is central to community policing. “[CompStat] undermines the principles and practices of community policing because true community policing requires a discipline and a protocol that is sustained,” he said.

While either approach can theoretically result in the same practices, such as a foot beat patrol in a given neighborhood, Zimring said the reasoning behind it depends on the model. “CompStat to begin with is completely crime-driven,” Zimring said. “The reason you have it is to reduce crimes. It involves computerized mapping of crimes. It involves allocating resources to so-called hot spots, and it involves the police department imposing its own priorities as opposed to implementing community priorities.”

The Board of Supervisors will consider Mirkarimi’s measure and SFPD budget in July, airing a debate that could continue on to the November ballot, when voters would decide whether to maintain their faith in CompStat and the SFPD or ask for more community policing and foot patrols.